diff --git a/CheckGuard Models/wholeimage/amount/finetune_lora_llava_mistral.sh b/CheckGuard Models/wholeimage/amount/finetune_lora_llava_mistral.sh
new file mode 100644
index 0000000000000000000000000000000000000000..3d69789851cd6bb59694dc85be15c81fbee7041d
--- /dev/null
+++ b/CheckGuard Models/wholeimage/amount/finetune_lora_llava_mistral.sh
@@ -0,0 +1,39 @@
+#!/bin/bash
+
+deepspeed llava/train/train_mem.py \
+ --lora_enable True --lora_r 128 --lora_alpha 256 --mm_projector_lr 2e-5 \
+ --deepspeed ./scripts/zero3.json \
+ --model_name_or_path liuhaotian/llava-v1.6-mistral-7b \
+ --version mistral_instruct \
+ --data_path /home/larry5/project/LLaVA-1.6-ft/data/peft/amount/modified_path_to_train_val_human-gpt-whole-check.json \
+ --image_folder /home/larry5/project/LLaVA-1.6-ft/data/data/ \
+ --vision_tower openai/clip-vit-large-patch14-336 \
+ --mm_projector_type mlp2x_gelu \
+ --mm_vision_select_layer -2 \
+ --mm_use_im_start_end False \
+ --mm_use_im_patch_token False \
+ --mm_patch_merge_type spatial_unpad \
+ --image_aspect_ratio anyres \
+ --group_by_modality_length False \
+ --bf16 False \
+ --fp16 True \
+ --output_dir /home/larry5/project/LLaVA-1.6-ft/scripts_peft/mistral/lora/llava-lora-mistral-r128a256/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model \
+ --num_train_epochs 1 \
+ --per_device_train_batch_size 10 \
+ --per_device_eval_batch_size 1 \
+ --gradient_accumulation_steps 1 \
+ --evaluation_strategy "no" \
+ --save_strategy "steps" \
+ --save_steps 500 \
+ --save_total_limit 5 \
+ --learning_rate 2e-5 \
+ --weight_decay 0. \
+ --warmup_ratio 0.05 \
+ --lr_scheduler_type "cosine" \
+ --logging_steps 1 \
+ --tf32 True \
+ --model_max_length 4096 \
+ --gradient_checkpointing True \
+ --dataloader_num_workers 4 \
+ --lazy_preprocess True \
+ --report_to wandb \
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/README.md b/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..bdb138eee6972419f6d60676388b52fd99ec478e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/README.md
@@ -0,0 +1,202 @@
+---
+library_name: peft
+base_model: liuhaotian/llava-v1.6-mistral-7b
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.10.0
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/adapter_config.json b/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..4bed3c21a88535e71ce0984d0e3e6b1f2fbfe658
--- /dev/null
+++ b/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/adapter_config.json
@@ -0,0 +1,34 @@
+{
+ "alpha_pattern": {},
+ "auto_mapping": null,
+ "base_model_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "bias": "none",
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 256,
+ "lora_dropout": 0.05,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "r": 128,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "up_proj",
+ "k_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "v_proj",
+ "gate_proj"
+ ],
+ "task_type": "CAUSAL_LM",
+ "use_dora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors b/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..76151f76d7d9149ae26aff57d012d19e73357194
--- /dev/null
+++ b/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf2247d527585fc799edb906388c1be818fe3bb61e79cbe1b59d3311b2b6e5e9
+size 708924928
diff --git a/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/config.json b/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..93e133af45036a778791b5679a8953a4f6a35a33
--- /dev/null
+++ b/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/config.json
@@ -0,0 +1,70 @@
+{
+ "_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "architectures": [
+ "LlavaMistralForCausalLM"
+ ],
+ "attention_dropout": 0.0,
+ "bos_token_id": 1,
+ "eos_token_id": 2,
+ "freeze_mm_mlp_adapter": false,
+ "freeze_mm_vision_resampler": false,
+ "hidden_act": "silu",
+ "hidden_size": 4096,
+ "image_aspect_ratio": "anyres",
+ "image_crop_resolution": 224,
+ "image_grid_pinpoints": [
+ [
+ 336,
+ 672
+ ],
+ [
+ 672,
+ 336
+ ],
+ [
+ 672,
+ 672
+ ],
+ [
+ 1008,
+ 336
+ ],
+ [
+ 336,
+ 1008
+ ]
+ ],
+ "image_split_resolution": 224,
+ "initializer_range": 0.02,
+ "intermediate_size": 14336,
+ "max_position_embeddings": 32768,
+ "mm_hidden_size": 1024,
+ "mm_patch_merge_type": "spatial_unpad",
+ "mm_projector_lr": 2e-05,
+ "mm_projector_type": "mlp2x_gelu",
+ "mm_resampler_type": null,
+ "mm_use_im_patch_token": false,
+ "mm_use_im_start_end": false,
+ "mm_vision_select_feature": "patch",
+ "mm_vision_select_layer": -2,
+ "mm_vision_tower": "openai/clip-vit-large-patch14-336",
+ "mm_vision_tower_lr": 2e-06,
+ "model_type": "llava_mistral",
+ "num_attention_heads": 32,
+ "num_hidden_layers": 32,
+ "num_key_value_heads": 8,
+ "rms_norm_eps": 1e-05,
+ "rope_theta": 1000000.0,
+ "sliding_window": null,
+ "tie_word_embeddings": false,
+ "tokenizer_model_max_length": 4096,
+ "tokenizer_padding_side": "right",
+ "torch_dtype": "bfloat16",
+ "transformers_version": "4.37.2",
+ "tune_mm_mlp_adapter": false,
+ "tune_mm_vision_resampler": false,
+ "unfreeze_mm_vision_tower": true,
+ "use_cache": true,
+ "use_mm_proj": true,
+ "vocab_size": 32000
+}
diff --git a/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin b/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin
new file mode 100644
index 0000000000000000000000000000000000000000..cd99cf2428d96ad5af062e0a9af0e361ee2567d7
--- /dev/null
+++ b/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:122abc8f0749c93d63088c8fbc3c18949d0e6fe8a9c9bc719442920c7224b9fc
+size 41961648
diff --git a/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/trainer_state.json b/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..3be7d6e597c15e421ee50d5053e16abd97a37d09
--- /dev/null
+++ b/CheckGuard Models/wholeimage/amount/llava-lora-mistral-r128a256-10BS-model/trainer_state.json
@@ -0,0 +1,2010 @@
+{
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.0,
+ "eval_steps": 500,
+ "global_step": 330,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "epoch": 0.0,
+ "learning_rate": 1.1764705882352942e-06,
+ "loss": 0.2385,
+ "step": 1
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.3529411764705885e-06,
+ "loss": 0.2477,
+ "step": 2
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.3529411764705885e-06,
+ "loss": 0.297,
+ "step": 3
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 3.529411764705883e-06,
+ "loss": 0.1694,
+ "step": 4
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 4.705882352941177e-06,
+ "loss": 0.1294,
+ "step": 5
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 5.882352941176471e-06,
+ "loss": 0.1461,
+ "step": 6
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.058823529411766e-06,
+ "loss": 0.1272,
+ "step": 7
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.058823529411766e-06,
+ "loss": 0.1176,
+ "step": 8
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 8.23529411764706e-06,
+ "loss": 0.0666,
+ "step": 9
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 9.411764705882354e-06,
+ "loss": 0.1199,
+ "step": 10
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.0588235294117648e-05,
+ "loss": 0.1216,
+ "step": 11
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.1764705882352942e-05,
+ "loss": 0.1258,
+ "step": 12
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.2941176470588238e-05,
+ "loss": 0.0381,
+ "step": 13
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.4117647058823532e-05,
+ "loss": 0.0318,
+ "step": 14
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.5294117647058822e-05,
+ "loss": 0.1278,
+ "step": 15
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.647058823529412e-05,
+ "loss": 0.1014,
+ "step": 16
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.7647058823529414e-05,
+ "loss": 0.0555,
+ "step": 17
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.8823529411764708e-05,
+ "loss": 0.0576,
+ "step": 18
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 2e-05,
+ "loss": 0.0987,
+ "step": 19
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9999496293646753e-05,
+ "loss": 0.1177,
+ "step": 20
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.999798522533102e-05,
+ "loss": 0.1623,
+ "step": 21
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9995466947279753e-05,
+ "loss": 0.0899,
+ "step": 22
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9991941713187477e-05,
+ "loss": 0.0615,
+ "step": 23
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9987409878190752e-05,
+ "loss": 0.068,
+ "step": 24
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.99818718988324e-05,
+ "loss": 0.0909,
+ "step": 25
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9975328333015497e-05,
+ "loss": 0.0658,
+ "step": 26
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9967779839947172e-05,
+ "loss": 0.0251,
+ "step": 27
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9959227180072216e-05,
+ "loss": 0.0526,
+ "step": 28
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9949671214996448e-05,
+ "loss": 0.0495,
+ "step": 29
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.993911290739993e-05,
+ "loss": 0.034,
+ "step": 30
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.992755332093999e-05,
+ "loss": 0.0678,
+ "step": 31
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9914993620144055e-05,
+ "loss": 0.063,
+ "step": 32
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.990143507029234e-05,
+ "loss": 0.0237,
+ "step": 33
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9886879037290385e-05,
+ "loss": 0.0773,
+ "step": 34
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9871326987531453e-05,
+ "loss": 0.0357,
+ "step": 35
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.98547804877488e-05,
+ "loss": 0.1064,
+ "step": 36
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.983724120485783e-05,
+ "loss": 0.083,
+ "step": 37
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9818710905788195e-05,
+ "loss": 0.1053,
+ "step": 38
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9799191457305767e-05,
+ "loss": 0.0343,
+ "step": 39
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.977868482582459e-05,
+ "loss": 0.1018,
+ "step": 40
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9757193077208776e-05,
+ "loss": 0.1488,
+ "step": 41
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9734718376564386e-05,
+ "loss": 0.0511,
+ "step": 42
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9711262988021322e-05,
+ "loss": 0.0643,
+ "step": 43
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.968682927450523e-05,
+ "loss": 0.0184,
+ "step": 44
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9661419697499455e-05,
+ "loss": 0.0483,
+ "step": 45
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9635036816797072e-05,
+ "loss": 0.1308,
+ "step": 46
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.960768329024301e-05,
+ "loss": 0.0618,
+ "step": 47
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.957936187346628e-05,
+ "loss": 0.0513,
+ "step": 48
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.955007541960241e-05,
+ "loss": 0.0517,
+ "step": 49
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9519826879005964e-05,
+ "loss": 0.0638,
+ "step": 50
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.948861929895336e-05,
+ "loss": 0.0841,
+ "step": 51
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.945645582333587e-05,
+ "loss": 0.0994,
+ "step": 52
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9423339692342885e-05,
+ "loss": 0.0906,
+ "step": 53
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9389274242135528e-05,
+ "loss": 0.2008,
+ "step": 54
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9354262904510544e-05,
+ "loss": 0.0152,
+ "step": 55
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9318309206554567e-05,
+ "loss": 0.0232,
+ "step": 56
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9281416770288806e-05,
+ "loss": 0.057,
+ "step": 57
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.924358931230418e-05,
+ "loss": 0.1069,
+ "step": 58
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.920483064338687e-05,
+ "loss": 0.034,
+ "step": 59
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9165144668134426e-05,
+ "loss": 0.052,
+ "step": 60
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9124535384562423e-05,
+ "loss": 0.1445,
+ "step": 61
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.9083006883701688e-05,
+ "loss": 0.0578,
+ "step": 62
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.904056334918617e-05,
+ "loss": 0.0426,
+ "step": 63
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8997209056831462e-05,
+ "loss": 0.0214,
+ "step": 64
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8952948374204066e-05,
+ "loss": 0.084,
+ "step": 65
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8907785760181392e-05,
+ "loss": 0.1055,
+ "step": 66
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8861725764502557e-05,
+ "loss": 0.0333,
+ "step": 67
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.881477302731006e-05,
+ "loss": 0.0334,
+ "step": 68
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.87669322786823e-05,
+ "loss": 0.0513,
+ "step": 69
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8718208338157082e-05,
+ "loss": 0.0324,
+ "step": 70
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.866860611424609e-05,
+ "loss": 0.033,
+ "step": 71
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8618130603940386e-05,
+ "loss": 0.0379,
+ "step": 72
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.856678689220701e-05,
+ "loss": 0.0403,
+ "step": 73
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.851458015147673e-05,
+ "loss": 0.0489,
+ "step": 74
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.846151564112294e-05,
+ "loss": 0.0599,
+ "step": 75
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.840759870693184e-05,
+ "loss": 0.0654,
+ "step": 76
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8352834780563888e-05,
+ "loss": 0.0526,
+ "step": 77
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8297229379006614e-05,
+ "loss": 0.0105,
+ "step": 78
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8240788104018824e-05,
+ "loss": 0.0394,
+ "step": 79
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8183516641566278e-05,
+ "loss": 0.0573,
+ "step": 80
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8125420761248878e-05,
+ "loss": 0.0478,
+ "step": 81
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.806650631571943e-05,
+ "loss": 0.0633,
+ "step": 82
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8006779240094024e-05,
+ "loss": 0.0423,
+ "step": 83
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7946245551354156e-05,
+ "loss": 0.0618,
+ "step": 84
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7884911347740556e-05,
+ "loss": 0.0658,
+ "step": 85
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.782278280813882e-05,
+ "loss": 0.0485,
+ "step": 86
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.775986619145697e-05,
+ "loss": 0.0468,
+ "step": 87
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7696167835994927e-05,
+ "loss": 0.0558,
+ "step": 88
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7631694158805945e-05,
+ "loss": 0.0518,
+ "step": 89
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7566451655050197e-05,
+ "loss": 0.0684,
+ "step": 90
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7500446897340408e-05,
+ "loss": 0.0304,
+ "step": 91
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7433686535079736e-05,
+ "loss": 0.0397,
+ "step": 92
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.736617729379191e-05,
+ "loss": 0.1084,
+ "step": 93
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7297925974443675e-05,
+ "loss": 0.0826,
+ "step": 94
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7228939452759666e-05,
+ "loss": 0.0309,
+ "step": 95
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7159224678529734e-05,
+ "loss": 0.0348,
+ "step": 96
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7088788674908817e-05,
+ "loss": 0.097,
+ "step": 97
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.7017638537709426e-05,
+ "loss": 0.0897,
+ "step": 98
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6945781434686783e-05,
+ "loss": 0.0614,
+ "step": 99
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6873224604816753e-05,
+ "loss": 0.08,
+ "step": 100
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.679997535756657e-05,
+ "loss": 0.0126,
+ "step": 101
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.672604107215848e-05,
+ "loss": 0.0644,
+ "step": 102
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6651429196826337e-05,
+ "loss": 0.0702,
+ "step": 103
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6576147248065268e-05,
+ "loss": 0.0809,
+ "step": 104
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6500202809874446e-05,
+ "loss": 0.0354,
+ "step": 105
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6423603532993074e-05,
+ "loss": 0.1429,
+ "step": 106
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.634635713412964e-05,
+ "loss": 0.1142,
+ "step": 107
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.626847139518452e-05,
+ "loss": 0.034,
+ "step": 108
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.618995416246601e-05,
+ "loss": 0.0818,
+ "step": 109
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6110813345899914e-05,
+ "loss": 0.0594,
+ "step": 110
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.6031056918232642e-05,
+ "loss": 0.0958,
+ "step": 111
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.595069291422807e-05,
+ "loss": 0.0418,
+ "step": 112
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.586972942985807e-05,
+ "loss": 0.0315,
+ "step": 113
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5788174621486936e-05,
+ "loss": 0.0435,
+ "step": 114
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.570603670504969e-05,
+ "loss": 0.0596,
+ "step": 115
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.570603670504969e-05,
+ "loss": 0.048,
+ "step": 116
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5623323955224404e-05,
+ "loss": 0.0352,
+ "step": 117
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5540044704598588e-05,
+ "loss": 0.0264,
+ "step": 118
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5456207342829777e-05,
+ "loss": 0.0378,
+ "step": 119
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5371820315800316e-05,
+ "loss": 0.0519,
+ "step": 120
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5286892124766546e-05,
+ "loss": 0.0559,
+ "step": 121
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5201431325502332e-05,
+ "loss": 0.0708,
+ "step": 122
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5115446527437193e-05,
+ "loss": 0.0823,
+ "step": 123
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.5028946392788934e-05,
+ "loss": 0.0345,
+ "step": 124
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4941939635691036e-05,
+ "loss": 0.111,
+ "step": 125
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4854435021314766e-05,
+ "loss": 0.0284,
+ "step": 126
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4766441364986162e-05,
+ "loss": 0.1226,
+ "step": 127
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.467796753129797e-05,
+ "loss": 0.022,
+ "step": 128
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4589022433216616e-05,
+ "loss": 0.0565,
+ "step": 129
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4499615031184297e-05,
+ "loss": 0.0875,
+ "step": 130
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4409754332216303e-05,
+ "loss": 0.0457,
+ "step": 131
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.431944938899363e-05,
+ "loss": 0.0776,
+ "step": 132
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4228709298950998e-05,
+ "loss": 0.0397,
+ "step": 133
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.4137543203360382e-05,
+ "loss": 0.0278,
+ "step": 134
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.4045960286410093e-05,
+ "loss": 0.0204,
+ "step": 135
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.395396977427955e-05,
+ "loss": 0.0531,
+ "step": 136
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3861580934209832e-05,
+ "loss": 0.0334,
+ "step": 137
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.376880307357009e-05,
+ "loss": 0.0389,
+ "step": 138
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3675645538919884e-05,
+ "loss": 0.0571,
+ "step": 139
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3582117715067628e-05,
+ "loss": 0.0352,
+ "step": 140
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3488229024125142e-05,
+ "loss": 0.0062,
+ "step": 141
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3393988924558445e-05,
+ "loss": 0.0489,
+ "step": 142
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3299406910234917e-05,
+ "loss": 0.068,
+ "step": 143
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3204492509466862e-05,
+ "loss": 0.0478,
+ "step": 144
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3109255284051615e-05,
+ "loss": 0.1145,
+ "step": 145
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3013704828308276e-05,
+ "loss": 0.0253,
+ "step": 146
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2917850768111171e-05,
+ "loss": 0.0138,
+ "step": 147
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.282170275992012e-05,
+ "loss": 0.1083,
+ "step": 148
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2725270489807637e-05,
+ "loss": 0.0708,
+ "step": 149
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2628563672483147e-05,
+ "loss": 0.0144,
+ "step": 150
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2531592050314308e-05,
+ "loss": 0.0628,
+ "step": 151
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2434365392345553e-05,
+ "loss": 0.0475,
+ "step": 152
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2336893493313946e-05,
+ "loss": 0.0161,
+ "step": 153
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.223918617266245e-05,
+ "loss": 0.084,
+ "step": 154
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2141253273550698e-05,
+ "loss": 0.0494,
+ "step": 155
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2043104661863386e-05,
+ "loss": 0.0293,
+ "step": 156
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1944750225216363e-05,
+ "loss": 0.0837,
+ "step": 157
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1846199871960557e-05,
+ "loss": 0.0479,
+ "step": 158
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1747463530183781e-05,
+ "loss": 0.0752,
+ "step": 159
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1648551146710557e-05,
+ "loss": 0.0242,
+ "step": 160
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1549472686100079e-05,
+ "loss": 0.0322,
+ "step": 161
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.145023812964237e-05,
+ "loss": 0.0812,
+ "step": 162
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1350857474352734e-05,
+ "loss": 0.0133,
+ "step": 163
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1251340731964664e-05,
+ "loss": 0.093,
+ "step": 164
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1151697927921242e-05,
+ "loss": 0.0377,
+ "step": 165
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1051939100365154e-05,
+ "loss": 0.0561,
+ "step": 166
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0952074299127451e-05,
+ "loss": 0.0556,
+ "step": 167
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0852113584715103e-05,
+ "loss": 0.0567,
+ "step": 168
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0752067027297486e-05,
+ "loss": 0.0722,
+ "step": 169
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.065194470569193e-05,
+ "loss": 0.0419,
+ "step": 170
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0551756706348331e-05,
+ "loss": 0.04,
+ "step": 171
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0451513122333042e-05,
+ "loss": 0.0227,
+ "step": 172
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.035122405231209e-05,
+ "loss": 0.0136,
+ "step": 173
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0250899599533833e-05,
+ "loss": 0.0613,
+ "step": 174
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0150549870811108e-05,
+ "loss": 0.02,
+ "step": 175
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0050184975503104e-05,
+ "loss": 0.0232,
+ "step": 176
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.949815024496901e-06,
+ "loss": 0.0229,
+ "step": 177
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.849450129188895e-06,
+ "loss": 0.0483,
+ "step": 178
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.74910040046617e-06,
+ "loss": 0.0153,
+ "step": 179
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.648775947687914e-06,
+ "loss": 0.0121,
+ "step": 180
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.548486877666963e-06,
+ "loss": 0.0603,
+ "step": 181
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.448243293651676e-06,
+ "loss": 0.0291,
+ "step": 182
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.348055294308074e-06,
+ "loss": 0.0458,
+ "step": 183
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.247932972702514e-06,
+ "loss": 0.0438,
+ "step": 184
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.147886415284903e-06,
+ "loss": 0.0337,
+ "step": 185
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.047925700872552e-06,
+ "loss": 0.0527,
+ "step": 186
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.948060899634846e-06,
+ "loss": 0.0457,
+ "step": 187
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.848302072078762e-06,
+ "loss": 0.0354,
+ "step": 188
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.748659268035339e-06,
+ "loss": 0.0153,
+ "step": 189
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.649142525647271e-06,
+ "loss": 0.1229,
+ "step": 190
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.549761870357633e-06,
+ "loss": 0.0149,
+ "step": 191
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.450527313899923e-06,
+ "loss": 0.042,
+ "step": 192
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.351448853289448e-06,
+ "loss": 0.0146,
+ "step": 193
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.25253646981622e-06,
+ "loss": 0.0145,
+ "step": 194
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.153800128039441e-06,
+ "loss": 0.0691,
+ "step": 195
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.05524977478364e-06,
+ "loss": 0.0165,
+ "step": 196
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.956895338136618e-06,
+ "loss": 0.0776,
+ "step": 197
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.858746726449309e-06,
+ "loss": 0.046,
+ "step": 198
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.760813827337555e-06,
+ "loss": 0.0511,
+ "step": 199
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.663106506686057e-06,
+ "loss": 0.056,
+ "step": 200
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.565634607654453e-06,
+ "loss": 0.0084,
+ "step": 201
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.468407949685695e-06,
+ "loss": 0.0492,
+ "step": 202
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.371436327516854e-06,
+ "loss": 0.0472,
+ "step": 203
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.274729510192367e-06,
+ "loss": 0.0826,
+ "step": 204
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.1782972400798825e-06,
+ "loss": 0.062,
+ "step": 205
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.082149231888833e-06,
+ "loss": 0.0373,
+ "step": 206
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.986295171691727e-06,
+ "loss": 0.0227,
+ "step": 207
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.890744715948388e-06,
+ "loss": 0.029,
+ "step": 208
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.795507490533142e-06,
+ "loss": 0.0606,
+ "step": 209
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.700593089765086e-06,
+ "loss": 0.0784,
+ "step": 210
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.606011075441556e-06,
+ "loss": 0.0254,
+ "step": 211
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.511770975874862e-06,
+ "loss": 0.0669,
+ "step": 212
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.417882284932373e-06,
+ "loss": 0.0754,
+ "step": 213
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.324354461080121e-06,
+ "loss": 0.0171,
+ "step": 214
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.231196926429913e-06,
+ "loss": 0.054,
+ "step": 215
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.138419065790169e-06,
+ "loss": 0.0068,
+ "step": 216
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 6.046030225720456e-06,
+ "loss": 0.0406,
+ "step": 217
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.95403971358991e-06,
+ "loss": 0.0413,
+ "step": 218
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.86245679663962e-06,
+ "loss": 0.0359,
+ "step": 219
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.7712907010490036e-06,
+ "loss": 0.0398,
+ "step": 220
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.680550611006372e-06,
+ "loss": 0.0679,
+ "step": 221
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.590245667783701e-06,
+ "loss": 0.0356,
+ "step": 222
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.5003849688157075e-06,
+ "loss": 0.0169,
+ "step": 223
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.4109775667833866e-06,
+ "loss": 0.0432,
+ "step": 224
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.322032468702037e-06,
+ "loss": 0.012,
+ "step": 225
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.233558635013842e-06,
+ "loss": 0.0554,
+ "step": 226
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.145564978685234e-06,
+ "loss": 0.065,
+ "step": 227
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.058060364308965e-06,
+ "loss": 0.0139,
+ "step": 228
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.971053607211069e-06,
+ "loss": 0.0117,
+ "step": 229
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.884553472562809e-06,
+ "loss": 0.0763,
+ "step": 230
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.7985686744976714e-06,
+ "loss": 0.0341,
+ "step": 231
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.713107875233459e-06,
+ "loss": 0.0602,
+ "step": 232
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.628179684199685e-06,
+ "loss": 0.0102,
+ "step": 233
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.543792657170228e-06,
+ "loss": 0.0552,
+ "step": 234
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.459955295401415e-06,
+ "loss": 0.0356,
+ "step": 235
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.376676044775601e-06,
+ "loss": 0.0439,
+ "step": 236
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.293963294950313e-06,
+ "loss": 0.0109,
+ "step": 237
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.211825378513066e-06,
+ "loss": 0.0224,
+ "step": 238
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.130270570141931e-06,
+ "loss": 0.0378,
+ "step": 239
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 4.0493070857719305e-06,
+ "loss": 0.0714,
+ "step": 240
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.968943081767358e-06,
+ "loss": 0.0165,
+ "step": 241
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.889186654100089e-06,
+ "loss": 0.0637,
+ "step": 242
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.81004583753399e-06,
+ "loss": 0.0066,
+ "step": 243
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.7315286048154862e-06,
+ "loss": 0.0178,
+ "step": 244
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6536428658703594e-06,
+ "loss": 0.0485,
+ "step": 245
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.576396467006925e-06,
+ "loss": 0.0283,
+ "step": 246
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.4997971901255588e-06,
+ "loss": 0.0694,
+ "step": 247
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.4238527519347353e-06,
+ "loss": 0.0552,
+ "step": 248
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.3485708031736698e-06,
+ "loss": 0.0436,
+ "step": 249
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.2739589278415252e-06,
+ "loss": 0.0881,
+ "step": 250
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.2000246424334315e-06,
+ "loss": 0.039,
+ "step": 251
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.1267753951832523e-06,
+ "loss": 0.0411,
+ "step": 252
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 3.0542185653132216e-06,
+ "loss": 0.0676,
+ "step": 253
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.982361462290575e-06,
+ "loss": 0.0206,
+ "step": 254
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.9112113250911844e-06,
+ "loss": 0.0827,
+ "step": 255
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.8407753214702694e-06,
+ "loss": 0.0102,
+ "step": 256
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.7710605472403373e-06,
+ "loss": 0.0229,
+ "step": 257
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.702074025556327e-06,
+ "loss": 0.0196,
+ "step": 258
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.6338227062080924e-06,
+ "loss": 0.0625,
+ "step": 259
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.566313464920265e-06,
+ "loss": 0.0595,
+ "step": 260
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4995531026595952e-06,
+ "loss": 0.0147,
+ "step": 261
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4335483449498053e-06,
+ "loss": 0.0544,
+ "step": 262
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.3683058411940563e-06,
+ "loss": 0.0453,
+ "step": 263
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.3038321640050763e-06,
+ "loss": 0.0609,
+ "step": 264
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.2401338085430326e-06,
+ "loss": 0.0504,
+ "step": 265
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.177217191861183e-06,
+ "loss": 0.0248,
+ "step": 266
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.115088652259446e-06,
+ "loss": 0.0616,
+ "step": 267
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.053754448645846e-06,
+ "loss": 0.0408,
+ "step": 268
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.9932207599059782e-06,
+ "loss": 0.0444,
+ "step": 269
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.933493684280574e-06,
+ "loss": 0.0632,
+ "step": 270
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8745792387511241e-06,
+ "loss": 0.019,
+ "step": 271
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8164833584337216e-06,
+ "loss": 0.0266,
+ "step": 272
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.75921189598118e-06,
+ "loss": 0.0499,
+ "step": 273
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.7027706209933903e-06,
+ "loss": 0.0379,
+ "step": 274
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6471652194361131e-06,
+ "loss": 0.0092,
+ "step": 275
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.5924012930681643e-06,
+ "loss": 0.0326,
+ "step": 276
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.5384843588770626e-06,
+ "loss": 0.0179,
+ "step": 277
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4854198485232696e-06,
+ "loss": 0.0641,
+ "step": 278
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.433213107792991e-06,
+ "loss": 0.0285,
+ "step": 279
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3818693960596186e-06,
+ "loss": 0.0365,
+ "step": 280
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3313938857539133e-06,
+ "loss": 0.0873,
+ "step": 281
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.2817916618429194e-06,
+ "loss": 0.0148,
+ "step": 282
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.2330677213177034e-06,
+ "loss": 0.0127,
+ "step": 283
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1852269726899423e-06,
+ "loss": 0.0028,
+ "step": 284
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.138274235497443e-06,
+ "loss": 0.0241,
+ "step": 285
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.0922142398186097e-06,
+ "loss": 0.0466,
+ "step": 286
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.0470516257959351e-06,
+ "loss": 0.0273,
+ "step": 287
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.00279094316854e-06,
+ "loss": 0.0529,
+ "step": 288
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 9.594366508138352e-07,
+ "loss": 0.0648,
+ "step": 289
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 9.169931162983137e-07,
+ "loss": 0.0118,
+ "step": 290
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.754646154375801e-07,
+ "loss": 0.0321,
+ "step": 291
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.348553318655795e-07,
+ "loss": 0.0167,
+ "step": 292
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.951693566131325e-07,
+ "loss": 0.0204,
+ "step": 293
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.564106876958188e-07,
+ "loss": 0.0502,
+ "step": 294
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.185832297111939e-07,
+ "loss": 0.0075,
+ "step": 295
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.816907934454353e-07,
+ "loss": 0.053,
+ "step": 296
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.457370954894582e-07,
+ "loss": 0.0255,
+ "step": 297
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.107257578644721e-07,
+ "loss": 0.0422,
+ "step": 298
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.766603076571164e-07,
+ "loss": 0.0745,
+ "step": 299
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.43544176664137e-07,
+ "loss": 0.0217,
+ "step": 300
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.113807010466432e-07,
+ "loss": 0.0417,
+ "step": 301
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.801731209940375e-07,
+ "loss": 0.0324,
+ "step": 302
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.499245803975927e-07,
+ "loss": 0.0264,
+ "step": 303
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.206381265337189e-07,
+ "loss": 0.0285,
+ "step": 304
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.9231670975699354e-07,
+ "loss": 0.0603,
+ "step": 305
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.649631832029288e-07,
+ "loss": 0.0515,
+ "step": 306
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.385803025005463e-07,
+ "loss": 0.0133,
+ "step": 307
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.1317072549477246e-07,
+ "loss": 0.0106,
+ "step": 308
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.887370119786792e-07,
+ "loss": 0.0492,
+ "step": 309
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.6528162343561593e-07,
+ "loss": 0.0132,
+ "step": 310
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.4280692279122554e-07,
+ "loss": 0.0207,
+ "step": 311
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 2.2131517417540937e-07,
+ "loss": 0.0301,
+ "step": 312
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 2.00808542694233e-07,
+ "loss": 0.0294,
+ "step": 313
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.8128909421180506e-07,
+ "loss": 0.0349,
+ "step": 314
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.6275879514217052e-07,
+ "loss": 0.0405,
+ "step": 315
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.4521951225120345e-07,
+ "loss": 0.0464,
+ "step": 316
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.2867301246854757e-07,
+ "loss": 0.0258,
+ "step": 317
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.1312096270961525e-07,
+ "loss": 0.0217,
+ "step": 318
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 9.856492970766296e-08,
+ "loss": 0.0521,
+ "step": 319
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 8.50063798559475e-08,
+ "loss": 0.074,
+ "step": 320
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 7.244667906001202e-08,
+ "loss": 0.0091,
+ "step": 321
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 6.088709260007153e-08,
+ "loss": 0.0584,
+ "step": 322
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 5.032878500355498e-08,
+ "loss": 0.0114,
+ "step": 323
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 4.07728199277857e-08,
+ "loss": 0.0528,
+ "step": 324
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 3.2220160052828245e-08,
+ "loss": 0.0534,
+ "step": 325
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 2.467166698450485e-08,
+ "loss": 0.0912,
+ "step": 326
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.812810116760044e-08,
+ "loss": 0.0426,
+ "step": 327
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.2590121809247235e-08,
+ "loss": 0.0501,
+ "step": 328
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 8.05828681252452e-09,
+ "loss": 0.0135,
+ "step": 329
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 4.5330527202480656e-09,
+ "loss": 0.0375,
+ "step": 330
+ },
+ {
+ "epoch": 1.0,
+ "step": 330,
+ "total_flos": 1006103224320.0,
+ "train_loss": 0.05386477595011732,
+ "train_runtime": 2481.834,
+ "train_samples_per_second": 2.652,
+ "train_steps_per_second": 0.133
+ }
+ ],
+ "logging_steps": 1.0,
+ "max_steps": 330,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 1,
+ "save_steps": 500,
+ "total_flos": 1006103224320.0,
+ "train_batch_size": 10,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/CheckGuard Models/wholeimage/bank/finetune_lora_llava_mistral.sh b/CheckGuard Models/wholeimage/bank/finetune_lora_llava_mistral.sh
new file mode 100644
index 0000000000000000000000000000000000000000..ffa07fd7fef618ad97569cade264920f2ef2acdb
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/finetune_lora_llava_mistral.sh
@@ -0,0 +1,43 @@
+#!/bin/bash
+# Use first parameter as GPU IDs, default to "0,1,2,3" if not provided
+GPU_IDS=${1:-0,1,2,3}
+
+
+CUDA_VISIBLE_DEVICES=0,1,2,3 deepspeed --include localhost:"$GPU_IDS" --master_port 29601\
+ llava/train/train_mem.py \
+ --lora_enable True --lora_r 128 --lora_alpha 256 --mm_projector_lr 2e-5 \
+ --deepspeed ./scripts/zero3.json \
+ --model_name_or_path liuhaotian/llava-v1.6-mistral-7b \
+ --version mistral_instruct \
+ --data_path /home/larry5/project/LLaVA-1.6-ft/data/peft/bank/bank_dataset.json \
+ --image_folder /home/larry5/project/LLaVA-1.6-ft/data/data/ \
+ --vision_tower openai/clip-vit-large-patch14-336 \
+ --mm_projector_type mlp2x_gelu \
+ --mm_vision_select_layer -2 \
+ --mm_use_im_start_end False \
+ --mm_use_im_patch_token False \
+ --mm_patch_merge_type spatial_unpad \
+ --image_aspect_ratio anyres \
+ --group_by_modality_length False \
+ --bf16 False \
+ --fp16 True \
+ --output_dir /home/larry5/project/LLaVA-1.6-ft/scripts_peft/mistral/lora/llava-lora-mistral-r128a256/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model \
+ --num_train_epochs 1 \
+ --per_device_train_batch_size 10 \
+ --per_device_eval_batch_size 1 \
+ --gradient_accumulation_steps 1 \
+ --evaluation_strategy "no" \
+ --save_strategy "steps" \
+ --save_steps 500 \
+ --save_total_limit 5 \
+ --learning_rate 2e-5 \
+ --weight_decay 0. \
+ --warmup_ratio 0.05 \
+ --lr_scheduler_type "cosine" \
+ --logging_steps 1 \
+ --tf32 True \
+ --model_max_length 4096 \
+ --gradient_checkpointing True \
+ --dataloader_num_workers 4 \
+ --lazy_preprocess True \
+ --report_to wandb \
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/README.md b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..bdb138eee6972419f6d60676388b52fd99ec478e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/README.md
@@ -0,0 +1,202 @@
+---
+library_name: peft
+base_model: liuhaotian/llava-v1.6-mistral-7b
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.10.0
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/adapter_config.json b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..163278563d73cc786fd882d1e16e8f934a09391a
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/adapter_config.json
@@ -0,0 +1,34 @@
+{
+ "alpha_pattern": {},
+ "auto_mapping": null,
+ "base_model_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "bias": "none",
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 256,
+ "lora_dropout": 0.05,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "r": 128,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "q_proj",
+ "v_proj",
+ "gate_proj",
+ "up_proj",
+ "o_proj",
+ "k_proj",
+ "down_proj"
+ ],
+ "task_type": "CAUSAL_LM",
+ "use_dora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..a56d5fb1555a33cf575ad0f6981f99fb5236f94a
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b234ad29295da4f261427006e770781d152d27e2bd090a65ac32cfb8472dd11b
+size 708924928
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/README.md b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..bdb138eee6972419f6d60676388b52fd99ec478e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/README.md
@@ -0,0 +1,202 @@
+---
+library_name: peft
+base_model: liuhaotian/llava-v1.6-mistral-7b
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.10.0
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_config.json b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..163278563d73cc786fd882d1e16e8f934a09391a
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_config.json
@@ -0,0 +1,34 @@
+{
+ "alpha_pattern": {},
+ "auto_mapping": null,
+ "base_model_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "bias": "none",
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 256,
+ "lora_dropout": 0.05,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "r": 128,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "q_proj",
+ "v_proj",
+ "gate_proj",
+ "up_proj",
+ "o_proj",
+ "k_proj",
+ "down_proj"
+ ],
+ "task_type": "CAUSAL_LM",
+ "use_dora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_model.safetensors b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..a34ad4e28b0ea0354fbb4cf07335028cbdf7f970
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c26e454370836d2d8ee1827620ba4d532f3b135ebd1e7fcbb0263a086f241253
+size 1417762896
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_model_states.pt b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_model_states.pt
new file mode 100644
index 0000000000000000000000000000000000000000..2148fda79de7d1a6e7f0b5258183d9c00a5fddfc
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3183212032ce3f53bf011c0ea2d72e73d90e4ae83d758f3cb2661945c405d2e
+size 632242
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_optim_states.pt b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000000000000000000000000000000000000..317333c4d067d7497fbb235fb69d5ffa7996f1c2
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59f2e280a5a1fb1c30380e867fb4625232d56bf1130fb2ac1bda1c76272752dd
+size 4504787266
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/latest b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/latest
new file mode 100644
index 0000000000000000000000000000000000000000..f0b47ce15fff9a01b2a416a473b2148085048a50
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/latest
@@ -0,0 +1 @@
+global_step500
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/rng_state.pth b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..b50886c52853405327456894d56d2e49c5f3431b
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42d7cdbb5673ea29475539a9e027f8b9828b8bdf3f8f5a3383b13244fc3604a3
+size 14244
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/scheduler.pt b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..7f5d0e70c3dac2a7b8b5c10da62710a65f9f5497
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:74f73b67322f406ba2e53b1ed170e4b3c50a5de49d1b4aa38bda0b32a3724ada
+size 1064
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/special_tokens_map.json b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..14761dcf1466dc232bd41de9c21d4c617b15755e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/special_tokens_map.json
@@ -0,0 +1,24 @@
+{
+ "bos_token": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false
+ },
+ "eos_token": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false
+ },
+ "pad_token": "",
+ "unk_token": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false
+ }
+}
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer.model b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer.model
new file mode 100644
index 0000000000000000000000000000000000000000..8b443ef19c2a19acc3ac64fb9c3db4a72921dff6
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer.model
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
+size 493443
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer_config.json b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..d0ea5c3458cd84f0062b47fa0476bb328b3e208a
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer_config.json
@@ -0,0 +1,44 @@
+{
+ "add_bos_token": true,
+ "add_eos_token": false,
+ "added_tokens_decoder": {
+ "0": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "1": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "2": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ }
+ },
+ "additional_special_tokens": [],
+ "bos_token": "",
+ "chat_template": "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if message['role'] == 'user' %}{{ '[INST] ' + message['content'] + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ message['content'] + eos_token}}{% else %}{{ raise_exception('Only user and assistant roles are supported!') }}{% endif %}{% endfor %}",
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "",
+ "legacy": true,
+ "model_max_length": 4096,
+ "pad_token": "",
+ "padding_side": "right",
+ "sp_model_kwargs": {},
+ "spaces_between_special_tokens": false,
+ "tokenizer_class": "LlamaTokenizer",
+ "unk_token": "",
+ "use_default_system_prompt": false
+}
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/trainer_state.json b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..c5d268d55767a419e925adace6ec7313aad46dab
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/trainer_state.json
@@ -0,0 +1,3021 @@
+{
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.9009009009009009,
+ "eval_steps": 500,
+ "global_step": 500,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "epoch": 0.0,
+ "learning_rate": 7.142857142857143e-07,
+ "loss": 0.4237,
+ "step": 1
+ },
+ {
+ "epoch": 0.0,
+ "learning_rate": 1.4285714285714286e-06,
+ "loss": 0.3368,
+ "step": 2
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.1428571428571427e-06,
+ "loss": 0.214,
+ "step": 3
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.8571428571428573e-06,
+ "loss": 0.396,
+ "step": 4
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 3.5714285714285718e-06,
+ "loss": 0.305,
+ "step": 5
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 4.2857142857142855e-06,
+ "loss": 0.4049,
+ "step": 6
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 5e-06,
+ "loss": 0.108,
+ "step": 7
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 5.7142857142857145e-06,
+ "loss": 0.2286,
+ "step": 8
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 6.4285714285714295e-06,
+ "loss": 0.1443,
+ "step": 9
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.1428571428571436e-06,
+ "loss": 0.2252,
+ "step": 10
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.857142857142858e-06,
+ "loss": 0.0747,
+ "step": 11
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 8.571428571428571e-06,
+ "loss": 0.1084,
+ "step": 12
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 9.285714285714288e-06,
+ "loss": 0.2115,
+ "step": 13
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1e-05,
+ "loss": 0.4742,
+ "step": 14
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.0714285714285714e-05,
+ "loss": 0.083,
+ "step": 15
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.1428571428571429e-05,
+ "loss": 0.3392,
+ "step": 16
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.2142857142857142e-05,
+ "loss": 0.065,
+ "step": 17
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.2857142857142859e-05,
+ "loss": 0.1711,
+ "step": 18
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.3571428571428574e-05,
+ "loss": 0.0539,
+ "step": 19
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.4285714285714287e-05,
+ "loss": 0.0701,
+ "step": 20
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.5000000000000002e-05,
+ "loss": 0.0836,
+ "step": 21
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.5714285714285715e-05,
+ "loss": 0.1891,
+ "step": 22
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.642857142857143e-05,
+ "loss": 0.0422,
+ "step": 23
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.7142857142857142e-05,
+ "loss": 0.2094,
+ "step": 24
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.785714285714286e-05,
+ "loss": 0.139,
+ "step": 25
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.8571428571428575e-05,
+ "loss": 0.2214,
+ "step": 26
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.928571428571429e-05,
+ "loss": 0.1084,
+ "step": 27
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 2e-05,
+ "loss": 0.0898,
+ "step": 28
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.9999822316445652e-05,
+ "loss": 0.0359,
+ "step": 29
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.9999289272096886e-05,
+ "loss": 0.2648,
+ "step": 30
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9998400885896355e-05,
+ "loss": 0.4007,
+ "step": 31
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9997157189414373e-05,
+ "loss": 0.235,
+ "step": 32
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.999555822684783e-05,
+ "loss": 0.0273,
+ "step": 33
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.999360405501859e-05,
+ "loss": 0.0267,
+ "step": 34
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.99912947433715e-05,
+ "loss": 0.2619,
+ "step": 35
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9988630373971896e-05,
+ "loss": 0.4101,
+ "step": 36
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9985611041502704e-05,
+ "loss": 0.1302,
+ "step": 37
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9982236853261067e-05,
+ "loss": 0.118,
+ "step": 38
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9978507929154534e-05,
+ "loss": 0.0933,
+ "step": 39
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.997442440169681e-05,
+ "loss": 0.0104,
+ "step": 40
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9969986416003026e-05,
+ "loss": 0.1061,
+ "step": 41
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9965194129784597e-05,
+ "loss": 0.1575,
+ "step": 42
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.996004771334361e-05,
+ "loss": 0.1969,
+ "step": 43
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.996004771334361e-05,
+ "loss": 0.0492,
+ "step": 44
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9954547349566783e-05,
+ "loss": 0.3012,
+ "step": 45
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.994869323391895e-05,
+ "loss": 0.2185,
+ "step": 46
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.994248557443613e-05,
+ "loss": 0.1729,
+ "step": 47
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.993592459171812e-05,
+ "loss": 0.0354,
+ "step": 48
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9929010518920667e-05,
+ "loss": 0.3939,
+ "step": 49
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.992174360174717e-05,
+ "loss": 0.0505,
+ "step": 50
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9914124098439976e-05,
+ "loss": 0.0777,
+ "step": 51
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9914124098439976e-05,
+ "loss": 0.6651,
+ "step": 52
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9906152279771162e-05,
+ "loss": 0.15,
+ "step": 53
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9897828429032946e-05,
+ "loss": 0.1416,
+ "step": 54
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9889152842027607e-05,
+ "loss": 0.1195,
+ "step": 55
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9880125827056967e-05,
+ "loss": 0.0787,
+ "step": 56
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.987074770491145e-05,
+ "loss": 0.0681,
+ "step": 57
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.986101880885867e-05,
+ "loss": 0.1337,
+ "step": 58
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9850939484631598e-05,
+ "loss": 0.0961,
+ "step": 59
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.984051009041626e-05,
+ "loss": 0.116,
+ "step": 60
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.982973099683902e-05,
+ "loss": 0.3853,
+ "step": 61
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9818602586953414e-05,
+ "loss": 0.0875,
+ "step": 62
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9807125256226532e-05,
+ "loss": 0.3216,
+ "step": 63
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9795299412524948e-05,
+ "loss": 0.0752,
+ "step": 64
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9783125476100254e-05,
+ "loss": 0.1461,
+ "step": 65
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9770603879574108e-05,
+ "loss": 0.075,
+ "step": 66
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.975773506792287e-05,
+ "loss": 0.0685,
+ "step": 67
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.974451949846177e-05,
+ "loss": 0.0555,
+ "step": 68
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.973095764082869e-05,
+ "loss": 0.0171,
+ "step": 69
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9717049976967437e-05,
+ "loss": 0.0247,
+ "step": 70
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9702797001110642e-05,
+ "loss": 0.0839,
+ "step": 71
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9688199219762183e-05,
+ "loss": 0.4163,
+ "step": 72
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.96732571516792e-05,
+ "loss": 0.1461,
+ "step": 73
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9657971327853644e-05,
+ "loss": 0.1457,
+ "step": 74
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.964234229149342e-05,
+ "loss": 0.0482,
+ "step": 75
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.962637059800307e-05,
+ "loss": 0.0802,
+ "step": 76
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9610056814964053e-05,
+ "loss": 0.0697,
+ "step": 77
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.959340152211455e-05,
+ "loss": 0.0614,
+ "step": 78
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.95764053113289e-05,
+ "loss": 0.1004,
+ "step": 79
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9559068786596526e-05,
+ "loss": 0.0286,
+ "step": 80
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.954139256400049e-05,
+ "loss": 0.1162,
+ "step": 81
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.952337727169561e-05,
+ "loss": 0.0731,
+ "step": 82
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.950502354988612e-05,
+ "loss": 0.0286,
+ "step": 83
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.948633205080292e-05,
+ "loss": 0.2425,
+ "step": 84
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9467303438680414e-05,
+ "loss": 0.0505,
+ "step": 85
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.944793838973289e-05,
+ "loss": 0.0922,
+ "step": 86
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9428237592130487e-05,
+ "loss": 0.2949,
+ "step": 87
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.940820174597476e-05,
+ "loss": 0.2807,
+ "step": 88
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9387831563273775e-05,
+ "loss": 0.2377,
+ "step": 89
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9367127767916828e-05,
+ "loss": 0.2558,
+ "step": 90
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9346091095648712e-05,
+ "loss": 0.0871,
+ "step": 91
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.932472229404356e-05,
+ "loss": 0.2204,
+ "step": 92
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9303022122478303e-05,
+ "loss": 0.1174,
+ "step": 93
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9280991352105656e-05,
+ "loss": 0.2181,
+ "step": 94
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.925863076582674e-05,
+ "loss": 0.1251,
+ "step": 95
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9235941158263253e-05,
+ "loss": 0.2251,
+ "step": 96
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9212923335729206e-05,
+ "loss": 0.1236,
+ "step": 97
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.918957811620231e-05,
+ "loss": 0.0901,
+ "step": 98
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9165906329294875e-05,
+ "loss": 0.1002,
+ "step": 99
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9141908816224356e-05,
+ "loss": 0.4397,
+ "step": 100
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9117586429783433e-05,
+ "loss": 0.1141,
+ "step": 101
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.909294003430972e-05,
+ "loss": 0.1842,
+ "step": 102
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.906797050565505e-05,
+ "loss": 0.0985,
+ "step": 103
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.9042678731154337e-05,
+ "loss": 0.1533,
+ "step": 104
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.901706560959407e-05,
+ "loss": 0.145,
+ "step": 105
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8991132051180332e-05,
+ "loss": 0.1693,
+ "step": 106
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8964878977506496e-05,
+ "loss": 0.2012,
+ "step": 107
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8938307321520453e-05,
+ "loss": 0.1286,
+ "step": 108
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8911418027491453e-05,
+ "loss": 0.1396,
+ "step": 109
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8884212050976568e-05,
+ "loss": 0.0291,
+ "step": 110
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.885669035878672e-05,
+ "loss": 0.0317,
+ "step": 111
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.882885392895232e-05,
+ "loss": 0.1143,
+ "step": 112
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8800703750688536e-05,
+ "loss": 0.126,
+ "step": 113
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.877224082436011e-05,
+ "loss": 0.2017,
+ "step": 114
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8743466161445823e-05,
+ "loss": 0.0735,
+ "step": 115
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8714380784502553e-05,
+ "loss": 0.0527,
+ "step": 116
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8684985727128936e-05,
+ "loss": 0.1112,
+ "step": 117
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8655282033928618e-05,
+ "loss": 0.3129,
+ "step": 118
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8625270760473164e-05,
+ "loss": 0.2827,
+ "step": 119
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8594952973264512e-05,
+ "loss": 0.5608,
+ "step": 120
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.856432974969711e-05,
+ "loss": 0.1465,
+ "step": 121
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8533402178019596e-05,
+ "loss": 0.1322,
+ "step": 122
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8502171357296144e-05,
+ "loss": 0.0912,
+ "step": 123
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8470638397367397e-05,
+ "loss": 0.0419,
+ "step": 124
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8438804418811038e-05,
+ "loss": 0.0369,
+ "step": 125
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8406670552901958e-05,
+ "loss": 0.0529,
+ "step": 126
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.837423794157206e-05,
+ "loss": 0.1472,
+ "step": 127
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.834150773736967e-05,
+ "loss": 0.0425,
+ "step": 128
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8308481103418597e-05,
+ "loss": 0.1634,
+ "step": 129
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8275159213376783e-05,
+ "loss": 0.0485,
+ "step": 130
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.82415432513946e-05,
+ "loss": 0.0313,
+ "step": 131
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8207634412072765e-05,
+ "loss": 0.1792,
+ "step": 132
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.81734339004199e-05,
+ "loss": 0.1184,
+ "step": 133
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8138942931809702e-05,
+ "loss": 0.2756,
+ "step": 134
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8104162731937746e-05,
+ "loss": 0.0635,
+ "step": 135
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8069094536777938e-05,
+ "loss": 0.0158,
+ "step": 136
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8033739592538598e-05,
+ "loss": 0.2732,
+ "step": 137
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7998099155618147e-05,
+ "loss": 0.1428,
+ "step": 138
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7962174492560492e-05,
+ "loss": 0.0777,
+ "step": 139
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7925966880009998e-05,
+ "loss": 0.1644,
+ "step": 140
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7889477604666124e-05,
+ "loss": 0.0999,
+ "step": 141
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.785270796323769e-05,
+ "loss": 0.0446,
+ "step": 142
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7815659262396825e-05,
+ "loss": 0.0647,
+ "step": 143
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7778332818732492e-05,
+ "loss": 0.0521,
+ "step": 144
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7740729958703725e-05,
+ "loss": 0.2041,
+ "step": 145
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7702852018592493e-05,
+ "loss": 0.0149,
+ "step": 146
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7664700344456198e-05,
+ "loss": 0.0502,
+ "step": 147
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.762627629207986e-05,
+ "loss": 0.2027,
+ "step": 148
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.758758122692791e-05,
+ "loss": 0.0187,
+ "step": 149
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7548616524095697e-05,
+ "loss": 0.1248,
+ "step": 150
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7509383568260597e-05,
+ "loss": 0.0859,
+ "step": 151
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7469883753632817e-05,
+ "loss": 0.0822,
+ "step": 152
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.743011848390585e-05,
+ "loss": 0.2445,
+ "step": 153
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7390089172206594e-05,
+ "loss": 0.2662,
+ "step": 154
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7349797241045115e-05,
+ "loss": 0.0984,
+ "step": 155
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.730924412226413e-05,
+ "loss": 0.0317,
+ "step": 156
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.726843125698809e-05,
+ "loss": 0.1129,
+ "step": 157
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7227360095571992e-05,
+ "loss": 0.1882,
+ "step": 158
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7186032097549822e-05,
+ "loss": 0.1099,
+ "step": 159
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7144448731582698e-05,
+ "loss": 0.3506,
+ "step": 160
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7102611475406676e-05,
+ "loss": 0.0936,
+ "step": 161
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7060521815780225e-05,
+ "loss": 0.104,
+ "step": 162
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7018181248431416e-05,
+ "loss": 0.168,
+ "step": 163
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6975591278004747e-05,
+ "loss": 0.2726,
+ "step": 164
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6932753418007683e-05,
+ "loss": 0.0564,
+ "step": 165
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.688966919075687e-05,
+ "loss": 0.2981,
+ "step": 166
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.684634012732403e-05,
+ "loss": 0.0602,
+ "step": 167
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.680276776748157e-05,
+ "loss": 0.0364,
+ "step": 168
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6758953659647838e-05,
+ "loss": 0.096,
+ "step": 169
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6714899360832118e-05,
+ "loss": 0.2139,
+ "step": 170
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.667060643657929e-05,
+ "loss": 0.1666,
+ "step": 171
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.66260764609142e-05,
+ "loss": 0.0486,
+ "step": 172
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.658131101628571e-05,
+ "loss": 0.055,
+ "step": 173
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.653631169351049e-05,
+ "loss": 0.0953,
+ "step": 174
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6491080091716457e-05,
+ "loss": 0.1824,
+ "step": 175
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6445617818285974e-05,
+ "loss": 0.0226,
+ "step": 176
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6399926488798702e-05,
+ "loss": 0.0388,
+ "step": 177
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6354007726974205e-05,
+ "loss": 0.1149,
+ "step": 178
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.630786316461425e-05,
+ "loss": 0.1428,
+ "step": 179
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6261494441544805e-05,
+ "loss": 0.0445,
+ "step": 180
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6214903205557774e-05,
+ "loss": 0.0612,
+ "step": 181
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6168091112352443e-05,
+ "loss": 0.0826,
+ "step": 182
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.612105982547663e-05,
+ "loss": 0.0376,
+ "step": 183
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.607381101626758e-05,
+ "loss": 0.1441,
+ "step": 184
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6026346363792565e-05,
+ "loss": 0.1089,
+ "step": 185
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5978667554789216e-05,
+ "loss": 0.0845,
+ "step": 186
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5930776283605585e-05,
+ "loss": 0.0835,
+ "step": 187
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5882674252139928e-05,
+ "loss": 0.0762,
+ "step": 188
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5834363169780227e-05,
+ "loss": 0.067,
+ "step": 189
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.578584475334345e-05,
+ "loss": 0.0327,
+ "step": 190
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5737120727014535e-05,
+ "loss": 0.0254,
+ "step": 191
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5688192822285116e-05,
+ "loss": 0.028,
+ "step": 192
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5639062777892e-05,
+ "loss": 0.1708,
+ "step": 193
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5589732339755362e-05,
+ "loss": 0.0542,
+ "step": 194
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5540203260916728e-05,
+ "loss": 0.2358,
+ "step": 195
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5490477301476648e-05,
+ "loss": 0.1471,
+ "step": 196
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5440556228532168e-05,
+ "loss": 0.0414,
+ "step": 197
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5390441816114022e-05,
+ "loss": 0.0754,
+ "step": 198
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.534013584512359e-05,
+ "loss": 0.105,
+ "step": 199
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5289640103269626e-05,
+ "loss": 0.2052,
+ "step": 200
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5238956385004703e-05,
+ "loss": 0.2482,
+ "step": 201
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5188086491461467e-05,
+ "loss": 0.0967,
+ "step": 202
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5137032230388613e-05,
+ "loss": 0.1314,
+ "step": 203
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5085795416086655e-05,
+ "loss": 0.2313,
+ "step": 204
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5034377869343453e-05,
+ "loss": 0.1304,
+ "step": 205
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.4982781417369496e-05,
+ "loss": 0.2304,
+ "step": 206
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.4931007893732981e-05,
+ "loss": 0.0508,
+ "step": 207
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.4879059138294647e-05,
+ "loss": 0.1389,
+ "step": 208
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4826936997142399e-05,
+ "loss": 0.2129,
+ "step": 209
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4774643322525691e-05,
+ "loss": 0.0201,
+ "step": 210
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4722179972789725e-05,
+ "loss": 0.1064,
+ "step": 211
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.466954881230939e-05,
+ "loss": 0.0459,
+ "step": 212
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4616751711423016e-05,
+ "loss": 0.2229,
+ "step": 213
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4563790546365914e-05,
+ "loss": 0.1464,
+ "step": 214
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4510667199203697e-05,
+ "loss": 0.0558,
+ "step": 215
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4457383557765385e-05,
+ "loss": 0.0214,
+ "step": 216
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4403941515576344e-05,
+ "loss": 0.1551,
+ "step": 217
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4350342971790979e-05,
+ "loss": 0.2093,
+ "step": 218
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4296589831125234e-05,
+ "loss": 0.0453,
+ "step": 219
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4242684003788934e-05,
+ "loss": 0.0317,
+ "step": 220
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.418862740541788e-05,
+ "loss": 0.1334,
+ "step": 221
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4134421957005775e-05,
+ "loss": 0.0185,
+ "step": 222
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4080069584835971e-05,
+ "loss": 0.087,
+ "step": 223
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4025572220412998e-05,
+ "loss": 0.1747,
+ "step": 224
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3970931800393943e-05,
+ "loss": 0.1168,
+ "step": 225
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.391615026651961e-05,
+ "loss": 0.5095,
+ "step": 226
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3861229565545532e-05,
+ "loss": 0.1157,
+ "step": 227
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3806171649172782e-05,
+ "loss": 0.1201,
+ "step": 228
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3750978473978611e-05,
+ "loss": 0.2232,
+ "step": 229
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3695652001346928e-05,
+ "loss": 0.1718,
+ "step": 230
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.36401941973986e-05,
+ "loss": 0.0509,
+ "step": 231
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3584607032921566e-05,
+ "loss": 0.0333,
+ "step": 232
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3528892483300821e-05,
+ "loss": 0.1811,
+ "step": 233
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3473052528448203e-05,
+ "loss": 0.1771,
+ "step": 234
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3417089152732049e-05,
+ "loss": 0.1098,
+ "step": 235
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3361004344906652e-05,
+ "loss": 0.0566,
+ "step": 236
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.330480009804162e-05,
+ "loss": 0.2864,
+ "step": 237
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3248478409451017e-05,
+ "loss": 0.0166,
+ "step": 238
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3192041280622409e-05,
+ "loss": 0.2239,
+ "step": 239
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3135490717145726e-05,
+ "loss": 0.2247,
+ "step": 240
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3078828728641994e-05,
+ "loss": 0.1758,
+ "step": 241
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3022057328691915e-05,
+ "loss": 0.0618,
+ "step": 242
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2965178534764311e-05,
+ "loss": 0.1204,
+ "step": 243
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2908194368144437e-05,
+ "loss": 0.0233,
+ "step": 244
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.285110685386215e-05,
+ "loss": 0.0387,
+ "step": 245
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2793918020619937e-05,
+ "loss": 0.0791,
+ "step": 246
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2736629900720832e-05,
+ "loss": 0.0106,
+ "step": 247
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2679244529996182e-05,
+ "loss": 0.042,
+ "step": 248
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.262176394773332e-05,
+ "loss": 0.0725,
+ "step": 249
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.256419019660308e-05,
+ "loss": 0.0834,
+ "step": 250
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2506525322587207e-05,
+ "loss": 0.0432,
+ "step": 251
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2448771374905655e-05,
+ "loss": 0.177,
+ "step": 252
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2390930405943766e-05,
+ "loss": 0.0887,
+ "step": 253
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.233300447117933e-05,
+ "loss": 0.0152,
+ "step": 254
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2274995629109545e-05,
+ "loss": 0.0317,
+ "step": 255
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2216905941177854e-05,
+ "loss": 0.0268,
+ "step": 256
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.215873747170071e-05,
+ "loss": 0.1685,
+ "step": 257
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2100492287794186e-05,
+ "loss": 0.1403,
+ "step": 258
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2042172459300546e-05,
+ "loss": 0.0443,
+ "step": 259
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.198378005871467e-05,
+ "loss": 0.3589,
+ "step": 260
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.192531716111042e-05,
+ "loss": 0.0427,
+ "step": 261
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1866785844066884e-05,
+ "loss": 0.1103,
+ "step": 262
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1808188187594549e-05,
+ "loss": 0.2563,
+ "step": 263
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1749526274061394e-05,
+ "loss": 0.1494,
+ "step": 264
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1690802188118878e-05,
+ "loss": 0.1105,
+ "step": 265
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1632018016627859e-05,
+ "loss": 0.082,
+ "step": 266
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1573175848584455e-05,
+ "loss": 0.3555,
+ "step": 267
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1514277775045768e-05,
+ "loss": 0.0603,
+ "step": 268
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1455325889055616e-05,
+ "loss": 0.2883,
+ "step": 269
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1396322285570119e-05,
+ "loss": 0.054,
+ "step": 270
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1337269061383278e-05,
+ "loss": 0.0668,
+ "step": 271
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1278168315052445e-05,
+ "loss": 0.1454,
+ "step": 272
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1219022146823762e-05,
+ "loss": 0.0619,
+ "step": 273
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1159832658557498e-05,
+ "loss": 0.0449,
+ "step": 274
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1100601953653393e-05,
+ "loss": 0.0684,
+ "step": 275
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1041332136975874e-05,
+ "loss": 0.0273,
+ "step": 276
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.0982025314779287e-05,
+ "loss": 0.2375,
+ "step": 277
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.092268359463302e-05,
+ "loss": 0.0353,
+ "step": 278
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.086330908534663e-05,
+ "loss": 0.1224,
+ "step": 279
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.0803903896894877e-05,
+ "loss": 0.1297,
+ "step": 280
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0744470140342775e-05,
+ "loss": 0.4464,
+ "step": 281
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0685009927770542e-05,
+ "loss": 0.103,
+ "step": 282
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0625525372198564e-05,
+ "loss": 0.0881,
+ "step": 283
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.056601858751229e-05,
+ "loss": 0.075,
+ "step": 284
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0506491688387128e-05,
+ "loss": 0.0677,
+ "step": 285
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0446946790213275e-05,
+ "loss": 0.2301,
+ "step": 286
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0387386009020569e-05,
+ "loss": 0.0737,
+ "step": 287
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.032781146140326e-05,
+ "loss": 0.1262,
+ "step": 288
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0268225264444829e-05,
+ "loss": 0.0252,
+ "step": 289
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0208629535642726e-05,
+ "loss": 0.0192,
+ "step": 290
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0149026392833137e-05,
+ "loss": 0.257,
+ "step": 291
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0089417954115715e-05,
+ "loss": 0.1876,
+ "step": 292
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.002980633777831e-05,
+ "loss": 0.0341,
+ "step": 293
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.970193662221694e-06,
+ "loss": 0.232,
+ "step": 294
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.910582045884292e-06,
+ "loss": 0.1429,
+ "step": 295
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.850973607166865e-06,
+ "loss": 0.2432,
+ "step": 296
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.791370464357279e-06,
+ "loss": 0.0288,
+ "step": 297
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.731774735555174e-06,
+ "loss": 0.2272,
+ "step": 298
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.672188538596746e-06,
+ "loss": 0.1102,
+ "step": 299
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.612613990979436e-06,
+ "loss": 0.0529,
+ "step": 300
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.553053209786725e-06,
+ "loss": 0.1721,
+ "step": 301
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.493508311612874e-06,
+ "loss": 0.0046,
+ "step": 302
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.433981412487711e-06,
+ "loss": 0.043,
+ "step": 303
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.374474627801439e-06,
+ "loss": 0.0589,
+ "step": 304
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.314990072229461e-06,
+ "loss": 0.0114,
+ "step": 305
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.25552985965723e-06,
+ "loss": 0.1645,
+ "step": 306
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.196096103105127e-06,
+ "loss": 0.2002,
+ "step": 307
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.136690914653377e-06,
+ "loss": 0.057,
+ "step": 308
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.07731640536698e-06,
+ "loss": 0.1744,
+ "step": 309
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.017974685220716e-06,
+ "loss": 0.0343,
+ "step": 310
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.958667863024127e-06,
+ "loss": 0.0405,
+ "step": 311
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.899398046346608e-06,
+ "loss": 0.2055,
+ "step": 312
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.840167341442505e-06,
+ "loss": 0.0673,
+ "step": 313
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.78097785317624e-06,
+ "loss": 0.0291,
+ "step": 314
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.721831684947557e-06,
+ "loss": 0.2443,
+ "step": 315
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.662730938616724e-06,
+ "loss": 0.058,
+ "step": 316
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.603677714429888e-06,
+ "loss": 0.2347,
+ "step": 317
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.54467411094439e-06,
+ "loss": 0.0307,
+ "step": 318
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.485722224954237e-06,
+ "loss": 0.0094,
+ "step": 319
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.426824151415548e-06,
+ "loss": 0.0724,
+ "step": 320
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.367981983372143e-06,
+ "loss": 0.0816,
+ "step": 321
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.309197811881128e-06,
+ "loss": 0.0375,
+ "step": 322
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.250473725938608e-06,
+ "loss": 0.0106,
+ "step": 323
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.191811812405453e-06,
+ "loss": 0.0701,
+ "step": 324
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.133214155933118e-06,
+ "loss": 0.0134,
+ "step": 325
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.074682838889581e-06,
+ "loss": 0.1992,
+ "step": 326
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.01621994128533e-06,
+ "loss": 0.1688,
+ "step": 327
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.95782754069946e-06,
+ "loss": 0.2751,
+ "step": 328
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.899507712205818e-06,
+ "loss": 0.0192,
+ "step": 329
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.841262528299296e-06,
+ "loss": 0.0797,
+ "step": 330
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.783094058822147e-06,
+ "loss": 0.0867,
+ "step": 331
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.72500437089046e-06,
+ "loss": 0.0445,
+ "step": 332
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.666995528820673e-06,
+ "loss": 0.1654,
+ "step": 333
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.609069594056234e-06,
+ "loss": 0.0168,
+ "step": 334
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.551228625094349e-06,
+ "loss": 0.0779,
+ "step": 335
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.493474677412795e-06,
+ "loss": 0.0444,
+ "step": 336
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.435809803396923e-06,
+ "loss": 0.1839,
+ "step": 337
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.37823605226668e-06,
+ "loss": 0.3834,
+ "step": 338
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.320755470003822e-06,
+ "loss": 0.0261,
+ "step": 339
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.263370099279173e-06,
+ "loss": 0.0084,
+ "step": 340
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.2060819793800665e-06,
+ "loss": 0.0469,
+ "step": 341
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.148893146137852e-06,
+ "loss": 0.3605,
+ "step": 342
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.091805631855566e-06,
+ "loss": 0.0621,
+ "step": 343
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.034821465235693e-06,
+ "loss": 0.099,
+ "step": 344
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 6.977942671308087e-06,
+ "loss": 0.0641,
+ "step": 345
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 6.921171271358007e-06,
+ "loss": 0.0859,
+ "step": 346
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.864509282854272e-06,
+ "loss": 0.0564,
+ "step": 347
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.8079587193775935e-06,
+ "loss": 0.0405,
+ "step": 348
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.751521590548986e-06,
+ "loss": 0.101,
+ "step": 349
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.695199901958386e-06,
+ "loss": 0.1178,
+ "step": 350
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.638995655093351e-06,
+ "loss": 0.2406,
+ "step": 351
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.582910847267957e-06,
+ "loss": 0.1846,
+ "step": 352
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.526947471551799e-06,
+ "loss": 0.1374,
+ "step": 353
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.471107516699183e-06,
+ "loss": 0.0863,
+ "step": 354
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.415392967078438e-06,
+ "loss": 0.0755,
+ "step": 355
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.3598058026013995e-06,
+ "loss": 0.0732,
+ "step": 356
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.304347998653074e-06,
+ "loss": 0.0555,
+ "step": 357
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.24902152602139e-06,
+ "loss": 0.0475,
+ "step": 358
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.193828350827222e-06,
+ "loss": 0.036,
+ "step": 359
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.1387704344544684e-06,
+ "loss": 0.2679,
+ "step": 360
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.083849733480394e-06,
+ "loss": 0.0661,
+ "step": 361
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.0290681996060605e-06,
+ "loss": 0.0362,
+ "step": 362
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 5.974427779587004e-06,
+ "loss": 0.0815,
+ "step": 363
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.919930415164033e-06,
+ "loss": 0.0205,
+ "step": 364
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.865578042994227e-06,
+ "loss": 0.0065,
+ "step": 365
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.8113725945821245e-06,
+ "loss": 0.2377,
+ "step": 366
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.757315996211066e-06,
+ "loss": 0.0673,
+ "step": 367
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.703410168874768e-06,
+ "loss": 0.1033,
+ "step": 368
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.649657028209024e-06,
+ "loss": 0.1259,
+ "step": 369
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.5960584844236565e-06,
+ "loss": 0.0052,
+ "step": 370
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.542616442234618e-06,
+ "loss": 0.1048,
+ "step": 371
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.48933280079631e-06,
+ "loss": 0.3342,
+ "step": 372
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.436209453634087e-06,
+ "loss": 0.0725,
+ "step": 373
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.3832482885769855e-06,
+ "loss": 0.1597,
+ "step": 374
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.330451187690614e-06,
+ "loss": 0.2186,
+ "step": 375
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.277820027210279e-06,
+ "loss": 0.0521,
+ "step": 376
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.225356677474309e-06,
+ "loss": 0.0426,
+ "step": 377
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.1730630028576055e-06,
+ "loss": 0.1171,
+ "step": 378
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.120940861705357e-06,
+ "loss": 0.0551,
+ "step": 379
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.068992106267021e-06,
+ "loss": 0.1238,
+ "step": 380
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.017218582630507e-06,
+ "loss": 0.4425,
+ "step": 381
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.965622130656551e-06,
+ "loss": 0.1591,
+ "step": 382
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.914204583913349e-06,
+ "loss": 0.0568,
+ "step": 383
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.862967769611389e-06,
+ "loss": 0.0159,
+ "step": 384
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.8119135085385375e-06,
+ "loss": 0.055,
+ "step": 385
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.7610436149953e-06,
+ "loss": 0.0356,
+ "step": 386
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.710359896730379e-06,
+ "loss": 0.0969,
+ "step": 387
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.659864154876411e-06,
+ "loss": 0.1161,
+ "step": 388
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.609558183885979e-06,
+ "loss": 0.0437,
+ "step": 389
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.559443771467833e-06,
+ "loss": 0.1526,
+ "step": 390
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.509522698523352e-06,
+ "loss": 0.0183,
+ "step": 391
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.4597967390832745e-06,
+ "loss": 0.073,
+ "step": 392
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.4102676602446375e-06,
+ "loss": 0.0411,
+ "step": 393
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.360937222108002e-06,
+ "loss": 0.0524,
+ "step": 394
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.3118071777148865e-06,
+ "loss": 0.1156,
+ "step": 395
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.262879272985468e-06,
+ "loss": 0.0311,
+ "step": 396
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.21415524665655e-06,
+ "loss": 0.1253,
+ "step": 397
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.165636830219776e-06,
+ "loss": 0.0589,
+ "step": 398
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.117325747860077e-06,
+ "loss": 0.0248,
+ "step": 399
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.069223716394419e-06,
+ "loss": 0.0164,
+ "step": 400
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.021332445210785e-06,
+ "loss": 0.1801,
+ "step": 401
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 3.973653636207437e-06,
+ "loss": 0.107,
+ "step": 402
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.9261889837324245e-06,
+ "loss": 0.0477,
+ "step": 403
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.878940174523371e-06,
+ "loss": 0.0214,
+ "step": 404
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.8319088876475595e-06,
+ "loss": 0.1071,
+ "step": 405
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.785096794442229e-06,
+ "loss": 0.071,
+ "step": 406
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.7385055584552e-06,
+ "loss": 0.0623,
+ "step": 407
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6921368353857524e-06,
+ "loss": 0.0534,
+ "step": 408
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.645992273025797e-06,
+ "loss": 0.1143,
+ "step": 409
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6000735112012984e-06,
+ "loss": 0.1056,
+ "step": 410
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.5543821817140313e-06,
+ "loss": 0.0537,
+ "step": 411
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.5089199082835436e-06,
+ "loss": 0.0065,
+ "step": 412
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.463688306489511e-06,
+ "loss": 0.0995,
+ "step": 413
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.418688983714291e-06,
+ "loss": 0.0818,
+ "step": 414
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.373923539085805e-06,
+ "loss": 0.0481,
+ "step": 415
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.329393563420713e-06,
+ "loss": 0.1379,
+ "step": 416
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.285100639167883e-06,
+ "loss": 0.1759,
+ "step": 417
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.2410463403521653e-06,
+ "loss": 0.0599,
+ "step": 418
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.1972322325184347e-06,
+ "loss": 0.0898,
+ "step": 419
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.1536598726759747e-06,
+ "loss": 0.0079,
+ "step": 420
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.110330809243134e-06,
+ "loss": 0.0185,
+ "step": 421
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0672465819923215e-06,
+ "loss": 0.0792,
+ "step": 422
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0244087219952565e-06,
+ "loss": 0.1059,
+ "step": 423
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 2.981818751568586e-06,
+ "loss": 0.044,
+ "step": 424
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.939478184219777e-06,
+ "loss": 0.0766,
+ "step": 425
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.8973885245933287e-06,
+ "loss": 0.1558,
+ "step": 426
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.855551268417305e-06,
+ "loss": 0.0052,
+ "step": 427
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.813967902450179e-06,
+ "loss": 0.0747,
+ "step": 428
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.7726399044280107e-06,
+ "loss": 0.0868,
+ "step": 429
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.7315687430119097e-06,
+ "loss": 0.047,
+ "step": 430
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.6907558777358756e-06,
+ "loss": 0.0721,
+ "step": 431
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.650202758954886e-06,
+ "loss": 0.128,
+ "step": 432
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.6099108277934105e-06,
+ "loss": 0.08,
+ "step": 433
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.5698815160941494e-06,
+ "loss": 0.0901,
+ "step": 434
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.5301162463671845e-06,
+ "loss": 0.0965,
+ "step": 435
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4906164317394067e-06,
+ "loss": 0.062,
+ "step": 436
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.451383475904304e-06,
+ "loss": 0.0634,
+ "step": 437
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4124187730720916e-06,
+ "loss": 0.1525,
+ "step": 438
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.3737237079201437e-06,
+ "loss": 0.1071,
+ "step": 439
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.3352996555438036e-06,
+ "loss": 0.0409,
+ "step": 440
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.297147981407509e-06,
+ "loss": 0.1753,
+ "step": 441
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.2592700412962775e-06,
+ "loss": 0.175,
+ "step": 442
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.2216671812675118e-06,
+ "loss": 0.0348,
+ "step": 443
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.184340737603178e-06,
+ "loss": 0.105,
+ "step": 444
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.1472920367623094e-06,
+ "loss": 0.0477,
+ "step": 445
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.1105223953338805e-06,
+ "loss": 0.0176,
+ "step": 446
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0740331199900053e-06,
+ "loss": 0.6195,
+ "step": 447
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0378255074395094e-06,
+ "loss": 0.0913,
+ "step": 448
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.001900844381857e-06,
+ "loss": 0.0386,
+ "step": 449
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.9662604074614044e-06,
+ "loss": 0.1309,
+ "step": 450
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.9309054632220645e-06,
+ "loss": 0.0218,
+ "step": 451
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.895837268062256e-06,
+ "loss": 0.0185,
+ "step": 452
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8610570681903018e-06,
+ "loss": 0.3416,
+ "step": 453
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8265660995801004e-06,
+ "loss": 0.2817,
+ "step": 454
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7923655879272395e-06,
+ "loss": 0.0182,
+ "step": 455
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7584567486054039e-06,
+ "loss": 0.0665,
+ "step": 456
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7248407866232175e-06,
+ "loss": 0.0403,
+ "step": 457
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6915188965814034e-06,
+ "loss": 0.017,
+ "step": 458
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6915188965814034e-06,
+ "loss": 0.3175,
+ "step": 459
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6584922626303325e-06,
+ "loss": 0.0474,
+ "step": 460
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6257620584279454e-06,
+ "loss": 0.0881,
+ "step": 461
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.5933294470980443e-06,
+ "loss": 0.0475,
+ "step": 462
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.5611955811889645e-06,
+ "loss": 0.0473,
+ "step": 463
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.5293616026326053e-06,
+ "loss": 0.0143,
+ "step": 464
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4978286427038602e-06,
+ "loss": 0.1228,
+ "step": 465
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4665978219804056e-06,
+ "loss": 0.2635,
+ "step": 466
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.435670250302892e-06,
+ "loss": 0.0668,
+ "step": 467
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.405047026735491e-06,
+ "loss": 0.082,
+ "step": 468
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3747292395268407e-06,
+ "loss": 0.085,
+ "step": 469
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.344717966071385e-06,
+ "loss": 0.1178,
+ "step": 470
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3150142728710669e-06,
+ "loss": 0.0633,
+ "step": 471
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.2856192154974488e-06,
+ "loss": 0.0229,
+ "step": 472
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.2565338385541792e-06,
+ "loss": 0.0356,
+ "step": 473
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.2277591756398933e-06,
+ "loss": 0.1599,
+ "step": 474
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1992962493114645e-06,
+ "loss": 0.0168,
+ "step": 475
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.171146071047683e-06,
+ "loss": 0.0626,
+ "step": 476
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1433096412132838e-06,
+ "loss": 0.1343,
+ "step": 477
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1157879490234346e-06,
+ "loss": 0.0529,
+ "step": 478
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.088581972508549e-06,
+ "loss": 0.0556,
+ "step": 479
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.0616926784795511e-06,
+ "loss": 0.0903,
+ "step": 480
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.035121022493506e-06,
+ "loss": 0.0993,
+ "step": 481
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.0088679488196695e-06,
+ "loss": 0.0673,
+ "step": 482
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.829343904059342e-07,
+ "loss": 0.018,
+ "step": 483
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.573212688456635e-07,
+ "loss": 0.1005,
+ "step": 484
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.320294943449537e-07,
+ "loss": 0.0859,
+ "step": 485
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 9.070599656902801e-07,
+ "loss": 0.0361,
+ "step": 486
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.824135702165693e-07,
+ "loss": 0.0256,
+ "step": 487
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.580911837756467e-07,
+ "loss": 0.0652,
+ "step": 488
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.340936707051273e-07,
+ "loss": 0.103,
+ "step": 489
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.10421883797694e-07,
+ "loss": 0.0589,
+ "step": 490
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 7.87076664270795e-07,
+ "loss": 0.1919,
+ "step": 491
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.6405884173675e-07,
+ "loss": 0.1313,
+ "step": 492
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.413692341732582e-07,
+ "loss": 0.0657,
+ "step": 493
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.190086478943459e-07,
+ "loss": 0.1785,
+ "step": 494
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 6.969778775217007e-07,
+ "loss": 0.1866,
+ "step": 495
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 6.752777059564431e-07,
+ "loss": 0.0295,
+ "step": 496
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.539089043512914e-07,
+ "loss": 0.0316,
+ "step": 497
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.328722320831737e-07,
+ "loss": 0.0702,
+ "step": 498
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.121684367262271e-07,
+ "loss": 0.0271,
+ "step": 499
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 5.917982540252442e-07,
+ "loss": 0.0398,
+ "step": 500
+ }
+ ],
+ "logging_steps": 1.0,
+ "max_steps": 555,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 1,
+ "save_steps": 500,
+ "total_flos": 1080325140480.0,
+ "train_batch_size": 10,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/training_args.bin b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..086695c74164e41c129494184be863504cd808c1
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19850de0da1b4eac2d894ea196e517fb53755d5462a1803598ffe2ca47c1ad86
+size 6968
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/zero_to_fp32.py b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..c98caae31534368be22b67fc4ae906836c992a8d
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/zero_to_fp32.py
@@ -0,0 +1,587 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+ FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+ FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+ buffers: dict()
+ param_shapes: dict()
+ shared_params: list
+ ds_version: int
+ frozen_param_shapes: dict()
+ frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+ return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+ '''
+ alist.sort(key=natural_keys) sorts in human order
+ http://nedbatchelder.com/blog/200712/human_sorting.html
+ (See Toothy's implementation in the comments)
+ '''
+ return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+ if not os.path.isdir(checkpoint_dir):
+ raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+ # there should be only one file
+ if zero_stage <= 2:
+ file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+ elif zero_stage == 3:
+ file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+ if not os.path.exists(file):
+ raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+ return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+ # XXX: need to test that this simple glob rule works for multi-node setup too
+ ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+ if len(ckpt_files) == 0:
+ raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+ return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+ return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+ return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+ zero_model_states = []
+ for file in files:
+ state_dict = torch.load(file, map_location=device)
+
+ if BUFFER_NAMES not in state_dict:
+ raise ValueError(f"{file} is not a model state checkpoint")
+ buffer_names = state_dict[BUFFER_NAMES]
+ if debug:
+ print("Found buffers:", buffer_names)
+
+ # recover just the buffers while restoring them to fp32 if they were saved in fp16
+ buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+ param_shapes = state_dict[PARAM_SHAPES]
+
+ # collect parameters that are included in param_shapes
+ param_names = []
+ for s in param_shapes:
+ for name in s.keys():
+ param_names.append(name)
+
+ # update with frozen parameters
+ frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+ if frozen_param_shapes is not None:
+ if debug:
+ print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+ param_names += list(frozen_param_shapes.keys())
+
+ # handle shared params
+ shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+ ds_version = state_dict.get(DS_VERSION, None)
+
+ frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+ z_model_state = zero_model_state(buffers=buffers,
+ param_shapes=param_shapes,
+ shared_params=shared_params,
+ ds_version=ds_version,
+ frozen_param_shapes=frozen_param_shapes,
+ frozen_param_fragments=frozen_param_fragments)
+ zero_model_states.append(z_model_state)
+
+ return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+ total_files = len(files)
+ state_dicts = []
+ for f in files:
+ state_dict = torch.load(f, map_location=device)
+ # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+ # and also handle the case where it was already removed by another helper script
+ state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+ state_dicts.append(state_dict)
+
+ if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+ raise ValueError(f"{files[0]} is not a zero checkpoint")
+ zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+ world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+ # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+ # parameters can be different from data parallelism for non-expert parameters. So we can just
+ # use the max of the partition_count to get the dp world_size.
+
+ if type(world_size) is list:
+ world_size = max(world_size)
+
+ if world_size != total_files:
+ raise ValueError(
+ f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+ "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+ )
+
+ # the groups are named differently in each stage
+ if zero_stage <= 2:
+ fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+ elif zero_stage == 3:
+ fp32_groups_key = FP32_FLAT_GROUPS
+ else:
+ raise ValueError(f"unknown zero stage {zero_stage}")
+
+ if zero_stage <= 2:
+ fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+ elif zero_stage == 3:
+ # if there is more than one param group, there will be multiple flattened tensors - one
+ # flattened tensor per group - for simplicity merge them into a single tensor
+ #
+ # XXX: could make the script more memory efficient for when there are multiple groups - it
+ # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+ fp32_flat_groups = [
+ torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+ ]
+
+ return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir):
+ """
+ Returns fp32 state_dict reconstructed from ds checkpoint
+
+ Args:
+ - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+ """
+ print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+ optim_files = get_optim_files(ds_checkpoint_dir)
+ zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+ print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+ model_files = get_model_state_files(ds_checkpoint_dir)
+
+ zero_model_states = parse_model_states(model_files)
+ print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+ if zero_stage <= 2:
+ return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+ elif zero_stage == 3:
+ return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+ if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+ return
+
+ frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+ frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+ if debug:
+ num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+ print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+ wanted_params = len(frozen_param_shapes)
+ wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+ avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+ print(f'Frozen params: Have {avail_numel} numels to process.')
+ print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+ total_params = 0
+ total_numel = 0
+ for name, shape in frozen_param_shapes.items():
+ total_params += 1
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+
+ state_dict[name] = frozen_param_fragments[name]
+
+ if debug:
+ print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+ print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+ param_shapes = zero_model_states[0].param_shapes
+
+ # Reconstruction protocol:
+ #
+ # XXX: document this
+
+ if debug:
+ for i in range(world_size):
+ for j in range(len(fp32_flat_groups[0])):
+ print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+ # XXX: memory usage doubles here (zero2)
+ num_param_groups = len(fp32_flat_groups[0])
+ merged_single_partition_of_fp32_groups = []
+ for i in range(num_param_groups):
+ merged_partitions = [sd[i] for sd in fp32_flat_groups]
+ full_single_fp32_vector = torch.cat(merged_partitions, 0)
+ merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+ avail_numel = sum(
+ [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+ if debug:
+ wanted_params = sum([len(shapes) for shapes in param_shapes])
+ wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+ # not asserting if there is a mismatch due to possible padding
+ print(f"Have {avail_numel} numels to process.")
+ print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+ # params
+ # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+ # out-of-core computing solution
+ total_numel = 0
+ total_params = 0
+ for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+ offset = 0
+ avail_numel = full_single_fp32_vector.numel()
+ for name, shape in shapes.items():
+
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+ total_params += 1
+
+ if debug:
+ print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+ state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+ offset += unpartitioned_numel
+
+ # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+ # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+ # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+ # live optimizer object, so we are checking that the numbers are within the right range
+ align_to = 2 * world_size
+
+ def zero2_align(x):
+ return align_to * math.ceil(x / align_to)
+
+ if debug:
+ print(f"original offset={offset}, avail_numel={avail_numel}")
+
+ offset = zero2_align(offset)
+ avail_numel = zero2_align(avail_numel)
+
+ if debug:
+ print(f"aligned offset={offset}, avail_numel={avail_numel}")
+
+ # Sanity check
+ if offset != avail_numel:
+ raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+ print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+ state_dict = OrderedDict()
+
+ # buffers
+ buffers = zero_model_states[0].buffers
+ state_dict.update(buffers)
+ if debug:
+ print(f"added {len(buffers)} buffers")
+
+ _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+ _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+ # recover shared parameters
+ for pair in zero_model_states[0].shared_params:
+ if pair[1] in state_dict:
+ state_dict[pair[0]] = state_dict[pair[1]]
+
+ return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+ remainder = unpartitioned_numel % world_size
+ padding_numel = (world_size - remainder) if remainder else 0
+ partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+ return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+ if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+ return
+
+ if debug:
+ for i in range(world_size):
+ num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+ print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+ frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+ wanted_params = len(frozen_param_shapes)
+ wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+ avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+ print(f'Frozen params: Have {avail_numel} numels to process.')
+ print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+ total_params = 0
+ total_numel = 0
+ for name, shape in zero_model_states[0].frozen_param_shapes.items():
+ total_params += 1
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+
+ param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+ state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+ partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+ if debug:
+ print(
+ f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+ )
+
+ print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+ param_shapes = zero_model_states[0].param_shapes
+ avail_numel = fp32_flat_groups[0].numel() * world_size
+ # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+ # param, re-consolidating each param, while dealing with padding if any
+
+ # merge list of dicts, preserving order
+ param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+ if debug:
+ for i in range(world_size):
+ print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+ wanted_params = len(param_shapes)
+ wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+ # not asserting if there is a mismatch due to possible padding
+ avail_numel = fp32_flat_groups[0].numel() * world_size
+ print(f"Trainable params: Have {avail_numel} numels to process.")
+ print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+ # params
+ # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+ # out-of-core computing solution
+ offset = 0
+ total_numel = 0
+ total_params = 0
+ for name, shape in param_shapes.items():
+
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+ total_params += 1
+
+ partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+ if debug:
+ print(
+ f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+ )
+
+ # XXX: memory usage doubles here
+ state_dict[name] = torch.cat(
+ tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+ 0).narrow(0, 0, unpartitioned_numel).view(shape)
+ offset += partitioned_numel
+
+ offset *= world_size
+
+ # Sanity check
+ if offset != avail_numel:
+ raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+ print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+ state_dict = OrderedDict()
+
+ # buffers
+ buffers = zero_model_states[0].buffers
+ state_dict.update(buffers)
+ if debug:
+ print(f"added {len(buffers)} buffers")
+
+ _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+ _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+ # recover shared parameters
+ for pair in zero_model_states[0].shared_params:
+ if pair[1] in state_dict:
+ state_dict[pair[0]] = state_dict[pair[1]]
+
+ return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None):
+ """
+ Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+ ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+ via a model hub.
+
+ Args:
+ - ``checkpoint_dir``: path to the desired checkpoint folder
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+
+ Returns:
+ - pytorch ``state_dict``
+
+ Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+ you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+ the checkpoint.
+
+ A typical usage might be ::
+
+ from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+ # do the training and checkpoint saving
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+ model = model.cpu() # move to cpu
+ model.load_state_dict(state_dict)
+ # submit to model hub or save the model to share with others
+
+ In this example the ``model`` will no longer be usable in the deepspeed context of the same
+ application. i.e. you will need to re-initialize the deepspeed engine, since
+ ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+ If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+ """
+ if tag is None:
+ latest_path = os.path.join(checkpoint_dir, 'latest')
+ if os.path.isfile(latest_path):
+ with open(latest_path, 'r') as fd:
+ tag = fd.read().strip()
+ else:
+ raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+ ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+ if not os.path.isdir(ds_checkpoint_dir):
+ raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+ return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None):
+ """
+ Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+ loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+ Args:
+ - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+ - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+ """
+
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+ print(f"Saving fp32 state dict to {output_file}")
+ torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+ """
+ 1. Put the provided model to cpu
+ 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+ 3. Load it into the provided model
+
+ Args:
+ - ``model``: the model object to update
+ - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+ Returns:
+ - ``model`: modified model
+
+ Make sure you have plenty of CPU memory available before you call this function. If you don't
+ have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+ conveniently placed for you in the checkpoint folder.
+
+ A typical usage might be ::
+
+ from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+ model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+ # submit to model hub or save the model to share with others
+
+ Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+ of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+ ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+ """
+ logger.info(f"Extracting fp32 weights")
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+ logger.info(f"Overwriting model with fp32 weights")
+ model = model.cpu()
+ model.load_state_dict(state_dict, strict=False)
+
+ return model
+
+
+if __name__ == "__main__":
+
+ parser = argparse.ArgumentParser()
+ parser.add_argument("checkpoint_dir",
+ type=str,
+ help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+ parser.add_argument(
+ "output_file",
+ type=str,
+ help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+ parser.add_argument("-t",
+ "--tag",
+ type=str,
+ default=None,
+ help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+ parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+ args = parser.parse_args()
+
+ debug = args.debug
+
+ convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, args.output_file, tag=args.tag)
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/config.json b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..93e133af45036a778791b5679a8953a4f6a35a33
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/config.json
@@ -0,0 +1,70 @@
+{
+ "_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "architectures": [
+ "LlavaMistralForCausalLM"
+ ],
+ "attention_dropout": 0.0,
+ "bos_token_id": 1,
+ "eos_token_id": 2,
+ "freeze_mm_mlp_adapter": false,
+ "freeze_mm_vision_resampler": false,
+ "hidden_act": "silu",
+ "hidden_size": 4096,
+ "image_aspect_ratio": "anyres",
+ "image_crop_resolution": 224,
+ "image_grid_pinpoints": [
+ [
+ 336,
+ 672
+ ],
+ [
+ 672,
+ 336
+ ],
+ [
+ 672,
+ 672
+ ],
+ [
+ 1008,
+ 336
+ ],
+ [
+ 336,
+ 1008
+ ]
+ ],
+ "image_split_resolution": 224,
+ "initializer_range": 0.02,
+ "intermediate_size": 14336,
+ "max_position_embeddings": 32768,
+ "mm_hidden_size": 1024,
+ "mm_patch_merge_type": "spatial_unpad",
+ "mm_projector_lr": 2e-05,
+ "mm_projector_type": "mlp2x_gelu",
+ "mm_resampler_type": null,
+ "mm_use_im_patch_token": false,
+ "mm_use_im_start_end": false,
+ "mm_vision_select_feature": "patch",
+ "mm_vision_select_layer": -2,
+ "mm_vision_tower": "openai/clip-vit-large-patch14-336",
+ "mm_vision_tower_lr": 2e-06,
+ "model_type": "llava_mistral",
+ "num_attention_heads": 32,
+ "num_hidden_layers": 32,
+ "num_key_value_heads": 8,
+ "rms_norm_eps": 1e-05,
+ "rope_theta": 1000000.0,
+ "sliding_window": null,
+ "tie_word_embeddings": false,
+ "tokenizer_model_max_length": 4096,
+ "tokenizer_padding_side": "right",
+ "torch_dtype": "bfloat16",
+ "transformers_version": "4.37.2",
+ "tune_mm_mlp_adapter": false,
+ "tune_mm_vision_resampler": false,
+ "unfreeze_mm_vision_tower": true,
+ "use_cache": true,
+ "use_mm_proj": true,
+ "vocab_size": 32000
+}
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin
new file mode 100644
index 0000000000000000000000000000000000000000..e035ed370a46949146eb7164d79ef88097eacbfb
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1fde24a5cdabb49bef91a9dd1ee36c3b2ed72791efc564f6476124852a334852
+size 41961648
diff --git a/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/trainer_state.json b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..09272574ba783ea775f320dc35885ea3bdebbd9e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank/llava-lora-mistral-r128a256-10BS-model/trainer_state.json
@@ -0,0 +1,3360 @@
+{
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.0,
+ "eval_steps": 500,
+ "global_step": 555,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "epoch": 0.0,
+ "learning_rate": 7.142857142857143e-07,
+ "loss": 0.4237,
+ "step": 1
+ },
+ {
+ "epoch": 0.0,
+ "learning_rate": 1.4285714285714286e-06,
+ "loss": 0.3368,
+ "step": 2
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.1428571428571427e-06,
+ "loss": 0.214,
+ "step": 3
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.8571428571428573e-06,
+ "loss": 0.396,
+ "step": 4
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 3.5714285714285718e-06,
+ "loss": 0.305,
+ "step": 5
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 4.2857142857142855e-06,
+ "loss": 0.4049,
+ "step": 6
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 5e-06,
+ "loss": 0.108,
+ "step": 7
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 5.7142857142857145e-06,
+ "loss": 0.2286,
+ "step": 8
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 6.4285714285714295e-06,
+ "loss": 0.1443,
+ "step": 9
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.1428571428571436e-06,
+ "loss": 0.2252,
+ "step": 10
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.857142857142858e-06,
+ "loss": 0.0747,
+ "step": 11
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 8.571428571428571e-06,
+ "loss": 0.1084,
+ "step": 12
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 9.285714285714288e-06,
+ "loss": 0.2115,
+ "step": 13
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1e-05,
+ "loss": 0.4742,
+ "step": 14
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.0714285714285714e-05,
+ "loss": 0.083,
+ "step": 15
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.1428571428571429e-05,
+ "loss": 0.3392,
+ "step": 16
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.2142857142857142e-05,
+ "loss": 0.065,
+ "step": 17
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.2857142857142859e-05,
+ "loss": 0.1711,
+ "step": 18
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.3571428571428574e-05,
+ "loss": 0.0539,
+ "step": 19
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.4285714285714287e-05,
+ "loss": 0.0701,
+ "step": 20
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.5000000000000002e-05,
+ "loss": 0.0836,
+ "step": 21
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.5714285714285715e-05,
+ "loss": 0.1891,
+ "step": 22
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.642857142857143e-05,
+ "loss": 0.0422,
+ "step": 23
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.7142857142857142e-05,
+ "loss": 0.2094,
+ "step": 24
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.785714285714286e-05,
+ "loss": 0.139,
+ "step": 25
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.8571428571428575e-05,
+ "loss": 0.2214,
+ "step": 26
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.928571428571429e-05,
+ "loss": 0.1084,
+ "step": 27
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 2e-05,
+ "loss": 0.0898,
+ "step": 28
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.9999822316445652e-05,
+ "loss": 0.0359,
+ "step": 29
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.9999289272096886e-05,
+ "loss": 0.2648,
+ "step": 30
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9998400885896355e-05,
+ "loss": 0.4007,
+ "step": 31
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9997157189414373e-05,
+ "loss": 0.235,
+ "step": 32
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.999555822684783e-05,
+ "loss": 0.0273,
+ "step": 33
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.999360405501859e-05,
+ "loss": 0.0267,
+ "step": 34
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.99912947433715e-05,
+ "loss": 0.2619,
+ "step": 35
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9988630373971896e-05,
+ "loss": 0.4101,
+ "step": 36
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9985611041502704e-05,
+ "loss": 0.1302,
+ "step": 37
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9982236853261067e-05,
+ "loss": 0.118,
+ "step": 38
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9978507929154534e-05,
+ "loss": 0.0933,
+ "step": 39
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.997442440169681e-05,
+ "loss": 0.0104,
+ "step": 40
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9969986416003026e-05,
+ "loss": 0.1061,
+ "step": 41
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9965194129784597e-05,
+ "loss": 0.1575,
+ "step": 42
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.996004771334361e-05,
+ "loss": 0.1969,
+ "step": 43
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.996004771334361e-05,
+ "loss": 0.0492,
+ "step": 44
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9954547349566783e-05,
+ "loss": 0.3012,
+ "step": 45
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.994869323391895e-05,
+ "loss": 0.2185,
+ "step": 46
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.994248557443613e-05,
+ "loss": 0.1729,
+ "step": 47
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.993592459171812e-05,
+ "loss": 0.0354,
+ "step": 48
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9929010518920667e-05,
+ "loss": 0.3939,
+ "step": 49
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.992174360174717e-05,
+ "loss": 0.0505,
+ "step": 50
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9914124098439976e-05,
+ "loss": 0.0777,
+ "step": 51
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9914124098439976e-05,
+ "loss": 0.6651,
+ "step": 52
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9906152279771162e-05,
+ "loss": 0.15,
+ "step": 53
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9897828429032946e-05,
+ "loss": 0.1416,
+ "step": 54
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9889152842027607e-05,
+ "loss": 0.1195,
+ "step": 55
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9880125827056967e-05,
+ "loss": 0.0787,
+ "step": 56
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.987074770491145e-05,
+ "loss": 0.0681,
+ "step": 57
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.986101880885867e-05,
+ "loss": 0.1337,
+ "step": 58
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9850939484631598e-05,
+ "loss": 0.0961,
+ "step": 59
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.984051009041626e-05,
+ "loss": 0.116,
+ "step": 60
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.982973099683902e-05,
+ "loss": 0.3853,
+ "step": 61
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9818602586953414e-05,
+ "loss": 0.0875,
+ "step": 62
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9807125256226532e-05,
+ "loss": 0.3216,
+ "step": 63
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9795299412524948e-05,
+ "loss": 0.0752,
+ "step": 64
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9783125476100254e-05,
+ "loss": 0.1461,
+ "step": 65
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9770603879574108e-05,
+ "loss": 0.075,
+ "step": 66
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.975773506792287e-05,
+ "loss": 0.0685,
+ "step": 67
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.974451949846177e-05,
+ "loss": 0.0555,
+ "step": 68
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.973095764082869e-05,
+ "loss": 0.0171,
+ "step": 69
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9717049976967437e-05,
+ "loss": 0.0247,
+ "step": 70
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9702797001110642e-05,
+ "loss": 0.0839,
+ "step": 71
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9688199219762183e-05,
+ "loss": 0.4163,
+ "step": 72
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.96732571516792e-05,
+ "loss": 0.1461,
+ "step": 73
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9657971327853644e-05,
+ "loss": 0.1457,
+ "step": 74
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.964234229149342e-05,
+ "loss": 0.0482,
+ "step": 75
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.962637059800307e-05,
+ "loss": 0.0802,
+ "step": 76
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9610056814964053e-05,
+ "loss": 0.0697,
+ "step": 77
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.959340152211455e-05,
+ "loss": 0.0614,
+ "step": 78
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.95764053113289e-05,
+ "loss": 0.1004,
+ "step": 79
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9559068786596526e-05,
+ "loss": 0.0286,
+ "step": 80
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.954139256400049e-05,
+ "loss": 0.1162,
+ "step": 81
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.952337727169561e-05,
+ "loss": 0.0731,
+ "step": 82
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.950502354988612e-05,
+ "loss": 0.0286,
+ "step": 83
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.948633205080292e-05,
+ "loss": 0.2425,
+ "step": 84
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9467303438680414e-05,
+ "loss": 0.0505,
+ "step": 85
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.944793838973289e-05,
+ "loss": 0.0922,
+ "step": 86
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9428237592130487e-05,
+ "loss": 0.2949,
+ "step": 87
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.940820174597476e-05,
+ "loss": 0.2807,
+ "step": 88
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9387831563273775e-05,
+ "loss": 0.2377,
+ "step": 89
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9367127767916828e-05,
+ "loss": 0.2558,
+ "step": 90
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9346091095648712e-05,
+ "loss": 0.0871,
+ "step": 91
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.932472229404356e-05,
+ "loss": 0.2204,
+ "step": 92
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9303022122478303e-05,
+ "loss": 0.1174,
+ "step": 93
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9280991352105656e-05,
+ "loss": 0.2181,
+ "step": 94
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.925863076582674e-05,
+ "loss": 0.1251,
+ "step": 95
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9235941158263253e-05,
+ "loss": 0.2251,
+ "step": 96
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9212923335729206e-05,
+ "loss": 0.1236,
+ "step": 97
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.918957811620231e-05,
+ "loss": 0.0901,
+ "step": 98
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9165906329294875e-05,
+ "loss": 0.1002,
+ "step": 99
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9141908816224356e-05,
+ "loss": 0.4397,
+ "step": 100
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9117586429783433e-05,
+ "loss": 0.1141,
+ "step": 101
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.909294003430972e-05,
+ "loss": 0.1842,
+ "step": 102
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.906797050565505e-05,
+ "loss": 0.0985,
+ "step": 103
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.9042678731154337e-05,
+ "loss": 0.1533,
+ "step": 104
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.901706560959407e-05,
+ "loss": 0.145,
+ "step": 105
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8991132051180332e-05,
+ "loss": 0.1693,
+ "step": 106
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8964878977506496e-05,
+ "loss": 0.2012,
+ "step": 107
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8938307321520453e-05,
+ "loss": 0.1286,
+ "step": 108
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8911418027491453e-05,
+ "loss": 0.1396,
+ "step": 109
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8884212050976568e-05,
+ "loss": 0.0291,
+ "step": 110
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.885669035878672e-05,
+ "loss": 0.0317,
+ "step": 111
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.882885392895232e-05,
+ "loss": 0.1143,
+ "step": 112
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8800703750688536e-05,
+ "loss": 0.126,
+ "step": 113
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.877224082436011e-05,
+ "loss": 0.2017,
+ "step": 114
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8743466161445823e-05,
+ "loss": 0.0735,
+ "step": 115
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8714380784502553e-05,
+ "loss": 0.0527,
+ "step": 116
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8684985727128936e-05,
+ "loss": 0.1112,
+ "step": 117
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8655282033928618e-05,
+ "loss": 0.3129,
+ "step": 118
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8625270760473164e-05,
+ "loss": 0.2827,
+ "step": 119
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8594952973264512e-05,
+ "loss": 0.5608,
+ "step": 120
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.856432974969711e-05,
+ "loss": 0.1465,
+ "step": 121
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8533402178019596e-05,
+ "loss": 0.1322,
+ "step": 122
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8502171357296144e-05,
+ "loss": 0.0912,
+ "step": 123
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8470638397367397e-05,
+ "loss": 0.0419,
+ "step": 124
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8438804418811038e-05,
+ "loss": 0.0369,
+ "step": 125
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8406670552901958e-05,
+ "loss": 0.0529,
+ "step": 126
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.837423794157206e-05,
+ "loss": 0.1472,
+ "step": 127
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.834150773736967e-05,
+ "loss": 0.0425,
+ "step": 128
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8308481103418597e-05,
+ "loss": 0.1634,
+ "step": 129
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8275159213376783e-05,
+ "loss": 0.0485,
+ "step": 130
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.82415432513946e-05,
+ "loss": 0.0313,
+ "step": 131
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8207634412072765e-05,
+ "loss": 0.1792,
+ "step": 132
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.81734339004199e-05,
+ "loss": 0.1184,
+ "step": 133
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8138942931809702e-05,
+ "loss": 0.2756,
+ "step": 134
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8104162731937746e-05,
+ "loss": 0.0635,
+ "step": 135
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8069094536777938e-05,
+ "loss": 0.0158,
+ "step": 136
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8033739592538598e-05,
+ "loss": 0.2732,
+ "step": 137
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7998099155618147e-05,
+ "loss": 0.1428,
+ "step": 138
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7962174492560492e-05,
+ "loss": 0.0777,
+ "step": 139
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7925966880009998e-05,
+ "loss": 0.1644,
+ "step": 140
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7889477604666124e-05,
+ "loss": 0.0999,
+ "step": 141
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.785270796323769e-05,
+ "loss": 0.0446,
+ "step": 142
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7815659262396825e-05,
+ "loss": 0.0647,
+ "step": 143
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7778332818732492e-05,
+ "loss": 0.0521,
+ "step": 144
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7740729958703725e-05,
+ "loss": 0.2041,
+ "step": 145
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7702852018592493e-05,
+ "loss": 0.0149,
+ "step": 146
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7664700344456198e-05,
+ "loss": 0.0502,
+ "step": 147
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.762627629207986e-05,
+ "loss": 0.2027,
+ "step": 148
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.758758122692791e-05,
+ "loss": 0.0187,
+ "step": 149
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7548616524095697e-05,
+ "loss": 0.1248,
+ "step": 150
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7509383568260597e-05,
+ "loss": 0.0859,
+ "step": 151
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7469883753632817e-05,
+ "loss": 0.0822,
+ "step": 152
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.743011848390585e-05,
+ "loss": 0.2445,
+ "step": 153
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7390089172206594e-05,
+ "loss": 0.2662,
+ "step": 154
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7349797241045115e-05,
+ "loss": 0.0984,
+ "step": 155
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.730924412226413e-05,
+ "loss": 0.0317,
+ "step": 156
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.726843125698809e-05,
+ "loss": 0.1129,
+ "step": 157
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7227360095571992e-05,
+ "loss": 0.1882,
+ "step": 158
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7186032097549822e-05,
+ "loss": 0.1099,
+ "step": 159
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7144448731582698e-05,
+ "loss": 0.3506,
+ "step": 160
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7102611475406676e-05,
+ "loss": 0.0936,
+ "step": 161
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7060521815780225e-05,
+ "loss": 0.104,
+ "step": 162
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7018181248431416e-05,
+ "loss": 0.168,
+ "step": 163
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6975591278004747e-05,
+ "loss": 0.2726,
+ "step": 164
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6932753418007683e-05,
+ "loss": 0.0564,
+ "step": 165
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.688966919075687e-05,
+ "loss": 0.2981,
+ "step": 166
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.684634012732403e-05,
+ "loss": 0.0602,
+ "step": 167
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.680276776748157e-05,
+ "loss": 0.0364,
+ "step": 168
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6758953659647838e-05,
+ "loss": 0.096,
+ "step": 169
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6714899360832118e-05,
+ "loss": 0.2139,
+ "step": 170
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.667060643657929e-05,
+ "loss": 0.1666,
+ "step": 171
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.66260764609142e-05,
+ "loss": 0.0486,
+ "step": 172
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.658131101628571e-05,
+ "loss": 0.055,
+ "step": 173
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.653631169351049e-05,
+ "loss": 0.0953,
+ "step": 174
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6491080091716457e-05,
+ "loss": 0.1824,
+ "step": 175
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6445617818285974e-05,
+ "loss": 0.0226,
+ "step": 176
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6399926488798702e-05,
+ "loss": 0.0388,
+ "step": 177
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6354007726974205e-05,
+ "loss": 0.1149,
+ "step": 178
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.630786316461425e-05,
+ "loss": 0.1428,
+ "step": 179
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6261494441544805e-05,
+ "loss": 0.0445,
+ "step": 180
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6214903205557774e-05,
+ "loss": 0.0612,
+ "step": 181
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6168091112352443e-05,
+ "loss": 0.0826,
+ "step": 182
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.612105982547663e-05,
+ "loss": 0.0376,
+ "step": 183
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.607381101626758e-05,
+ "loss": 0.1441,
+ "step": 184
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6026346363792565e-05,
+ "loss": 0.1089,
+ "step": 185
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5978667554789216e-05,
+ "loss": 0.0845,
+ "step": 186
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5930776283605585e-05,
+ "loss": 0.0835,
+ "step": 187
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5882674252139928e-05,
+ "loss": 0.0762,
+ "step": 188
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5834363169780227e-05,
+ "loss": 0.067,
+ "step": 189
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.578584475334345e-05,
+ "loss": 0.0327,
+ "step": 190
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5737120727014535e-05,
+ "loss": 0.0254,
+ "step": 191
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5688192822285116e-05,
+ "loss": 0.028,
+ "step": 192
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5639062777892e-05,
+ "loss": 0.1708,
+ "step": 193
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5589732339755362e-05,
+ "loss": 0.0542,
+ "step": 194
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5540203260916728e-05,
+ "loss": 0.2358,
+ "step": 195
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5490477301476648e-05,
+ "loss": 0.1471,
+ "step": 196
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5440556228532168e-05,
+ "loss": 0.0414,
+ "step": 197
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5390441816114022e-05,
+ "loss": 0.0754,
+ "step": 198
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.534013584512359e-05,
+ "loss": 0.105,
+ "step": 199
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5289640103269626e-05,
+ "loss": 0.2052,
+ "step": 200
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5238956385004703e-05,
+ "loss": 0.2482,
+ "step": 201
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5188086491461467e-05,
+ "loss": 0.0967,
+ "step": 202
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5137032230388613e-05,
+ "loss": 0.1314,
+ "step": 203
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5085795416086655e-05,
+ "loss": 0.2313,
+ "step": 204
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5034377869343453e-05,
+ "loss": 0.1304,
+ "step": 205
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.4982781417369496e-05,
+ "loss": 0.2304,
+ "step": 206
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.4931007893732981e-05,
+ "loss": 0.0508,
+ "step": 207
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.4879059138294647e-05,
+ "loss": 0.1389,
+ "step": 208
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4826936997142399e-05,
+ "loss": 0.2129,
+ "step": 209
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4774643322525691e-05,
+ "loss": 0.0201,
+ "step": 210
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4722179972789725e-05,
+ "loss": 0.1064,
+ "step": 211
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.466954881230939e-05,
+ "loss": 0.0459,
+ "step": 212
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4616751711423016e-05,
+ "loss": 0.2229,
+ "step": 213
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4563790546365914e-05,
+ "loss": 0.1464,
+ "step": 214
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4510667199203697e-05,
+ "loss": 0.0558,
+ "step": 215
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4457383557765385e-05,
+ "loss": 0.0214,
+ "step": 216
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4403941515576344e-05,
+ "loss": 0.1551,
+ "step": 217
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4350342971790979e-05,
+ "loss": 0.2093,
+ "step": 218
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4296589831125234e-05,
+ "loss": 0.0453,
+ "step": 219
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4242684003788934e-05,
+ "loss": 0.0317,
+ "step": 220
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.418862740541788e-05,
+ "loss": 0.1334,
+ "step": 221
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4134421957005775e-05,
+ "loss": 0.0185,
+ "step": 222
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4080069584835971e-05,
+ "loss": 0.087,
+ "step": 223
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4025572220412998e-05,
+ "loss": 0.1747,
+ "step": 224
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3970931800393943e-05,
+ "loss": 0.1168,
+ "step": 225
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.391615026651961e-05,
+ "loss": 0.5095,
+ "step": 226
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3861229565545532e-05,
+ "loss": 0.1157,
+ "step": 227
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3806171649172782e-05,
+ "loss": 0.1201,
+ "step": 228
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3750978473978611e-05,
+ "loss": 0.2232,
+ "step": 229
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3695652001346928e-05,
+ "loss": 0.1718,
+ "step": 230
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.36401941973986e-05,
+ "loss": 0.0509,
+ "step": 231
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3584607032921566e-05,
+ "loss": 0.0333,
+ "step": 232
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3528892483300821e-05,
+ "loss": 0.1811,
+ "step": 233
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3473052528448203e-05,
+ "loss": 0.1771,
+ "step": 234
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3417089152732049e-05,
+ "loss": 0.1098,
+ "step": 235
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3361004344906652e-05,
+ "loss": 0.0566,
+ "step": 236
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.330480009804162e-05,
+ "loss": 0.2864,
+ "step": 237
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3248478409451017e-05,
+ "loss": 0.0166,
+ "step": 238
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3192041280622409e-05,
+ "loss": 0.2239,
+ "step": 239
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3135490717145726e-05,
+ "loss": 0.2247,
+ "step": 240
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3078828728641994e-05,
+ "loss": 0.1758,
+ "step": 241
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3022057328691915e-05,
+ "loss": 0.0618,
+ "step": 242
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2965178534764311e-05,
+ "loss": 0.1204,
+ "step": 243
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2908194368144437e-05,
+ "loss": 0.0233,
+ "step": 244
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.285110685386215e-05,
+ "loss": 0.0387,
+ "step": 245
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2793918020619937e-05,
+ "loss": 0.0791,
+ "step": 246
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2736629900720832e-05,
+ "loss": 0.0106,
+ "step": 247
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2679244529996182e-05,
+ "loss": 0.042,
+ "step": 248
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.262176394773332e-05,
+ "loss": 0.0725,
+ "step": 249
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.256419019660308e-05,
+ "loss": 0.0834,
+ "step": 250
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2506525322587207e-05,
+ "loss": 0.0432,
+ "step": 251
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2448771374905655e-05,
+ "loss": 0.177,
+ "step": 252
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2390930405943766e-05,
+ "loss": 0.0887,
+ "step": 253
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.233300447117933e-05,
+ "loss": 0.0152,
+ "step": 254
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2274995629109545e-05,
+ "loss": 0.0317,
+ "step": 255
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2216905941177854e-05,
+ "loss": 0.0268,
+ "step": 256
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.215873747170071e-05,
+ "loss": 0.1685,
+ "step": 257
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2100492287794186e-05,
+ "loss": 0.1403,
+ "step": 258
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2042172459300546e-05,
+ "loss": 0.0443,
+ "step": 259
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.198378005871467e-05,
+ "loss": 0.3589,
+ "step": 260
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.192531716111042e-05,
+ "loss": 0.0427,
+ "step": 261
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1866785844066884e-05,
+ "loss": 0.1103,
+ "step": 262
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1808188187594549e-05,
+ "loss": 0.2563,
+ "step": 263
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1749526274061394e-05,
+ "loss": 0.1494,
+ "step": 264
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1690802188118878e-05,
+ "loss": 0.1105,
+ "step": 265
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1632018016627859e-05,
+ "loss": 0.082,
+ "step": 266
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1573175848584455e-05,
+ "loss": 0.3555,
+ "step": 267
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1514277775045768e-05,
+ "loss": 0.0603,
+ "step": 268
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1455325889055616e-05,
+ "loss": 0.2883,
+ "step": 269
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1396322285570119e-05,
+ "loss": 0.054,
+ "step": 270
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1337269061383278e-05,
+ "loss": 0.0668,
+ "step": 271
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1278168315052445e-05,
+ "loss": 0.1454,
+ "step": 272
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1219022146823762e-05,
+ "loss": 0.0619,
+ "step": 273
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1159832658557498e-05,
+ "loss": 0.0449,
+ "step": 274
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1100601953653393e-05,
+ "loss": 0.0684,
+ "step": 275
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1041332136975874e-05,
+ "loss": 0.0273,
+ "step": 276
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.0982025314779287e-05,
+ "loss": 0.2375,
+ "step": 277
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.092268359463302e-05,
+ "loss": 0.0353,
+ "step": 278
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.086330908534663e-05,
+ "loss": 0.1224,
+ "step": 279
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.0803903896894877e-05,
+ "loss": 0.1297,
+ "step": 280
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0744470140342775e-05,
+ "loss": 0.4464,
+ "step": 281
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0685009927770542e-05,
+ "loss": 0.103,
+ "step": 282
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0625525372198564e-05,
+ "loss": 0.0881,
+ "step": 283
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.056601858751229e-05,
+ "loss": 0.075,
+ "step": 284
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0506491688387128e-05,
+ "loss": 0.0677,
+ "step": 285
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0446946790213275e-05,
+ "loss": 0.2301,
+ "step": 286
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0387386009020569e-05,
+ "loss": 0.0737,
+ "step": 287
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.032781146140326e-05,
+ "loss": 0.1262,
+ "step": 288
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0268225264444829e-05,
+ "loss": 0.0252,
+ "step": 289
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0208629535642726e-05,
+ "loss": 0.0192,
+ "step": 290
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0149026392833137e-05,
+ "loss": 0.257,
+ "step": 291
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0089417954115715e-05,
+ "loss": 0.1876,
+ "step": 292
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.002980633777831e-05,
+ "loss": 0.0341,
+ "step": 293
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.970193662221694e-06,
+ "loss": 0.232,
+ "step": 294
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.910582045884292e-06,
+ "loss": 0.1429,
+ "step": 295
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.850973607166865e-06,
+ "loss": 0.2432,
+ "step": 296
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.791370464357279e-06,
+ "loss": 0.0288,
+ "step": 297
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.731774735555174e-06,
+ "loss": 0.2272,
+ "step": 298
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.672188538596746e-06,
+ "loss": 0.1102,
+ "step": 299
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.612613990979436e-06,
+ "loss": 0.0529,
+ "step": 300
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.553053209786725e-06,
+ "loss": 0.1721,
+ "step": 301
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.493508311612874e-06,
+ "loss": 0.0046,
+ "step": 302
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.433981412487711e-06,
+ "loss": 0.043,
+ "step": 303
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.374474627801439e-06,
+ "loss": 0.0589,
+ "step": 304
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.314990072229461e-06,
+ "loss": 0.0114,
+ "step": 305
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.25552985965723e-06,
+ "loss": 0.1645,
+ "step": 306
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.196096103105127e-06,
+ "loss": 0.2002,
+ "step": 307
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.136690914653377e-06,
+ "loss": 0.057,
+ "step": 308
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.07731640536698e-06,
+ "loss": 0.1744,
+ "step": 309
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.017974685220716e-06,
+ "loss": 0.0343,
+ "step": 310
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.958667863024127e-06,
+ "loss": 0.0405,
+ "step": 311
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.899398046346608e-06,
+ "loss": 0.2055,
+ "step": 312
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.840167341442505e-06,
+ "loss": 0.0673,
+ "step": 313
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.78097785317624e-06,
+ "loss": 0.0291,
+ "step": 314
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.721831684947557e-06,
+ "loss": 0.2443,
+ "step": 315
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.662730938616724e-06,
+ "loss": 0.058,
+ "step": 316
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.603677714429888e-06,
+ "loss": 0.2347,
+ "step": 317
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.54467411094439e-06,
+ "loss": 0.0307,
+ "step": 318
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.485722224954237e-06,
+ "loss": 0.0094,
+ "step": 319
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.426824151415548e-06,
+ "loss": 0.0724,
+ "step": 320
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.367981983372143e-06,
+ "loss": 0.0816,
+ "step": 321
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.309197811881128e-06,
+ "loss": 0.0375,
+ "step": 322
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.250473725938608e-06,
+ "loss": 0.0106,
+ "step": 323
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.191811812405453e-06,
+ "loss": 0.0701,
+ "step": 324
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.133214155933118e-06,
+ "loss": 0.0134,
+ "step": 325
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.074682838889581e-06,
+ "loss": 0.1992,
+ "step": 326
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.01621994128533e-06,
+ "loss": 0.1688,
+ "step": 327
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.95782754069946e-06,
+ "loss": 0.2751,
+ "step": 328
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.899507712205818e-06,
+ "loss": 0.0192,
+ "step": 329
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.841262528299296e-06,
+ "loss": 0.0797,
+ "step": 330
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.783094058822147e-06,
+ "loss": 0.0867,
+ "step": 331
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.72500437089046e-06,
+ "loss": 0.0445,
+ "step": 332
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.666995528820673e-06,
+ "loss": 0.1654,
+ "step": 333
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.609069594056234e-06,
+ "loss": 0.0168,
+ "step": 334
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.551228625094349e-06,
+ "loss": 0.0779,
+ "step": 335
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.493474677412795e-06,
+ "loss": 0.0444,
+ "step": 336
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.435809803396923e-06,
+ "loss": 0.1839,
+ "step": 337
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.37823605226668e-06,
+ "loss": 0.3834,
+ "step": 338
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.320755470003822e-06,
+ "loss": 0.0261,
+ "step": 339
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.263370099279173e-06,
+ "loss": 0.0084,
+ "step": 340
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.2060819793800665e-06,
+ "loss": 0.0469,
+ "step": 341
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.148893146137852e-06,
+ "loss": 0.3605,
+ "step": 342
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.091805631855566e-06,
+ "loss": 0.0621,
+ "step": 343
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.034821465235693e-06,
+ "loss": 0.099,
+ "step": 344
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 6.977942671308087e-06,
+ "loss": 0.0641,
+ "step": 345
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 6.921171271358007e-06,
+ "loss": 0.0859,
+ "step": 346
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.864509282854272e-06,
+ "loss": 0.0564,
+ "step": 347
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.8079587193775935e-06,
+ "loss": 0.0405,
+ "step": 348
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.751521590548986e-06,
+ "loss": 0.101,
+ "step": 349
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.695199901958386e-06,
+ "loss": 0.1178,
+ "step": 350
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.638995655093351e-06,
+ "loss": 0.2406,
+ "step": 351
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.582910847267957e-06,
+ "loss": 0.1846,
+ "step": 352
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.526947471551799e-06,
+ "loss": 0.1374,
+ "step": 353
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.471107516699183e-06,
+ "loss": 0.0863,
+ "step": 354
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.415392967078438e-06,
+ "loss": 0.0755,
+ "step": 355
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.3598058026013995e-06,
+ "loss": 0.0732,
+ "step": 356
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.304347998653074e-06,
+ "loss": 0.0555,
+ "step": 357
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.24902152602139e-06,
+ "loss": 0.0475,
+ "step": 358
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.193828350827222e-06,
+ "loss": 0.036,
+ "step": 359
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.1387704344544684e-06,
+ "loss": 0.2679,
+ "step": 360
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.083849733480394e-06,
+ "loss": 0.0661,
+ "step": 361
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.0290681996060605e-06,
+ "loss": 0.0362,
+ "step": 362
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 5.974427779587004e-06,
+ "loss": 0.0815,
+ "step": 363
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.919930415164033e-06,
+ "loss": 0.0205,
+ "step": 364
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.865578042994227e-06,
+ "loss": 0.0065,
+ "step": 365
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.8113725945821245e-06,
+ "loss": 0.2377,
+ "step": 366
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.757315996211066e-06,
+ "loss": 0.0673,
+ "step": 367
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.703410168874768e-06,
+ "loss": 0.1033,
+ "step": 368
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.649657028209024e-06,
+ "loss": 0.1259,
+ "step": 369
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.5960584844236565e-06,
+ "loss": 0.0052,
+ "step": 370
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.542616442234618e-06,
+ "loss": 0.1048,
+ "step": 371
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.48933280079631e-06,
+ "loss": 0.3342,
+ "step": 372
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.436209453634087e-06,
+ "loss": 0.0725,
+ "step": 373
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.3832482885769855e-06,
+ "loss": 0.1597,
+ "step": 374
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.330451187690614e-06,
+ "loss": 0.2186,
+ "step": 375
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.277820027210279e-06,
+ "loss": 0.0521,
+ "step": 376
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.225356677474309e-06,
+ "loss": 0.0426,
+ "step": 377
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.1730630028576055e-06,
+ "loss": 0.1171,
+ "step": 378
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.120940861705357e-06,
+ "loss": 0.0551,
+ "step": 379
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.068992106267021e-06,
+ "loss": 0.1238,
+ "step": 380
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.017218582630507e-06,
+ "loss": 0.4425,
+ "step": 381
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.965622130656551e-06,
+ "loss": 0.1591,
+ "step": 382
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.914204583913349e-06,
+ "loss": 0.0568,
+ "step": 383
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.862967769611389e-06,
+ "loss": 0.0159,
+ "step": 384
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.8119135085385375e-06,
+ "loss": 0.055,
+ "step": 385
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.7610436149953e-06,
+ "loss": 0.0356,
+ "step": 386
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.710359896730379e-06,
+ "loss": 0.0969,
+ "step": 387
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.659864154876411e-06,
+ "loss": 0.1161,
+ "step": 388
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.609558183885979e-06,
+ "loss": 0.0437,
+ "step": 389
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.559443771467833e-06,
+ "loss": 0.1526,
+ "step": 390
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.509522698523352e-06,
+ "loss": 0.0183,
+ "step": 391
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.4597967390832745e-06,
+ "loss": 0.073,
+ "step": 392
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.4102676602446375e-06,
+ "loss": 0.0411,
+ "step": 393
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.360937222108002e-06,
+ "loss": 0.0524,
+ "step": 394
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.3118071777148865e-06,
+ "loss": 0.1156,
+ "step": 395
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.262879272985468e-06,
+ "loss": 0.0311,
+ "step": 396
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.21415524665655e-06,
+ "loss": 0.1253,
+ "step": 397
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.165636830219776e-06,
+ "loss": 0.0589,
+ "step": 398
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.117325747860077e-06,
+ "loss": 0.0248,
+ "step": 399
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.069223716394419e-06,
+ "loss": 0.0164,
+ "step": 400
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.021332445210785e-06,
+ "loss": 0.1801,
+ "step": 401
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 3.973653636207437e-06,
+ "loss": 0.107,
+ "step": 402
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.9261889837324245e-06,
+ "loss": 0.0477,
+ "step": 403
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.878940174523371e-06,
+ "loss": 0.0214,
+ "step": 404
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.8319088876475595e-06,
+ "loss": 0.1071,
+ "step": 405
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.785096794442229e-06,
+ "loss": 0.071,
+ "step": 406
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.7385055584552e-06,
+ "loss": 0.0623,
+ "step": 407
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6921368353857524e-06,
+ "loss": 0.0534,
+ "step": 408
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.645992273025797e-06,
+ "loss": 0.1143,
+ "step": 409
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6000735112012984e-06,
+ "loss": 0.1056,
+ "step": 410
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.5543821817140313e-06,
+ "loss": 0.0537,
+ "step": 411
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.5089199082835436e-06,
+ "loss": 0.0065,
+ "step": 412
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.463688306489511e-06,
+ "loss": 0.0995,
+ "step": 413
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.418688983714291e-06,
+ "loss": 0.0818,
+ "step": 414
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.373923539085805e-06,
+ "loss": 0.0481,
+ "step": 415
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.329393563420713e-06,
+ "loss": 0.1379,
+ "step": 416
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.285100639167883e-06,
+ "loss": 0.1759,
+ "step": 417
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.2410463403521653e-06,
+ "loss": 0.0599,
+ "step": 418
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.1972322325184347e-06,
+ "loss": 0.0898,
+ "step": 419
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.1536598726759747e-06,
+ "loss": 0.0079,
+ "step": 420
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.110330809243134e-06,
+ "loss": 0.0185,
+ "step": 421
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0672465819923215e-06,
+ "loss": 0.0792,
+ "step": 422
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0244087219952565e-06,
+ "loss": 0.1059,
+ "step": 423
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 2.981818751568586e-06,
+ "loss": 0.044,
+ "step": 424
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.939478184219777e-06,
+ "loss": 0.0766,
+ "step": 425
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.8973885245933287e-06,
+ "loss": 0.1558,
+ "step": 426
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.855551268417305e-06,
+ "loss": 0.0052,
+ "step": 427
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.813967902450179e-06,
+ "loss": 0.0747,
+ "step": 428
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.7726399044280107e-06,
+ "loss": 0.0868,
+ "step": 429
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.7315687430119097e-06,
+ "loss": 0.047,
+ "step": 430
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.6907558777358756e-06,
+ "loss": 0.0721,
+ "step": 431
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.650202758954886e-06,
+ "loss": 0.128,
+ "step": 432
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.6099108277934105e-06,
+ "loss": 0.08,
+ "step": 433
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.5698815160941494e-06,
+ "loss": 0.0901,
+ "step": 434
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.5301162463671845e-06,
+ "loss": 0.0965,
+ "step": 435
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4906164317394067e-06,
+ "loss": 0.062,
+ "step": 436
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.451383475904304e-06,
+ "loss": 0.0634,
+ "step": 437
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4124187730720916e-06,
+ "loss": 0.1525,
+ "step": 438
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.3737237079201437e-06,
+ "loss": 0.1071,
+ "step": 439
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.3352996555438036e-06,
+ "loss": 0.0409,
+ "step": 440
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.297147981407509e-06,
+ "loss": 0.1753,
+ "step": 441
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.2592700412962775e-06,
+ "loss": 0.175,
+ "step": 442
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.2216671812675118e-06,
+ "loss": 0.0348,
+ "step": 443
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.184340737603178e-06,
+ "loss": 0.105,
+ "step": 444
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.1472920367623094e-06,
+ "loss": 0.0477,
+ "step": 445
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.1105223953338805e-06,
+ "loss": 0.0176,
+ "step": 446
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0740331199900053e-06,
+ "loss": 0.6195,
+ "step": 447
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0378255074395094e-06,
+ "loss": 0.0913,
+ "step": 448
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.001900844381857e-06,
+ "loss": 0.0386,
+ "step": 449
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.9662604074614044e-06,
+ "loss": 0.1309,
+ "step": 450
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.9309054632220645e-06,
+ "loss": 0.0218,
+ "step": 451
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.895837268062256e-06,
+ "loss": 0.0185,
+ "step": 452
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8610570681903018e-06,
+ "loss": 0.3416,
+ "step": 453
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8265660995801004e-06,
+ "loss": 0.2817,
+ "step": 454
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7923655879272395e-06,
+ "loss": 0.0182,
+ "step": 455
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7584567486054039e-06,
+ "loss": 0.0665,
+ "step": 456
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7248407866232175e-06,
+ "loss": 0.0403,
+ "step": 457
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6915188965814034e-06,
+ "loss": 0.017,
+ "step": 458
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6915188965814034e-06,
+ "loss": 0.3175,
+ "step": 459
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6584922626303325e-06,
+ "loss": 0.0474,
+ "step": 460
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6257620584279454e-06,
+ "loss": 0.0881,
+ "step": 461
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.5933294470980443e-06,
+ "loss": 0.0475,
+ "step": 462
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.5611955811889645e-06,
+ "loss": 0.0473,
+ "step": 463
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.5293616026326053e-06,
+ "loss": 0.0143,
+ "step": 464
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4978286427038602e-06,
+ "loss": 0.1228,
+ "step": 465
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4665978219804056e-06,
+ "loss": 0.2635,
+ "step": 466
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.435670250302892e-06,
+ "loss": 0.0668,
+ "step": 467
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.405047026735491e-06,
+ "loss": 0.082,
+ "step": 468
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3747292395268407e-06,
+ "loss": 0.085,
+ "step": 469
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.344717966071385e-06,
+ "loss": 0.1178,
+ "step": 470
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3150142728710669e-06,
+ "loss": 0.0633,
+ "step": 471
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.2856192154974488e-06,
+ "loss": 0.0229,
+ "step": 472
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.2565338385541792e-06,
+ "loss": 0.0356,
+ "step": 473
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.2277591756398933e-06,
+ "loss": 0.1599,
+ "step": 474
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1992962493114645e-06,
+ "loss": 0.0168,
+ "step": 475
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.171146071047683e-06,
+ "loss": 0.0626,
+ "step": 476
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1433096412132838e-06,
+ "loss": 0.1343,
+ "step": 477
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1157879490234346e-06,
+ "loss": 0.0529,
+ "step": 478
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.088581972508549e-06,
+ "loss": 0.0556,
+ "step": 479
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.0616926784795511e-06,
+ "loss": 0.0903,
+ "step": 480
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.035121022493506e-06,
+ "loss": 0.0993,
+ "step": 481
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.0088679488196695e-06,
+ "loss": 0.0673,
+ "step": 482
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.829343904059342e-07,
+ "loss": 0.018,
+ "step": 483
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.573212688456635e-07,
+ "loss": 0.1005,
+ "step": 484
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.320294943449537e-07,
+ "loss": 0.0859,
+ "step": 485
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 9.070599656902801e-07,
+ "loss": 0.0361,
+ "step": 486
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.824135702165693e-07,
+ "loss": 0.0256,
+ "step": 487
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.580911837756467e-07,
+ "loss": 0.0652,
+ "step": 488
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.340936707051273e-07,
+ "loss": 0.103,
+ "step": 489
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.10421883797694e-07,
+ "loss": 0.0589,
+ "step": 490
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 7.87076664270795e-07,
+ "loss": 0.1919,
+ "step": 491
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.6405884173675e-07,
+ "loss": 0.1313,
+ "step": 492
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.413692341732582e-07,
+ "loss": 0.0657,
+ "step": 493
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.190086478943459e-07,
+ "loss": 0.1785,
+ "step": 494
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 6.969778775217007e-07,
+ "loss": 0.1866,
+ "step": 495
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 6.752777059564431e-07,
+ "loss": 0.0295,
+ "step": 496
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.539089043512914e-07,
+ "loss": 0.0316,
+ "step": 497
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.328722320831737e-07,
+ "loss": 0.0702,
+ "step": 498
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.121684367262271e-07,
+ "loss": 0.0271,
+ "step": 499
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 5.917982540252442e-07,
+ "loss": 0.0398,
+ "step": 500
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 5.71762407869515e-07,
+ "loss": 0.2131,
+ "step": 501
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 5.520616102671128e-07,
+ "loss": 0.217,
+ "step": 502
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.326965613195867e-07,
+ "loss": 0.0161,
+ "step": 503
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.136679491970809e-07,
+ "loss": 0.0338,
+ "step": 504
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 4.949764501138832e-07,
+ "loss": 0.0178,
+ "step": 505
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 4.766227283043912e-07,
+ "loss": 0.0056,
+ "step": 506
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 4.5860743599951186e-07,
+ "loss": 0.1492,
+ "step": 507
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.4093121340347824e-07,
+ "loss": 0.1143,
+ "step": 508
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.235946886711018e-07,
+ "loss": 0.1318,
+ "step": 509
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.0659847788544926e-07,
+ "loss": 0.0485,
+ "step": 510
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.899431850359503e-07,
+ "loss": 0.0155,
+ "step": 511
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.736294019969311e-07,
+ "loss": 0.0895,
+ "step": 512
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.5765770850658244e-07,
+ "loss": 0.0913,
+ "step": 513
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.420286721463562e-07,
+ "loss": 0.0394,
+ "step": 514
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.2674284832080127e-07,
+ "loss": 0.0125,
+ "step": 515
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.118007802378198e-07,
+ "loss": 0.082,
+ "step": 516
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 2.972029988893621e-07,
+ "loss": 0.2907,
+ "step": 517
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 2.8295002303256546e-07,
+ "loss": 0.0428,
+ "step": 518
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.6904235917131094e-07,
+ "loss": 0.1136,
+ "step": 519
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.554805015382289e-07,
+ "loss": 0.2987,
+ "step": 520
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.422649320771331e-07,
+ "loss": 0.1616,
+ "step": 521
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.293961204258932e-07,
+ "loss": 0.1623,
+ "step": 522
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.1687452389974829e-07,
+ "loss": 0.0428,
+ "step": 523
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.0470058747505516e-07,
+ "loss": 0.0309,
+ "step": 524
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.9287474377347238e-07,
+ "loss": 0.0533,
+ "step": 525
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.8139741304658566e-07,
+ "loss": 0.0565,
+ "step": 526
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.7026900316098217e-07,
+ "loss": 0.0545,
+ "step": 527
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.5948990958374543e-07,
+ "loss": 0.1014,
+ "step": 528
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.490605153684066e-07,
+ "loss": 0.2859,
+ "step": 529
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.3898119114133192e-07,
+ "loss": 0.2631,
+ "step": 530
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.292522950885533e-07,
+ "loss": 0.1341,
+ "step": 531
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.1987417294303748e-07,
+ "loss": 0.1001,
+ "step": 532
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.1084715797239798e-07,
+ "loss": 0.2334,
+ "step": 533
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.0217157096705676e-07,
+ "loss": 0.0793,
+ "step": 534
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 9.384772022884015e-08,
+ "loss": 0.0426,
+ "step": 535
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 8.587590156002635e-08,
+ "loss": 0.0174,
+ "step": 536
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 7.825639825282949e-08,
+ "loss": 0.1224,
+ "step": 537
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 7.098948107933656e-08,
+ "loss": 0.0296,
+ "step": 538
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 6.407540828188175e-08,
+ "loss": 0.0831,
+ "step": 539
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 5.7514425563870436e-08,
+ "loss": 0.0264,
+ "step": 540
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 5.1306766081048456e-08,
+ "loss": 0.0982,
+ "step": 541
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 4.545265043321645e-08,
+ "loss": 0.0393,
+ "step": 542
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 3.9952286656389506e-08,
+ "loss": 0.0695,
+ "step": 543
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 3.480587021540527e-08,
+ "loss": 0.1458,
+ "step": 544
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 3.001358399697618e-08,
+ "loss": 0.0065,
+ "step": 545
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 2.557559830319245e-08,
+ "loss": 0.1492,
+ "step": 546
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 2.1492070845468005e-08,
+ "loss": 0.094,
+ "step": 547
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.7763146738938307e-08,
+ "loss": 0.2497,
+ "step": 548
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.4388958497300043e-08,
+ "loss": 0.0833,
+ "step": 549
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.1369626028104874e-08,
+ "loss": 0.0528,
+ "step": 550
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 8.705256628499525e-09,
+ "loss": 0.0411,
+ "step": 551
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 6.39594498140883e-09,
+ "loss": 0.0253,
+ "step": 552
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 4.4417731521717576e-09,
+ "loss": 0.2015,
+ "step": 553
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 2.842810585627076e-09,
+ "loss": 0.0077,
+ "step": 554
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 1.5991141036475478e-09,
+ "loss": 0.1521,
+ "step": 555
+ },
+ {
+ "epoch": 1.0,
+ "step": 555,
+ "total_flos": 1196575985664.0,
+ "train_loss": 0.11872713004180172,
+ "train_runtime": 4113.0695,
+ "train_samples_per_second": 1.347,
+ "train_steps_per_second": 0.135
+ }
+ ],
+ "logging_steps": 1.0,
+ "max_steps": 555,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 1,
+ "save_steps": 500,
+ "total_flos": 1196575985664.0,
+ "train_batch_size": 10,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/CheckGuard Models/wholeimage/bank_no/finetune_lora_llava_mistral.sh b/CheckGuard Models/wholeimage/bank_no/finetune_lora_llava_mistral.sh
new file mode 100644
index 0000000000000000000000000000000000000000..a3fe33e51671ffc85a84fab135581b03b8095a37
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/finetune_lora_llava_mistral.sh
@@ -0,0 +1,43 @@
+#!/bin/bash
+# Use first parameter as GPU IDs, default to "0,1,2,3" if not provided
+GPU_IDS=${1:-0,1,2,3}
+
+
+CUDA_VISIBLE_DEVICES=0,1,2,3 deepspeed --include localhost:"$GPU_IDS" --master_port 29602\
+ llava/train/train_mem.py \
+ --lora_enable True --lora_r 128 --lora_alpha 256 --mm_projector_lr 2e-5 \
+ --deepspeed ./scripts/zero3.json \
+ --model_name_or_path liuhaotian/llava-v1.6-mistral-7b \
+ --version mistral_instruct \
+ --data_path /home/larry5/project/LLaVA-1.6-ft/data/peft/bank_no/bank_no_dataset.json \
+ --image_folder /home/larry5/project/LLaVA-1.6-ft/data/data/ \
+ --vision_tower openai/clip-vit-large-patch14-336 \
+ --mm_projector_type mlp2x_gelu \
+ --mm_vision_select_layer -2 \
+ --mm_use_im_start_end False \
+ --mm_use_im_patch_token False \
+ --mm_patch_merge_type spatial_unpad \
+ --image_aspect_ratio anyres \
+ --group_by_modality_length False \
+ --bf16 False \
+ --fp16 True \
+ --output_dir /home/larry5/project/LLaVA-1.6-ft/scripts_peft/mistral/lora/llava-lora-mistral-r128a256/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model \
+ --num_train_epochs 1 \
+ --per_device_train_batch_size 10 \
+ --per_device_eval_batch_size 1 \
+ --gradient_accumulation_steps 1 \
+ --evaluation_strategy "no" \
+ --save_strategy "steps" \
+ --save_steps 500 \
+ --save_total_limit 5 \
+ --learning_rate 2e-5 \
+ --weight_decay 0. \
+ --warmup_ratio 0.05 \
+ --lr_scheduler_type "cosine" \
+ --logging_steps 1 \
+ --tf32 True \
+ --model_max_length 4096 \
+ --gradient_checkpointing True \
+ --dataloader_num_workers 4 \
+ --lazy_preprocess True \
+ --report_to wandb \
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/README.md b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..bdb138eee6972419f6d60676388b52fd99ec478e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/README.md
@@ -0,0 +1,202 @@
+---
+library_name: peft
+base_model: liuhaotian/llava-v1.6-mistral-7b
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.10.0
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/adapter_config.json b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..d5f15e977cddb84e185e90736f37a523f49d60a3
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/adapter_config.json
@@ -0,0 +1,34 @@
+{
+ "alpha_pattern": {},
+ "auto_mapping": null,
+ "base_model_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "bias": "none",
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 256,
+ "lora_dropout": 0.05,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "r": 128,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "v_proj",
+ "q_proj",
+ "up_proj",
+ "k_proj",
+ "o_proj",
+ "gate_proj",
+ "down_proj"
+ ],
+ "task_type": "CAUSAL_LM",
+ "use_dora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..6794c7b26c12da85cf1c14635fd74fb41fd33371
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6eea7f9e47bb7d2f074c81e49ccf9648c1394c7fbb7e851b9ac64e47efa2c03b
+size 708924928
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/README.md b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..bdb138eee6972419f6d60676388b52fd99ec478e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/README.md
@@ -0,0 +1,202 @@
+---
+library_name: peft
+base_model: liuhaotian/llava-v1.6-mistral-7b
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.10.0
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_config.json b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..d5f15e977cddb84e185e90736f37a523f49d60a3
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_config.json
@@ -0,0 +1,34 @@
+{
+ "alpha_pattern": {},
+ "auto_mapping": null,
+ "base_model_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "bias": "none",
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 256,
+ "lora_dropout": 0.05,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "r": 128,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "v_proj",
+ "q_proj",
+ "up_proj",
+ "k_proj",
+ "o_proj",
+ "gate_proj",
+ "down_proj"
+ ],
+ "task_type": "CAUSAL_LM",
+ "use_dora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_model.safetensors b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..a312c3b4122fa707998f6937cc59f787cd537e86
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:645c293e131efb974f8c218e7d69c93bf50c753554c806d1cf561baa77311585
+size 1417762896
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_model_states.pt b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_model_states.pt
new file mode 100644
index 0000000000000000000000000000000000000000..2148fda79de7d1a6e7f0b5258183d9c00a5fddfc
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3183212032ce3f53bf011c0ea2d72e73d90e4ae83d758f3cb2661945c405d2e
+size 632242
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_optim_states.pt b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000000000000000000000000000000000000..d2577799dcf426ba6536f827d460216ca98db03a
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:67835d032628ab68661627ea5db2a21c8defdf7306ff43ec6d2d034f2a3add64
+size 4504787266
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/latest b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/latest
new file mode 100644
index 0000000000000000000000000000000000000000..f0b47ce15fff9a01b2a416a473b2148085048a50
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/latest
@@ -0,0 +1 @@
+global_step500
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/rng_state.pth b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..1f078010075b06c4b35cef3a20eef2119ad7e065
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1691c008dc15394c290eec92c6d96f1d3cc3096220a9fdad0f2210c4f3699fd5
+size 14244
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/scheduler.pt b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..77ccb4bfa6c448fe73ca8ad7989c41505cb6d3d2
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b37e2b05185c6152f2a40fb75a789b697d3a87176492c5cbb481ba82522c2163
+size 1064
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/special_tokens_map.json b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..14761dcf1466dc232bd41de9c21d4c617b15755e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/special_tokens_map.json
@@ -0,0 +1,24 @@
+{
+ "bos_token": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false
+ },
+ "eos_token": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false
+ },
+ "pad_token": "",
+ "unk_token": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false
+ }
+}
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer.model b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer.model
new file mode 100644
index 0000000000000000000000000000000000000000..8b443ef19c2a19acc3ac64fb9c3db4a72921dff6
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer.model
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
+size 493443
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer_config.json b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..d0ea5c3458cd84f0062b47fa0476bb328b3e208a
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer_config.json
@@ -0,0 +1,44 @@
+{
+ "add_bos_token": true,
+ "add_eos_token": false,
+ "added_tokens_decoder": {
+ "0": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "1": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "2": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ }
+ },
+ "additional_special_tokens": [],
+ "bos_token": "",
+ "chat_template": "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if message['role'] == 'user' %}{{ '[INST] ' + message['content'] + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ message['content'] + eos_token}}{% else %}{{ raise_exception('Only user and assistant roles are supported!') }}{% endif %}{% endfor %}",
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "",
+ "legacy": true,
+ "model_max_length": 4096,
+ "pad_token": "",
+ "padding_side": "right",
+ "sp_model_kwargs": {},
+ "spaces_between_special_tokens": false,
+ "tokenizer_class": "LlamaTokenizer",
+ "unk_token": "",
+ "use_default_system_prompt": false
+}
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/trainer_state.json b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..8ec01250d70779c6350965e3194f8401eda3e3c1
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/trainer_state.json
@@ -0,0 +1,3021 @@
+{
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.9242144177449169,
+ "eval_steps": 500,
+ "global_step": 500,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "epoch": 0.0,
+ "learning_rate": 7.142857142857143e-07,
+ "loss": 0.6789,
+ "step": 1
+ },
+ {
+ "epoch": 0.0,
+ "learning_rate": 1.4285714285714286e-06,
+ "loss": 0.8481,
+ "step": 2
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.1428571428571427e-06,
+ "loss": 0.663,
+ "step": 3
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.8571428571428573e-06,
+ "loss": 0.679,
+ "step": 4
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 3.5714285714285718e-06,
+ "loss": 1.0166,
+ "step": 5
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 4.2857142857142855e-06,
+ "loss": 0.4693,
+ "step": 6
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 5e-06,
+ "loss": 0.4891,
+ "step": 7
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 5.7142857142857145e-06,
+ "loss": 0.5523,
+ "step": 8
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 6.4285714285714295e-06,
+ "loss": 0.2909,
+ "step": 9
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.1428571428571436e-06,
+ "loss": 0.2598,
+ "step": 10
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.857142857142858e-06,
+ "loss": 0.2532,
+ "step": 11
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.857142857142858e-06,
+ "loss": 0.4867,
+ "step": 12
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 8.571428571428571e-06,
+ "loss": 0.4145,
+ "step": 13
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 8.571428571428571e-06,
+ "loss": 0.3161,
+ "step": 14
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 9.285714285714288e-06,
+ "loss": 0.1836,
+ "step": 15
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1e-05,
+ "loss": 0.3355,
+ "step": 16
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.0714285714285714e-05,
+ "loss": 0.2286,
+ "step": 17
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.1428571428571429e-05,
+ "loss": 0.3594,
+ "step": 18
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.2142857142857142e-05,
+ "loss": 0.2981,
+ "step": 19
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.2857142857142859e-05,
+ "loss": 0.3021,
+ "step": 20
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.3571428571428574e-05,
+ "loss": 0.3866,
+ "step": 21
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.4285714285714287e-05,
+ "loss": 0.2409,
+ "step": 22
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.5000000000000002e-05,
+ "loss": 0.1397,
+ "step": 23
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.5714285714285715e-05,
+ "loss": 0.1416,
+ "step": 24
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.642857142857143e-05,
+ "loss": 0.1838,
+ "step": 25
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.7142857142857142e-05,
+ "loss": 0.1505,
+ "step": 26
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.785714285714286e-05,
+ "loss": 0.3278,
+ "step": 27
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.8571428571428575e-05,
+ "loss": 0.2567,
+ "step": 28
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.928571428571429e-05,
+ "loss": 0.1218,
+ "step": 29
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 2e-05,
+ "loss": 0.2288,
+ "step": 30
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9999812486015525e-05,
+ "loss": 0.1348,
+ "step": 31
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9999249951094388e-05,
+ "loss": 0.3734,
+ "step": 32
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.999831241633323e-05,
+ "loss": 0.3169,
+ "step": 33
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9996999916892222e-05,
+ "loss": 0.1066,
+ "step": 34
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9995312501993765e-05,
+ "loss": 0.4434,
+ "step": 35
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9993250234920638e-05,
+ "loss": 0.198,
+ "step": 36
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9990813193013625e-05,
+ "loss": 0.115,
+ "step": 37
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9988001467668613e-05,
+ "loss": 0.2676,
+ "step": 38
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9984815164333163e-05,
+ "loss": 0.2201,
+ "step": 39
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9981254402502568e-05,
+ "loss": 0.1945,
+ "step": 40
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.997731931571535e-05,
+ "loss": 0.1391,
+ "step": 41
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9973010051548274e-05,
+ "loss": 0.2697,
+ "step": 42
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9968326771610797e-05,
+ "loss": 0.1562,
+ "step": 43
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9963269651539018e-05,
+ "loss": 0.2204,
+ "step": 44
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9957838880989076e-05,
+ "loss": 0.2729,
+ "step": 45
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9952034663630064e-05,
+ "loss": 0.441,
+ "step": 46
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9952034663630064e-05,
+ "loss": 0.1401,
+ "step": 47
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9945857217136365e-05,
+ "loss": 0.3727,
+ "step": 48
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9939306773179498e-05,
+ "loss": 0.3269,
+ "step": 49
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9932383577419432e-05,
+ "loss": 0.0801,
+ "step": 50
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9925087889495374e-05,
+ "loss": 0.2772,
+ "step": 51
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9917419983016025e-05,
+ "loss": 0.2253,
+ "step": 52
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9909380145549325e-05,
+ "loss": 0.2318,
+ "step": 53
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9900968678611664e-05,
+ "loss": 0.1809,
+ "step": 54
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.989218589765658e-05,
+ "loss": 0.1155,
+ "step": 55
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9883032132062926e-05,
+ "loss": 0.2356,
+ "step": 56
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9873507725122505e-05,
+ "loss": 0.1194,
+ "step": 57
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9863613034027224e-05,
+ "loss": 0.3272,
+ "step": 58
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.985334842985567e-05,
+ "loss": 0.183,
+ "step": 59
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9842714297559212e-05,
+ "loss": 0.1217,
+ "step": 60
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9831711035947552e-05,
+ "loss": 0.1388,
+ "step": 61
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9820339057673773e-05,
+ "loss": 0.2112,
+ "step": 62
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9808598789218866e-05,
+ "loss": 0.0917,
+ "step": 63
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.979649067087574e-05,
+ "loss": 0.1585,
+ "step": 64
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9784015156732693e-05,
+ "loss": 0.1446,
+ "step": 65
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.97711727146564e-05,
+ "loss": 0.3511,
+ "step": 66
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9757963826274357e-05,
+ "loss": 0.1019,
+ "step": 67
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9744388986956824e-05,
+ "loss": 0.1165,
+ "step": 68
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.973044870579824e-05,
+ "loss": 0.2189,
+ "step": 69
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.971614350559814e-05,
+ "loss": 0.1254,
+ "step": 70
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.970147392284154e-05,
+ "loss": 0.0627,
+ "step": 71
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9686440507678827e-05,
+ "loss": 0.0952,
+ "step": 72
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.967104382390511e-05,
+ "loss": 0.1867,
+ "step": 73
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9655284448939094e-05,
+ "loss": 0.2003,
+ "step": 74
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9639162973801426e-05,
+ "loss": 0.1188,
+ "step": 75
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9622680003092503e-05,
+ "loss": 0.1111,
+ "step": 76
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.960583615496984e-05,
+ "loss": 0.1203,
+ "step": 77
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9588632061124837e-05,
+ "loss": 0.1599,
+ "step": 78
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9571068366759143e-05,
+ "loss": 0.209,
+ "step": 79
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9553145730560415e-05,
+ "loss": 0.2183,
+ "step": 80
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.953486482467764e-05,
+ "loss": 0.1351,
+ "step": 81
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.951622633469592e-05,
+ "loss": 0.128,
+ "step": 82
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9497230959610757e-05,
+ "loss": 0.2241,
+ "step": 83
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9477879411801843e-05,
+ "loss": 0.0991,
+ "step": 84
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9458172417006347e-05,
+ "loss": 0.1165,
+ "step": 85
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9438110714291697e-05,
+ "loss": 0.0792,
+ "step": 86
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9417695056027847e-05,
+ "loss": 0.121,
+ "step": 87
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9396926207859085e-05,
+ "loss": 0.2727,
+ "step": 88
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9375804948675308e-05,
+ "loss": 0.1947,
+ "step": 89
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.935433207058281e-05,
+ "loss": 0.2155,
+ "step": 90
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.933250837887457e-05,
+ "loss": 0.0525,
+ "step": 91
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9310334692000077e-05,
+ "loss": 0.2401,
+ "step": 92
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9287811841534598e-05,
+ "loss": 0.0743,
+ "step": 93
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9264940672148018e-05,
+ "loss": 0.1659,
+ "step": 94
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9241722041573166e-05,
+ "loss": 0.1184,
+ "step": 95
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9218156820573618e-05,
+ "loss": 0.1207,
+ "step": 96
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9194245892911077e-05,
+ "loss": 0.1292,
+ "step": 97
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.916999015531221e-05,
+ "loss": 0.2059,
+ "step": 98
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9145390517435013e-05,
+ "loss": 0.1682,
+ "step": 99
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9120447901834708e-05,
+ "loss": 0.1403,
+ "step": 100
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.9095163243929143e-05,
+ "loss": 0.1752,
+ "step": 101
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.906953749196371e-05,
+ "loss": 0.1616,
+ "step": 102
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.9043571606975776e-05,
+ "loss": 0.1127,
+ "step": 103
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.901726656275866e-05,
+ "loss": 0.2236,
+ "step": 104
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8990623345825084e-05,
+ "loss": 0.2308,
+ "step": 105
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8963642955370203e-05,
+ "loss": 0.1739,
+ "step": 106
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8936326403234125e-05,
+ "loss": 0.1762,
+ "step": 107
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.890867471386395e-05,
+ "loss": 0.1457,
+ "step": 108
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.888068892427538e-05,
+ "loss": 0.2768,
+ "step": 109
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8852370084013783e-05,
+ "loss": 0.1389,
+ "step": 110
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.882371925511488e-05,
+ "loss": 0.2747,
+ "step": 111
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.879473751206489e-05,
+ "loss": 0.0542,
+ "step": 112
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8765425941760237e-05,
+ "loss": 0.1414,
+ "step": 113
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8735785643466786e-05,
+ "loss": 0.2482,
+ "step": 114
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8705817728778626e-05,
+ "loss": 0.1602,
+ "step": 115
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.867552332157637e-05,
+ "loss": 0.1342,
+ "step": 116
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8644903557985027e-05,
+ "loss": 0.077,
+ "step": 117
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8613959586331364e-05,
+ "loss": 0.0818,
+ "step": 118
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8582692567100866e-05,
+ "loss": 0.1443,
+ "step": 119
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.855110367289421e-05,
+ "loss": 0.1148,
+ "step": 120
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.851919408838327e-05,
+ "loss": 0.1661,
+ "step": 121
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8486965010266726e-05,
+ "loss": 0.1676,
+ "step": 122
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.845441764722514e-05,
+ "loss": 0.1288,
+ "step": 123
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.842155321987566e-05,
+ "loss": 0.0725,
+ "step": 124
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8388372960726228e-05,
+ "loss": 0.1258,
+ "step": 125
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8354878114129368e-05,
+ "loss": 0.068,
+ "step": 126
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8321069936235503e-05,
+ "loss": 0.1698,
+ "step": 127
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8286949694945864e-05,
+ "loss": 0.2038,
+ "step": 128
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8252518669864935e-05,
+ "loss": 0.0274,
+ "step": 129
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.821777815225245e-05,
+ "loss": 0.0564,
+ "step": 130
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8182729444974993e-05,
+ "loss": 0.1182,
+ "step": 131
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8147373862457107e-05,
+ "loss": 0.3175,
+ "step": 132
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8111712730632024e-05,
+ "loss": 0.1017,
+ "step": 133
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.807574738689193e-05,
+ "loss": 0.3348,
+ "step": 134
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8039479180037803e-05,
+ "loss": 0.3129,
+ "step": 135
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.800290947022884e-05,
+ "loss": 0.1095,
+ "step": 136
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7966039628931447e-05,
+ "loss": 0.1922,
+ "step": 137
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7928871038867785e-05,
+ "loss": 0.1022,
+ "step": 138
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.789140509396394e-05,
+ "loss": 0.2318,
+ "step": 139
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7853643199297632e-05,
+ "loss": 0.2374,
+ "step": 140
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7815586771045535e-05,
+ "loss": 0.1194,
+ "step": 141
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.777723723643014e-05,
+ "loss": 0.1914,
+ "step": 142
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.773859603366626e-05,
+ "loss": 0.0431,
+ "step": 143
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.769966461190707e-05,
+ "loss": 0.081,
+ "step": 144
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.766044443118978e-05,
+ "loss": 0.2162,
+ "step": 145
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.762093696238086e-05,
+ "loss": 0.1151,
+ "step": 146
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7581143687120877e-05,
+ "loss": 0.184,
+ "step": 147
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7541066097768965e-05,
+ "loss": 0.1963,
+ "step": 148
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.750070569734681e-05,
+ "loss": 0.1318,
+ "step": 149
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7460063999482314e-05,
+ "loss": 0.1163,
+ "step": 150
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7419142528352815e-05,
+ "loss": 0.1013,
+ "step": 151
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.737794281862794e-05,
+ "loss": 0.0957,
+ "step": 152
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7336466415412028e-05,
+ "loss": 0.2023,
+ "step": 153
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.729471487418621e-05,
+ "loss": 0.1398,
+ "step": 154
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7252689760750053e-05,
+ "loss": 0.1238,
+ "step": 155
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.721039265116285e-05,
+ "loss": 0.2201,
+ "step": 156
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7167825131684516e-05,
+ "loss": 0.0698,
+ "step": 157
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7124988798716084e-05,
+ "loss": 0.0312,
+ "step": 158
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7081885258739846e-05,
+ "loss": 0.1443,
+ "step": 159
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.7038516128259118e-05,
+ "loss": 0.1349,
+ "step": 160
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6994883033737582e-05,
+ "loss": 0.0751,
+ "step": 161
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.695098761153832e-05,
+ "loss": 0.0543,
+ "step": 162
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6906831507862446e-05,
+ "loss": 0.0533,
+ "step": 163
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.686241637868734e-05,
+ "loss": 0.1328,
+ "step": 164
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6817743889704564e-05,
+ "loss": 0.3057,
+ "step": 165
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6772815716257414e-05,
+ "loss": 0.1642,
+ "step": 166
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.672763354327804e-05,
+ "loss": 0.1479,
+ "step": 167
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6682199065224307e-05,
+ "loss": 0.1163,
+ "step": 168
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6636513986016215e-05,
+ "loss": 0.0395,
+ "step": 169
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6590580018972012e-05,
+ "loss": 0.0456,
+ "step": 170
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6544398886743934e-05,
+ "loss": 0.2018,
+ "step": 171
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.64979723212536e-05,
+ "loss": 0.1655,
+ "step": 172
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6451302063627067e-05,
+ "loss": 0.1805,
+ "step": 173
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6404389864129533e-05,
+ "loss": 0.2445,
+ "step": 174
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6357237482099682e-05,
+ "loss": 0.134,
+ "step": 175
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6309846685883726e-05,
+ "loss": 0.0976,
+ "step": 176
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6262219252769065e-05,
+ "loss": 0.0984,
+ "step": 177
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.621435696891765e-05,
+ "loss": 0.0495,
+ "step": 178
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6166261629298996e-05,
+ "loss": 0.1005,
+ "step": 179
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6117935037622848e-05,
+ "loss": 0.1399,
+ "step": 180
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.606937900627157e-05,
+ "loss": 0.2105,
+ "step": 181
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.6020595356232137e-05,
+ "loss": 0.142,
+ "step": 182
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5971585917027864e-05,
+ "loss": 0.0791,
+ "step": 183
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5922352526649803e-05,
+ "loss": 0.2,
+ "step": 184
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.587289703148779e-05,
+ "loss": 0.1317,
+ "step": 185
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5823221286261217e-05,
+ "loss": 0.1656,
+ "step": 186
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5773327153949465e-05,
+ "loss": 0.3358,
+ "step": 187
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.572321650572205e-05,
+ "loss": 0.2216,
+ "step": 188
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.567289122086843e-05,
+ "loss": 0.0937,
+ "step": 189
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5622353186727542e-05,
+ "loss": 0.0995,
+ "step": 190
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.557160429861702e-05,
+ "loss": 0.2324,
+ "step": 191
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5520646459762102e-05,
+ "loss": 0.2847,
+ "step": 192
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5469481581224274e-05,
+ "loss": 0.1242,
+ "step": 193
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5418111581829575e-05,
+ "loss": 0.1771,
+ "step": 194
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.536653838809667e-05,
+ "loss": 0.2115,
+ "step": 195
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.531476393416456e-05,
+ "loss": 0.074,
+ "step": 196
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5262790161720082e-05,
+ "loss": 0.0893,
+ "step": 197
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5210619019925066e-05,
+ "loss": 0.0644,
+ "step": 198
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5158252465343242e-05,
+ "loss": 0.2146,
+ "step": 199
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5105692461866874e-05,
+ "loss": 0.2579,
+ "step": 200
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.50529409806431e-05,
+ "loss": 0.0806,
+ "step": 201
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5000000000000002e-05,
+ "loss": 0.0806,
+ "step": 202
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4946871505372426e-05,
+ "loss": 0.132,
+ "step": 203
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4893557489227518e-05,
+ "loss": 0.1438,
+ "step": 204
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4840059950989992e-05,
+ "loss": 0.1703,
+ "step": 205
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.478638089696716e-05,
+ "loss": 0.0903,
+ "step": 206
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4732522340273686e-05,
+ "loss": 0.1515,
+ "step": 207
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.467848630075608e-05,
+ "loss": 0.2156,
+ "step": 208
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4624274804916958e-05,
+ "loss": 0.0783,
+ "step": 209
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.456988988583904e-05,
+ "loss": 0.1432,
+ "step": 210
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4515333583108896e-05,
+ "loss": 0.1716,
+ "step": 211
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4460607942740468e-05,
+ "loss": 0.2328,
+ "step": 212
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4405715017098333e-05,
+ "loss": 0.1317,
+ "step": 213
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4350656864820733e-05,
+ "loss": 0.097,
+ "step": 214
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4295435550742372e-05,
+ "loss": 0.1547,
+ "step": 215
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4240053145816968e-05,
+ "loss": 0.0737,
+ "step": 216
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4184511727039612e-05,
+ "loss": 0.0926,
+ "step": 217
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4128813377368851e-05,
+ "loss": 0.0824,
+ "step": 218
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4072960185648576e-05,
+ "loss": 0.1236,
+ "step": 219
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.4016954246529697e-05,
+ "loss": 0.157,
+ "step": 220
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.396079766039157e-05,
+ "loss": 0.1241,
+ "step": 221
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3904492533263243e-05,
+ "loss": 0.1243,
+ "step": 222
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3848040976744459e-05,
+ "loss": 0.1429,
+ "step": 223
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3791445107926478e-05,
+ "loss": 0.0321,
+ "step": 224
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3734707049312674e-05,
+ "loss": 0.0398,
+ "step": 225
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3677828928738934e-05,
+ "loss": 0.2625,
+ "step": 226
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3620812879293864e-05,
+ "loss": 0.0926,
+ "step": 227
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3563661039238785e-05,
+ "loss": 0.06,
+ "step": 228
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3506375551927546e-05,
+ "loss": 0.2397,
+ "step": 229
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3448958565726144e-05,
+ "loss": 0.157,
+ "step": 230
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3391412233932148e-05,
+ "loss": 0.1105,
+ "step": 231
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3333738714693958e-05,
+ "loss": 0.0877,
+ "step": 232
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3275940170929845e-05,
+ "loss": 0.1821,
+ "step": 233
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3218018770246858e-05,
+ "loss": 0.0166,
+ "step": 234
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3159976684859528e-05,
+ "loss": 0.118,
+ "step": 235
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3101816091508389e-05,
+ "loss": 0.2289,
+ "step": 236
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3043539171378362e-05,
+ "loss": 0.0518,
+ "step": 237
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2985148110016947e-05,
+ "loss": 0.1012,
+ "step": 238
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.292664509725226e-05,
+ "loss": 0.2009,
+ "step": 239
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2868032327110904e-05,
+ "loss": 0.252,
+ "step": 240
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2809311997735697e-05,
+ "loss": 0.2044,
+ "step": 241
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2750486311303218e-05,
+ "loss": 0.1908,
+ "step": 242
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2691557473941246e-05,
+ "loss": 0.3064,
+ "step": 243
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2632527695645993e-05,
+ "loss": 0.091,
+ "step": 244
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.257339919019925e-05,
+ "loss": 0.0606,
+ "step": 245
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2514174175085346e-05,
+ "loss": 0.147,
+ "step": 246
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2454854871407993e-05,
+ "loss": 0.2029,
+ "step": 247
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.239544350380699e-05,
+ "loss": 0.0851,
+ "step": 248
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2335942300374788e-05,
+ "loss": 0.0904,
+ "step": 249
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2276353492572937e-05,
+ "loss": 0.0721,
+ "step": 250
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2216679315148388e-05,
+ "loss": 0.1488,
+ "step": 251
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2156922006049703e-05,
+ "loss": 0.1927,
+ "step": 252
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2097083806343104e-05,
+ "loss": 0.029,
+ "step": 253
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2037166960128443e-05,
+ "loss": 0.0301,
+ "step": 254
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1977173714455034e-05,
+ "loss": 0.1231,
+ "step": 255
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1917106319237386e-05,
+ "loss": 0.0348,
+ "step": 256
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1856967027170818e-05,
+ "loss": 0.0869,
+ "step": 257
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1796758093646989e-05,
+ "loss": 0.1164,
+ "step": 258
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1736481776669307e-05,
+ "loss": 0.0388,
+ "step": 259
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1676140336768236e-05,
+ "loss": 0.0433,
+ "step": 260
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.161573603691655e-05,
+ "loss": 0.1996,
+ "step": 261
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1555271142444433e-05,
+ "loss": 0.2182,
+ "step": 262
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1494747920954545e-05,
+ "loss": 0.0509,
+ "step": 263
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1434168642236964e-05,
+ "loss": 0.1078,
+ "step": 264
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1373535578184083e-05,
+ "loss": 0.0412,
+ "step": 265
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1312851002705383e-05,
+ "loss": 0.2425,
+ "step": 266
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1252117191642175e-05,
+ "loss": 0.1119,
+ "step": 267
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1191336422682237e-05,
+ "loss": 0.0455,
+ "step": 268
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1130510975274408e-05,
+ "loss": 0.2613,
+ "step": 269
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1069643130543084e-05,
+ "loss": 0.0651,
+ "step": 270
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1008735171202685e-05,
+ "loss": 0.1155,
+ "step": 271
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.0947789381472035e-05,
+ "loss": 0.0661,
+ "step": 272
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.0886808046988716e-05,
+ "loss": 0.0881,
+ "step": 273
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0825793454723325e-05,
+ "loss": 0.1123,
+ "step": 274
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0764747892893724e-05,
+ "loss": 0.14,
+ "step": 275
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0703673650879219e-05,
+ "loss": 0.0889,
+ "step": 276
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0642573019134703e-05,
+ "loss": 0.1333,
+ "step": 277
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0581448289104759e-05,
+ "loss": 0.0608,
+ "step": 278
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0520301753137725e-05,
+ "loss": 0.2882,
+ "step": 279
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.045913570439972e-05,
+ "loss": 0.0661,
+ "step": 280
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0397952436788643e-05,
+ "loss": 0.107,
+ "step": 281
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0336754244848156e-05,
+ "loss": 0.0499,
+ "step": 282
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0275543423681622e-05,
+ "loss": 0.237,
+ "step": 283
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0214322268866033e-05,
+ "loss": 0.0301,
+ "step": 284
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0153093076365923e-05,
+ "loss": 0.0904,
+ "step": 285
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0091858142447266e-05,
+ "loss": 0.0165,
+ "step": 286
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0030619763591348e-05,
+ "loss": 0.0791,
+ "step": 287
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.969380236408656e-06,
+ "loss": 0.1997,
+ "step": 288
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.908141857552737e-06,
+ "loss": 0.0155,
+ "step": 289
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.846906923634079e-06,
+ "loss": 0.0457,
+ "step": 290
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.785677731133972e-06,
+ "loss": 0.0203,
+ "step": 291
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.724456576318383e-06,
+ "loss": 0.2384,
+ "step": 292
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.663245755151847e-06,
+ "loss": 0.1459,
+ "step": 293
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.602047563211359e-06,
+ "loss": 0.2249,
+ "step": 294
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.540864295600282e-06,
+ "loss": 0.037,
+ "step": 295
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.479698246862277e-06,
+ "loss": 0.145,
+ "step": 296
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.418551710895243e-06,
+ "loss": 0.1501,
+ "step": 297
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.3574269808653e-06,
+ "loss": 0.0727,
+ "step": 298
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.296326349120786e-06,
+ "loss": 0.0992,
+ "step": 299
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.23525210710628e-06,
+ "loss": 0.2516,
+ "step": 300
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.174206545276678e-06,
+ "loss": 0.0628,
+ "step": 301
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.113191953011287e-06,
+ "loss": 0.132,
+ "step": 302
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.052210618527966e-06,
+ "loss": 0.0908,
+ "step": 303
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.991264828797319e-06,
+ "loss": 0.1432,
+ "step": 304
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.93035686945692e-06,
+ "loss": 0.0493,
+ "step": 305
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.869489024725595e-06,
+ "loss": 0.0578,
+ "step": 306
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.808663577317765e-06,
+ "loss": 0.0909,
+ "step": 307
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.747882808357828e-06,
+ "loss": 0.0646,
+ "step": 308
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.687148997294622e-06,
+ "loss": 0.1308,
+ "step": 309
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.626464421815919e-06,
+ "loss": 0.0729,
+ "step": 310
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.565831357763039e-06,
+ "loss": 0.1871,
+ "step": 311
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.505252079045459e-06,
+ "loss": 0.1577,
+ "step": 312
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.444728857555572e-06,
+ "loss": 0.1844,
+ "step": 313
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.384263963083453e-06,
+ "loss": 0.1673,
+ "step": 314
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.323859663231768e-06,
+ "loss": 0.1898,
+ "step": 315
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.263518223330698e-06,
+ "loss": 0.1106,
+ "step": 316
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.203241906353014e-06,
+ "loss": 0.0476,
+ "step": 317
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.143032972829184e-06,
+ "loss": 0.1432,
+ "step": 318
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.082893680762619e-06,
+ "loss": 0.0249,
+ "step": 319
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.022826285544967e-06,
+ "loss": 0.0762,
+ "step": 320
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.962833039871562e-06,
+ "loss": 0.1468,
+ "step": 321
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.902916193656898e-06,
+ "loss": 0.0272,
+ "step": 322
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.843077993950302e-06,
+ "loss": 0.0495,
+ "step": 323
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.783320684851613e-06,
+ "loss": 0.1958,
+ "step": 324
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.72364650742707e-06,
+ "loss": 0.0869,
+ "step": 325
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.664057699625215e-06,
+ "loss": 0.2957,
+ "step": 326
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.604556496193015e-06,
+ "loss": 0.0833,
+ "step": 327
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.545145128592009e-06,
+ "loss": 0.0978,
+ "step": 328
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.485825824914658e-06,
+ "loss": 0.1941,
+ "step": 329
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.426600809800753e-06,
+ "loss": 0.0384,
+ "step": 330
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.367472304354011e-06,
+ "loss": 0.0872,
+ "step": 331
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.308442526058757e-06,
+ "loss": 0.1051,
+ "step": 332
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.249513688696786e-06,
+ "loss": 0.0918,
+ "step": 333
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.190688002264308e-06,
+ "loss": 0.2169,
+ "step": 334
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.131967672889101e-06,
+ "loss": 0.1647,
+ "step": 335
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.073354902747742e-06,
+ "loss": 0.0585,
+ "step": 336
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.014851889983058e-06,
+ "loss": 0.1743,
+ "step": 337
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 6.956460828621641e-06,
+ "loss": 0.3001,
+ "step": 338
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.898183908491617e-06,
+ "loss": 0.0977,
+ "step": 339
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.840023315140476e-06,
+ "loss": 0.0549,
+ "step": 340
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.781981229753145e-06,
+ "loss": 0.0738,
+ "step": 341
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.7240598290701585e-06,
+ "loss": 0.027,
+ "step": 342
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.666261285306048e-06,
+ "loss": 0.0647,
+ "step": 343
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.608587766067853e-06,
+ "loss": 0.0531,
+ "step": 344
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.551041434273862e-06,
+ "loss": 0.0582,
+ "step": 345
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.4936244480724575e-06,
+ "loss": 0.2357,
+ "step": 346
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.4363389607612204e-06,
+ "loss": 0.0614,
+ "step": 347
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.379187120706138e-06,
+ "loss": 0.1516,
+ "step": 348
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.322171071261071e-06,
+ "loss": 0.2906,
+ "step": 349
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.265292950687329e-06,
+ "loss": 0.0402,
+ "step": 350
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.208554892073528e-06,
+ "loss": 0.0895,
+ "step": 351
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.151959023255545e-06,
+ "loss": 0.109,
+ "step": 352
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.095507466736763e-06,
+ "loss": 0.1338,
+ "step": 353
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.039202339608432e-06,
+ "loss": 0.0541,
+ "step": 354
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.983045753470308e-06,
+ "loss": 0.0614,
+ "step": 355
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.927039814351426e-06,
+ "loss": 0.2844,
+ "step": 356
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.871186622631155e-06,
+ "loss": 0.1412,
+ "step": 357
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.815488272960388e-06,
+ "loss": 0.0575,
+ "step": 358
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.759946854183036e-06,
+ "loss": 0.1047,
+ "step": 359
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.704564449257635e-06,
+ "loss": 0.2065,
+ "step": 360
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.649343135179271e-06,
+ "loss": 0.0995,
+ "step": 361
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.59428498290167e-06,
+ "loss": 0.1517,
+ "step": 362
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.539392057259536e-06,
+ "loss": 0.1122,
+ "step": 363
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.484666416891109e-06,
+ "loss": 0.0992,
+ "step": 364
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.430110114160965e-06,
+ "loss": 0.1303,
+ "step": 365
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.375725195083046e-06,
+ "loss": 0.1192,
+ "step": 366
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.321513699243924e-06,
+ "loss": 0.0991,
+ "step": 367
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.267477659726319e-06,
+ "loss": 0.077,
+ "step": 368
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.213619103032845e-06,
+ "loss": 0.1052,
+ "step": 369
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.159940049010015e-06,
+ "loss": 0.2359,
+ "step": 370
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.106442510772489e-06,
+ "loss": 0.0501,
+ "step": 371
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.053128494627578e-06,
+ "loss": 0.0803,
+ "step": 372
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.000000000000003e-06,
+ "loss": 0.2073,
+ "step": 373
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.947059019356904e-06,
+ "loss": 0.0479,
+ "step": 374
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.89430753813313e-06,
+ "loss": 0.125,
+ "step": 375
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.8417475346567635e-06,
+ "loss": 0.0715,
+ "step": 376
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.78938098007494e-06,
+ "loss": 0.0242,
+ "step": 377
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.737209838279923e-06,
+ "loss": 0.1242,
+ "step": 378
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.685236065835443e-06,
+ "loss": 0.1771,
+ "step": 379
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.633461611903336e-06,
+ "loss": 0.1037,
+ "step": 380
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.581888418170429e-06,
+ "loss": 0.0733,
+ "step": 381
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.530518418775734e-06,
+ "loss": 0.0565,
+ "step": 382
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.479353540237903e-06,
+ "loss": 0.1092,
+ "step": 383
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.4283957013829845e-06,
+ "loss": 0.0371,
+ "step": 384
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.3776468132724605e-06,
+ "loss": 0.1105,
+ "step": 385
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.327108779131573e-06,
+ "loss": 0.1856,
+ "step": 386
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.276783494277954e-06,
+ "loss": 0.1237,
+ "step": 387
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.226672846050538e-06,
+ "loss": 0.2521,
+ "step": 388
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.176778713738787e-06,
+ "loss": 0.0565,
+ "step": 389
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.127102968512214e-06,
+ "loss": 0.0518,
+ "step": 390
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.077647473350201e-06,
+ "loss": 0.0735,
+ "step": 391
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.028414082972141e-06,
+ "loss": 0.0786,
+ "step": 392
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.9794046437678705e-06,
+ "loss": 0.025,
+ "step": 393
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.930620993728434e-06,
+ "loss": 0.2235,
+ "step": 394
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.882064962377154e-06,
+ "loss": 0.1307,
+ "step": 395
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.83373837070101e-06,
+ "loss": 0.0224,
+ "step": 396
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.7856430310823546e-06,
+ "loss": 0.1109,
+ "step": 397
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.737780747230941e-06,
+ "loss": 0.0624,
+ "step": 398
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6901533141162804e-06,
+ "loss": 0.055,
+ "step": 399
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6427625179003223e-06,
+ "loss": 0.2079,
+ "step": 400
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.595610135870472e-06,
+ "loss": 0.2215,
+ "step": 401
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.548697936372937e-06,
+ "loss": 0.1016,
+ "step": 402
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.5020276787464058e-06,
+ "loss": 0.1229,
+ "step": 403
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.455601113256073e-06,
+ "loss": 0.0759,
+ "step": 404
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.4094199810279926e-06,
+ "loss": 0.1667,
+ "step": 405
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.3634860139837877e-06,
+ "loss": 0.048,
+ "step": 406
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.317800934775696e-06,
+ "loss": 0.0543,
+ "step": 407
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.2723664567219627e-06,
+ "loss": 0.1656,
+ "step": 408
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.2271842837425917e-06,
+ "loss": 0.0409,
+ "step": 409
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.1822561102954373e-06,
+ "loss": 0.1173,
+ "step": 410
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.1375836213126653e-06,
+ "loss": 0.0964,
+ "step": 411
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0931684921375572e-06,
+ "loss": 0.0432,
+ "step": 412
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0490123884616795e-06,
+ "loss": 0.1451,
+ "step": 413
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 3.0051169662624224e-06,
+ "loss": 0.1226,
+ "step": 414
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.9614838717408866e-06,
+ "loss": 0.096,
+ "step": 415
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.918114741260156e-06,
+ "loss": 0.1152,
+ "step": 416
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.8750112012839215e-06,
+ "loss": 0.0575,
+ "step": 417
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.8321748683154893e-06,
+ "loss": 0.097,
+ "step": 418
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.7896073488371535e-06,
+ "loss": 0.0513,
+ "step": 419
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.7473102392499517e-06,
+ "loss": 0.0566,
+ "step": 420
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.7052851258137936e-06,
+ "loss": 0.0193,
+ "step": 421
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.663533584587974e-06,
+ "loss": 0.1507,
+ "step": 422
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.622057181372063e-06,
+ "loss": 0.0208,
+ "step": 423
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.580857471647186e-06,
+ "loss": 0.0893,
+ "step": 424
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.539936000517689e-06,
+ "loss": 0.0988,
+ "step": 425
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4992943026531935e-06,
+ "loss": 0.0368,
+ "step": 426
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4589339022310386e-06,
+ "loss": 0.0911,
+ "step": 427
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4188563128791255e-06,
+ "loss": 0.1093,
+ "step": 428
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.379063037619146e-06,
+ "loss": 0.0717,
+ "step": 429
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.339555568810221e-06,
+ "loss": 0.1486,
+ "step": 430
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.300335388092929e-06,
+ "loss": 0.1174,
+ "step": 431
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.261403966333742e-06,
+ "loss": 0.2022,
+ "step": 432
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.2227627635698624e-06,
+ "loss": 0.0376,
+ "step": 433
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.1844132289544684e-06,
+ "loss": 0.3022,
+ "step": 434
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.1463568007023706e-06,
+ "loss": 0.0121,
+ "step": 435
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.1085949060360654e-06,
+ "loss": 0.1441,
+ "step": 436
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0711289611322204e-06,
+ "loss": 0.0457,
+ "step": 437
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0339603710685574e-06,
+ "loss": 0.0324,
+ "step": 438
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.9970905297711606e-06,
+ "loss": 0.045,
+ "step": 439
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.9605208199621993e-06,
+ "loss": 0.0644,
+ "step": 440
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.924252613108073e-06,
+ "loss": 0.0743,
+ "step": 441
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8882872693679787e-06,
+ "loss": 0.054,
+ "step": 442
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8526261375428955e-06,
+ "loss": 0.1679,
+ "step": 443
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8172705550250093e-06,
+ "loss": 0.0666,
+ "step": 444
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7822218477475496e-06,
+ "loss": 0.2,
+ "step": 445
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7474813301350668e-06,
+ "loss": 0.1191,
+ "step": 446
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.7130503050541368e-06,
+ "loss": 0.1166,
+ "step": 447
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6789300637645e-06,
+ "loss": 0.0089,
+ "step": 448
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6451218858706374e-06,
+ "loss": 0.0848,
+ "step": 449
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6116270392737753e-06,
+ "loss": 0.1263,
+ "step": 450
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.578446780124344e-06,
+ "loss": 0.1338,
+ "step": 451
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.5455823527748626e-06,
+ "loss": 0.0566,
+ "step": 452
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.5130349897332764e-06,
+ "loss": 0.0618,
+ "step": 453
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4808059116167306e-06,
+ "loss": 0.0259,
+ "step": 454
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4488963271057943e-06,
+ "loss": 0.1682,
+ "step": 455
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4173074328991376e-06,
+ "loss": 0.0967,
+ "step": 456
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.3860404136686411e-06,
+ "loss": 0.0799,
+ "step": 457
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.355096442014977e-06,
+ "loss": 0.1426,
+ "step": 458
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3244766784236307e-06,
+ "loss": 0.1401,
+ "step": 459
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.294182271221377e-06,
+ "loss": 0.0526,
+ "step": 460
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.2642143565332154e-06,
+ "loss": 0.1516,
+ "step": 461
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.2345740582397647e-06,
+ "loss": 0.0326,
+ "step": 462
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.2052624879351105e-06,
+ "loss": 0.0517,
+ "step": 463
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.176280744885121e-06,
+ "loss": 0.094,
+ "step": 464
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1476299159862204e-06,
+ "loss": 0.0684,
+ "step": 465
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1193110757246251e-06,
+ "loss": 0.0845,
+ "step": 466
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.09132528613605e-06,
+ "loss": 0.1105,
+ "step": 467
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.0636735967658785e-06,
+ "loss": 0.0947,
+ "step": 468
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.0363570446297999e-06,
+ "loss": 0.0685,
+ "step": 469
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.0093766541749206e-06,
+ "loss": 0.0902,
+ "step": 470
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.827334372413444e-07,
+ "loss": 0.0257,
+ "step": 471
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.564283930242258e-07,
+ "loss": 0.1048,
+ "step": 472
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.304625080362939e-07,
+ "loss": 0.1365,
+ "step": 473
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 9.048367560708604e-07,
+ "loss": 0.2323,
+ "step": 474
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.79552098165296e-07,
+ "loss": 0.0435,
+ "step": 475
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.546094825649909e-07,
+ "loss": 0.0644,
+ "step": 476
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.300098446877925e-07,
+ "loss": 0.0884,
+ "step": 477
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.057541070889229e-07,
+ "loss": 0.1381,
+ "step": 478
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.818431794263837e-07,
+ "loss": 0.0472,
+ "step": 479
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.582779584268374e-07,
+ "loss": 0.0606,
+ "step": 480
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.350593278519824e-07,
+ "loss": 0.0325,
+ "step": 481
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.121881584654056e-07,
+ "loss": 0.0391,
+ "step": 482
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 6.896653079999249e-07,
+ "loss": 0.0965,
+ "step": 483
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 6.67491621125429e-07,
+ "loss": 0.0288,
+ "step": 484
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.45667929417193e-07,
+ "loss": 0.0608,
+ "step": 485
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.241950513246931e-07,
+ "loss": 0.0619,
+ "step": 486
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.030737921409169e-07,
+ "loss": 0.2691,
+ "step": 487
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 5.823049439721562e-07,
+ "loss": 0.1071,
+ "step": 488
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 5.618892857083069e-07,
+ "loss": 0.1501,
+ "step": 489
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.418275829936537e-07,
+ "loss": 0.0807,
+ "step": 490
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.221205881981594e-07,
+ "loss": 0.0666,
+ "step": 491
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.027690403892461e-07,
+ "loss": 0.0993,
+ "step": 492
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 4.837736653040825e-07,
+ "loss": 0.2467,
+ "step": 493
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 4.6513517532236096e-07,
+ "loss": 0.0563,
+ "step": 494
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 4.468542694395861e-07,
+ "loss": 0.0792,
+ "step": 495
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.2893163324085886e-07,
+ "loss": 0.0648,
+ "step": 496
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.113679388751635e-07,
+ "loss": 0.3011,
+ "step": 497
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.941638450301644e-07,
+ "loss": 0.221,
+ "step": 498
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.773199969074959e-07,
+ "loss": 0.0961,
+ "step": 499
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.608370261985761e-07,
+ "loss": 0.0816,
+ "step": 500
+ }
+ ],
+ "logging_steps": 1.0,
+ "max_steps": 541,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 1,
+ "save_steps": 500,
+ "total_flos": 1194534973440.0,
+ "train_batch_size": 10,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/training_args.bin b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..6f6c18888b670fc1c3b31717a960603a6a87885d
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39a1b7853796df569dea25e1cfc44a0eebd9c31bc3cf60c516910866a1ef6ae7
+size 6968
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/zero_to_fp32.py b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..c98caae31534368be22b67fc4ae906836c992a8d
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/zero_to_fp32.py
@@ -0,0 +1,587 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+ FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+ FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+ buffers: dict()
+ param_shapes: dict()
+ shared_params: list
+ ds_version: int
+ frozen_param_shapes: dict()
+ frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+ return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+ '''
+ alist.sort(key=natural_keys) sorts in human order
+ http://nedbatchelder.com/blog/200712/human_sorting.html
+ (See Toothy's implementation in the comments)
+ '''
+ return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+ if not os.path.isdir(checkpoint_dir):
+ raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+ # there should be only one file
+ if zero_stage <= 2:
+ file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+ elif zero_stage == 3:
+ file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+ if not os.path.exists(file):
+ raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+ return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+ # XXX: need to test that this simple glob rule works for multi-node setup too
+ ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+ if len(ckpt_files) == 0:
+ raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+ return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+ return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+ return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+ zero_model_states = []
+ for file in files:
+ state_dict = torch.load(file, map_location=device)
+
+ if BUFFER_NAMES not in state_dict:
+ raise ValueError(f"{file} is not a model state checkpoint")
+ buffer_names = state_dict[BUFFER_NAMES]
+ if debug:
+ print("Found buffers:", buffer_names)
+
+ # recover just the buffers while restoring them to fp32 if they were saved in fp16
+ buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+ param_shapes = state_dict[PARAM_SHAPES]
+
+ # collect parameters that are included in param_shapes
+ param_names = []
+ for s in param_shapes:
+ for name in s.keys():
+ param_names.append(name)
+
+ # update with frozen parameters
+ frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+ if frozen_param_shapes is not None:
+ if debug:
+ print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+ param_names += list(frozen_param_shapes.keys())
+
+ # handle shared params
+ shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+ ds_version = state_dict.get(DS_VERSION, None)
+
+ frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+ z_model_state = zero_model_state(buffers=buffers,
+ param_shapes=param_shapes,
+ shared_params=shared_params,
+ ds_version=ds_version,
+ frozen_param_shapes=frozen_param_shapes,
+ frozen_param_fragments=frozen_param_fragments)
+ zero_model_states.append(z_model_state)
+
+ return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+ total_files = len(files)
+ state_dicts = []
+ for f in files:
+ state_dict = torch.load(f, map_location=device)
+ # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+ # and also handle the case where it was already removed by another helper script
+ state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+ state_dicts.append(state_dict)
+
+ if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+ raise ValueError(f"{files[0]} is not a zero checkpoint")
+ zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+ world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+ # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+ # parameters can be different from data parallelism for non-expert parameters. So we can just
+ # use the max of the partition_count to get the dp world_size.
+
+ if type(world_size) is list:
+ world_size = max(world_size)
+
+ if world_size != total_files:
+ raise ValueError(
+ f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+ "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+ )
+
+ # the groups are named differently in each stage
+ if zero_stage <= 2:
+ fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+ elif zero_stage == 3:
+ fp32_groups_key = FP32_FLAT_GROUPS
+ else:
+ raise ValueError(f"unknown zero stage {zero_stage}")
+
+ if zero_stage <= 2:
+ fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+ elif zero_stage == 3:
+ # if there is more than one param group, there will be multiple flattened tensors - one
+ # flattened tensor per group - for simplicity merge them into a single tensor
+ #
+ # XXX: could make the script more memory efficient for when there are multiple groups - it
+ # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+ fp32_flat_groups = [
+ torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+ ]
+
+ return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir):
+ """
+ Returns fp32 state_dict reconstructed from ds checkpoint
+
+ Args:
+ - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+ """
+ print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+ optim_files = get_optim_files(ds_checkpoint_dir)
+ zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+ print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+ model_files = get_model_state_files(ds_checkpoint_dir)
+
+ zero_model_states = parse_model_states(model_files)
+ print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+ if zero_stage <= 2:
+ return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+ elif zero_stage == 3:
+ return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+ if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+ return
+
+ frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+ frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+ if debug:
+ num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+ print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+ wanted_params = len(frozen_param_shapes)
+ wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+ avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+ print(f'Frozen params: Have {avail_numel} numels to process.')
+ print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+ total_params = 0
+ total_numel = 0
+ for name, shape in frozen_param_shapes.items():
+ total_params += 1
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+
+ state_dict[name] = frozen_param_fragments[name]
+
+ if debug:
+ print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+ print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+ param_shapes = zero_model_states[0].param_shapes
+
+ # Reconstruction protocol:
+ #
+ # XXX: document this
+
+ if debug:
+ for i in range(world_size):
+ for j in range(len(fp32_flat_groups[0])):
+ print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+ # XXX: memory usage doubles here (zero2)
+ num_param_groups = len(fp32_flat_groups[0])
+ merged_single_partition_of_fp32_groups = []
+ for i in range(num_param_groups):
+ merged_partitions = [sd[i] for sd in fp32_flat_groups]
+ full_single_fp32_vector = torch.cat(merged_partitions, 0)
+ merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+ avail_numel = sum(
+ [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+ if debug:
+ wanted_params = sum([len(shapes) for shapes in param_shapes])
+ wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+ # not asserting if there is a mismatch due to possible padding
+ print(f"Have {avail_numel} numels to process.")
+ print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+ # params
+ # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+ # out-of-core computing solution
+ total_numel = 0
+ total_params = 0
+ for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+ offset = 0
+ avail_numel = full_single_fp32_vector.numel()
+ for name, shape in shapes.items():
+
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+ total_params += 1
+
+ if debug:
+ print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+ state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+ offset += unpartitioned_numel
+
+ # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+ # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+ # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+ # live optimizer object, so we are checking that the numbers are within the right range
+ align_to = 2 * world_size
+
+ def zero2_align(x):
+ return align_to * math.ceil(x / align_to)
+
+ if debug:
+ print(f"original offset={offset}, avail_numel={avail_numel}")
+
+ offset = zero2_align(offset)
+ avail_numel = zero2_align(avail_numel)
+
+ if debug:
+ print(f"aligned offset={offset}, avail_numel={avail_numel}")
+
+ # Sanity check
+ if offset != avail_numel:
+ raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+ print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+ state_dict = OrderedDict()
+
+ # buffers
+ buffers = zero_model_states[0].buffers
+ state_dict.update(buffers)
+ if debug:
+ print(f"added {len(buffers)} buffers")
+
+ _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+ _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+ # recover shared parameters
+ for pair in zero_model_states[0].shared_params:
+ if pair[1] in state_dict:
+ state_dict[pair[0]] = state_dict[pair[1]]
+
+ return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+ remainder = unpartitioned_numel % world_size
+ padding_numel = (world_size - remainder) if remainder else 0
+ partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+ return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+ if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+ return
+
+ if debug:
+ for i in range(world_size):
+ num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+ print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+ frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+ wanted_params = len(frozen_param_shapes)
+ wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+ avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+ print(f'Frozen params: Have {avail_numel} numels to process.')
+ print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+ total_params = 0
+ total_numel = 0
+ for name, shape in zero_model_states[0].frozen_param_shapes.items():
+ total_params += 1
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+
+ param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+ state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+ partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+ if debug:
+ print(
+ f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+ )
+
+ print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+ param_shapes = zero_model_states[0].param_shapes
+ avail_numel = fp32_flat_groups[0].numel() * world_size
+ # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+ # param, re-consolidating each param, while dealing with padding if any
+
+ # merge list of dicts, preserving order
+ param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+ if debug:
+ for i in range(world_size):
+ print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+ wanted_params = len(param_shapes)
+ wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+ # not asserting if there is a mismatch due to possible padding
+ avail_numel = fp32_flat_groups[0].numel() * world_size
+ print(f"Trainable params: Have {avail_numel} numels to process.")
+ print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+ # params
+ # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+ # out-of-core computing solution
+ offset = 0
+ total_numel = 0
+ total_params = 0
+ for name, shape in param_shapes.items():
+
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+ total_params += 1
+
+ partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+ if debug:
+ print(
+ f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+ )
+
+ # XXX: memory usage doubles here
+ state_dict[name] = torch.cat(
+ tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+ 0).narrow(0, 0, unpartitioned_numel).view(shape)
+ offset += partitioned_numel
+
+ offset *= world_size
+
+ # Sanity check
+ if offset != avail_numel:
+ raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+ print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+ state_dict = OrderedDict()
+
+ # buffers
+ buffers = zero_model_states[0].buffers
+ state_dict.update(buffers)
+ if debug:
+ print(f"added {len(buffers)} buffers")
+
+ _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+ _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+ # recover shared parameters
+ for pair in zero_model_states[0].shared_params:
+ if pair[1] in state_dict:
+ state_dict[pair[0]] = state_dict[pair[1]]
+
+ return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None):
+ """
+ Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+ ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+ via a model hub.
+
+ Args:
+ - ``checkpoint_dir``: path to the desired checkpoint folder
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+
+ Returns:
+ - pytorch ``state_dict``
+
+ Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+ you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+ the checkpoint.
+
+ A typical usage might be ::
+
+ from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+ # do the training and checkpoint saving
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+ model = model.cpu() # move to cpu
+ model.load_state_dict(state_dict)
+ # submit to model hub or save the model to share with others
+
+ In this example the ``model`` will no longer be usable in the deepspeed context of the same
+ application. i.e. you will need to re-initialize the deepspeed engine, since
+ ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+ If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+ """
+ if tag is None:
+ latest_path = os.path.join(checkpoint_dir, 'latest')
+ if os.path.isfile(latest_path):
+ with open(latest_path, 'r') as fd:
+ tag = fd.read().strip()
+ else:
+ raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+ ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+ if not os.path.isdir(ds_checkpoint_dir):
+ raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+ return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None):
+ """
+ Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+ loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+ Args:
+ - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+ - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+ """
+
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+ print(f"Saving fp32 state dict to {output_file}")
+ torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+ """
+ 1. Put the provided model to cpu
+ 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+ 3. Load it into the provided model
+
+ Args:
+ - ``model``: the model object to update
+ - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+ Returns:
+ - ``model`: modified model
+
+ Make sure you have plenty of CPU memory available before you call this function. If you don't
+ have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+ conveniently placed for you in the checkpoint folder.
+
+ A typical usage might be ::
+
+ from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+ model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+ # submit to model hub or save the model to share with others
+
+ Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+ of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+ ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+ """
+ logger.info(f"Extracting fp32 weights")
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+ logger.info(f"Overwriting model with fp32 weights")
+ model = model.cpu()
+ model.load_state_dict(state_dict, strict=False)
+
+ return model
+
+
+if __name__ == "__main__":
+
+ parser = argparse.ArgumentParser()
+ parser.add_argument("checkpoint_dir",
+ type=str,
+ help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+ parser.add_argument(
+ "output_file",
+ type=str,
+ help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+ parser.add_argument("-t",
+ "--tag",
+ type=str,
+ default=None,
+ help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+ parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+ args = parser.parse_args()
+
+ debug = args.debug
+
+ convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, args.output_file, tag=args.tag)
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/config.json b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..93e133af45036a778791b5679a8953a4f6a35a33
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/config.json
@@ -0,0 +1,70 @@
+{
+ "_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "architectures": [
+ "LlavaMistralForCausalLM"
+ ],
+ "attention_dropout": 0.0,
+ "bos_token_id": 1,
+ "eos_token_id": 2,
+ "freeze_mm_mlp_adapter": false,
+ "freeze_mm_vision_resampler": false,
+ "hidden_act": "silu",
+ "hidden_size": 4096,
+ "image_aspect_ratio": "anyres",
+ "image_crop_resolution": 224,
+ "image_grid_pinpoints": [
+ [
+ 336,
+ 672
+ ],
+ [
+ 672,
+ 336
+ ],
+ [
+ 672,
+ 672
+ ],
+ [
+ 1008,
+ 336
+ ],
+ [
+ 336,
+ 1008
+ ]
+ ],
+ "image_split_resolution": 224,
+ "initializer_range": 0.02,
+ "intermediate_size": 14336,
+ "max_position_embeddings": 32768,
+ "mm_hidden_size": 1024,
+ "mm_patch_merge_type": "spatial_unpad",
+ "mm_projector_lr": 2e-05,
+ "mm_projector_type": "mlp2x_gelu",
+ "mm_resampler_type": null,
+ "mm_use_im_patch_token": false,
+ "mm_use_im_start_end": false,
+ "mm_vision_select_feature": "patch",
+ "mm_vision_select_layer": -2,
+ "mm_vision_tower": "openai/clip-vit-large-patch14-336",
+ "mm_vision_tower_lr": 2e-06,
+ "model_type": "llava_mistral",
+ "num_attention_heads": 32,
+ "num_hidden_layers": 32,
+ "num_key_value_heads": 8,
+ "rms_norm_eps": 1e-05,
+ "rope_theta": 1000000.0,
+ "sliding_window": null,
+ "tie_word_embeddings": false,
+ "tokenizer_model_max_length": 4096,
+ "tokenizer_padding_side": "right",
+ "torch_dtype": "bfloat16",
+ "transformers_version": "4.37.2",
+ "tune_mm_mlp_adapter": false,
+ "tune_mm_vision_resampler": false,
+ "unfreeze_mm_vision_tower": true,
+ "use_cache": true,
+ "use_mm_proj": true,
+ "vocab_size": 32000
+}
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin
new file mode 100644
index 0000000000000000000000000000000000000000..4782a8ced628946bd00542cbef1cbaa4c56f8be6
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0102c05ff05f99863f06b141ed5812df27620d7c8dd7551f8bac60d6b2f9f0e
+size 41961648
diff --git a/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/trainer_state.json b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..bf6f9339414af492906d46a281548c34a3f5ebae
--- /dev/null
+++ b/CheckGuard Models/wholeimage/bank_no/llava-lora-mistral-r128a256-10BS-model/trainer_state.json
@@ -0,0 +1,3276 @@
+{
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.0,
+ "eval_steps": 500,
+ "global_step": 541,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "epoch": 0.0,
+ "learning_rate": 7.142857142857143e-07,
+ "loss": 0.6789,
+ "step": 1
+ },
+ {
+ "epoch": 0.0,
+ "learning_rate": 1.4285714285714286e-06,
+ "loss": 0.8481,
+ "step": 2
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.1428571428571427e-06,
+ "loss": 0.663,
+ "step": 3
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.8571428571428573e-06,
+ "loss": 0.679,
+ "step": 4
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 3.5714285714285718e-06,
+ "loss": 1.0166,
+ "step": 5
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 4.2857142857142855e-06,
+ "loss": 0.4693,
+ "step": 6
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 5e-06,
+ "loss": 0.4891,
+ "step": 7
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 5.7142857142857145e-06,
+ "loss": 0.5523,
+ "step": 8
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 6.4285714285714295e-06,
+ "loss": 0.2909,
+ "step": 9
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.1428571428571436e-06,
+ "loss": 0.2598,
+ "step": 10
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.857142857142858e-06,
+ "loss": 0.2532,
+ "step": 11
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.857142857142858e-06,
+ "loss": 0.4867,
+ "step": 12
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 8.571428571428571e-06,
+ "loss": 0.4145,
+ "step": 13
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 8.571428571428571e-06,
+ "loss": 0.3161,
+ "step": 14
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 9.285714285714288e-06,
+ "loss": 0.1836,
+ "step": 15
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1e-05,
+ "loss": 0.3355,
+ "step": 16
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.0714285714285714e-05,
+ "loss": 0.2286,
+ "step": 17
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.1428571428571429e-05,
+ "loss": 0.3594,
+ "step": 18
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.2142857142857142e-05,
+ "loss": 0.2981,
+ "step": 19
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.2857142857142859e-05,
+ "loss": 0.3021,
+ "step": 20
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.3571428571428574e-05,
+ "loss": 0.3866,
+ "step": 21
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.4285714285714287e-05,
+ "loss": 0.2409,
+ "step": 22
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.5000000000000002e-05,
+ "loss": 0.1397,
+ "step": 23
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.5714285714285715e-05,
+ "loss": 0.1416,
+ "step": 24
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.642857142857143e-05,
+ "loss": 0.1838,
+ "step": 25
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.7142857142857142e-05,
+ "loss": 0.1505,
+ "step": 26
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.785714285714286e-05,
+ "loss": 0.3278,
+ "step": 27
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.8571428571428575e-05,
+ "loss": 0.2567,
+ "step": 28
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.928571428571429e-05,
+ "loss": 0.1218,
+ "step": 29
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 2e-05,
+ "loss": 0.2288,
+ "step": 30
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9999812486015525e-05,
+ "loss": 0.1348,
+ "step": 31
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9999249951094388e-05,
+ "loss": 0.3734,
+ "step": 32
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.999831241633323e-05,
+ "loss": 0.3169,
+ "step": 33
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9996999916892222e-05,
+ "loss": 0.1066,
+ "step": 34
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9995312501993765e-05,
+ "loss": 0.4434,
+ "step": 35
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9993250234920638e-05,
+ "loss": 0.198,
+ "step": 36
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9990813193013625e-05,
+ "loss": 0.115,
+ "step": 37
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9988001467668613e-05,
+ "loss": 0.2676,
+ "step": 38
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9984815164333163e-05,
+ "loss": 0.2201,
+ "step": 39
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9981254402502568e-05,
+ "loss": 0.1945,
+ "step": 40
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.997731931571535e-05,
+ "loss": 0.1391,
+ "step": 41
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9973010051548274e-05,
+ "loss": 0.2697,
+ "step": 42
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9968326771610797e-05,
+ "loss": 0.1562,
+ "step": 43
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9963269651539018e-05,
+ "loss": 0.2204,
+ "step": 44
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9957838880989076e-05,
+ "loss": 0.2729,
+ "step": 45
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9952034663630064e-05,
+ "loss": 0.441,
+ "step": 46
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9952034663630064e-05,
+ "loss": 0.1401,
+ "step": 47
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9945857217136365e-05,
+ "loss": 0.3727,
+ "step": 48
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9939306773179498e-05,
+ "loss": 0.3269,
+ "step": 49
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9932383577419432e-05,
+ "loss": 0.0801,
+ "step": 50
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9925087889495374e-05,
+ "loss": 0.2772,
+ "step": 51
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9917419983016025e-05,
+ "loss": 0.2253,
+ "step": 52
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9909380145549325e-05,
+ "loss": 0.2318,
+ "step": 53
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9900968678611664e-05,
+ "loss": 0.1809,
+ "step": 54
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.989218589765658e-05,
+ "loss": 0.1155,
+ "step": 55
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9883032132062926e-05,
+ "loss": 0.2356,
+ "step": 56
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9873507725122505e-05,
+ "loss": 0.1194,
+ "step": 57
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9863613034027224e-05,
+ "loss": 0.3272,
+ "step": 58
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.985334842985567e-05,
+ "loss": 0.183,
+ "step": 59
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9842714297559212e-05,
+ "loss": 0.1217,
+ "step": 60
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9831711035947552e-05,
+ "loss": 0.1388,
+ "step": 61
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9820339057673773e-05,
+ "loss": 0.2112,
+ "step": 62
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9808598789218866e-05,
+ "loss": 0.0917,
+ "step": 63
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.979649067087574e-05,
+ "loss": 0.1585,
+ "step": 64
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9784015156732693e-05,
+ "loss": 0.1446,
+ "step": 65
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.97711727146564e-05,
+ "loss": 0.3511,
+ "step": 66
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9757963826274357e-05,
+ "loss": 0.1019,
+ "step": 67
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9744388986956824e-05,
+ "loss": 0.1165,
+ "step": 68
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.973044870579824e-05,
+ "loss": 0.2189,
+ "step": 69
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.971614350559814e-05,
+ "loss": 0.1254,
+ "step": 70
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.970147392284154e-05,
+ "loss": 0.0627,
+ "step": 71
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9686440507678827e-05,
+ "loss": 0.0952,
+ "step": 72
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.967104382390511e-05,
+ "loss": 0.1867,
+ "step": 73
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9655284448939094e-05,
+ "loss": 0.2003,
+ "step": 74
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9639162973801426e-05,
+ "loss": 0.1188,
+ "step": 75
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9622680003092503e-05,
+ "loss": 0.1111,
+ "step": 76
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.960583615496984e-05,
+ "loss": 0.1203,
+ "step": 77
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9588632061124837e-05,
+ "loss": 0.1599,
+ "step": 78
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9571068366759143e-05,
+ "loss": 0.209,
+ "step": 79
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9553145730560415e-05,
+ "loss": 0.2183,
+ "step": 80
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.953486482467764e-05,
+ "loss": 0.1351,
+ "step": 81
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.951622633469592e-05,
+ "loss": 0.128,
+ "step": 82
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9497230959610757e-05,
+ "loss": 0.2241,
+ "step": 83
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9477879411801843e-05,
+ "loss": 0.0991,
+ "step": 84
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9458172417006347e-05,
+ "loss": 0.1165,
+ "step": 85
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9438110714291697e-05,
+ "loss": 0.0792,
+ "step": 86
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9417695056027847e-05,
+ "loss": 0.121,
+ "step": 87
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9396926207859085e-05,
+ "loss": 0.2727,
+ "step": 88
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9375804948675308e-05,
+ "loss": 0.1947,
+ "step": 89
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.935433207058281e-05,
+ "loss": 0.2155,
+ "step": 90
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.933250837887457e-05,
+ "loss": 0.0525,
+ "step": 91
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9310334692000077e-05,
+ "loss": 0.2401,
+ "step": 92
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9287811841534598e-05,
+ "loss": 0.0743,
+ "step": 93
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9264940672148018e-05,
+ "loss": 0.1659,
+ "step": 94
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9241722041573166e-05,
+ "loss": 0.1184,
+ "step": 95
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9218156820573618e-05,
+ "loss": 0.1207,
+ "step": 96
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9194245892911077e-05,
+ "loss": 0.1292,
+ "step": 97
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.916999015531221e-05,
+ "loss": 0.2059,
+ "step": 98
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9145390517435013e-05,
+ "loss": 0.1682,
+ "step": 99
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9120447901834708e-05,
+ "loss": 0.1403,
+ "step": 100
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.9095163243929143e-05,
+ "loss": 0.1752,
+ "step": 101
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.906953749196371e-05,
+ "loss": 0.1616,
+ "step": 102
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.9043571606975776e-05,
+ "loss": 0.1127,
+ "step": 103
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.901726656275866e-05,
+ "loss": 0.2236,
+ "step": 104
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8990623345825084e-05,
+ "loss": 0.2308,
+ "step": 105
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8963642955370203e-05,
+ "loss": 0.1739,
+ "step": 106
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8936326403234125e-05,
+ "loss": 0.1762,
+ "step": 107
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.890867471386395e-05,
+ "loss": 0.1457,
+ "step": 108
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.888068892427538e-05,
+ "loss": 0.2768,
+ "step": 109
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8852370084013783e-05,
+ "loss": 0.1389,
+ "step": 110
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.882371925511488e-05,
+ "loss": 0.2747,
+ "step": 111
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.879473751206489e-05,
+ "loss": 0.0542,
+ "step": 112
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8765425941760237e-05,
+ "loss": 0.1414,
+ "step": 113
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8735785643466786e-05,
+ "loss": 0.2482,
+ "step": 114
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8705817728778626e-05,
+ "loss": 0.1602,
+ "step": 115
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.867552332157637e-05,
+ "loss": 0.1342,
+ "step": 116
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8644903557985027e-05,
+ "loss": 0.077,
+ "step": 117
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8613959586331364e-05,
+ "loss": 0.0818,
+ "step": 118
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8582692567100866e-05,
+ "loss": 0.1443,
+ "step": 119
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.855110367289421e-05,
+ "loss": 0.1148,
+ "step": 120
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.851919408838327e-05,
+ "loss": 0.1661,
+ "step": 121
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8486965010266726e-05,
+ "loss": 0.1676,
+ "step": 122
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.845441764722514e-05,
+ "loss": 0.1288,
+ "step": 123
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.842155321987566e-05,
+ "loss": 0.0725,
+ "step": 124
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8388372960726228e-05,
+ "loss": 0.1258,
+ "step": 125
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8354878114129368e-05,
+ "loss": 0.068,
+ "step": 126
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8321069936235503e-05,
+ "loss": 0.1698,
+ "step": 127
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8286949694945864e-05,
+ "loss": 0.2038,
+ "step": 128
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8252518669864935e-05,
+ "loss": 0.0274,
+ "step": 129
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.821777815225245e-05,
+ "loss": 0.0564,
+ "step": 130
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8182729444974993e-05,
+ "loss": 0.1182,
+ "step": 131
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8147373862457107e-05,
+ "loss": 0.3175,
+ "step": 132
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8111712730632024e-05,
+ "loss": 0.1017,
+ "step": 133
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.807574738689193e-05,
+ "loss": 0.3348,
+ "step": 134
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8039479180037803e-05,
+ "loss": 0.3129,
+ "step": 135
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.800290947022884e-05,
+ "loss": 0.1095,
+ "step": 136
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7966039628931447e-05,
+ "loss": 0.1922,
+ "step": 137
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7928871038867785e-05,
+ "loss": 0.1022,
+ "step": 138
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.789140509396394e-05,
+ "loss": 0.2318,
+ "step": 139
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7853643199297632e-05,
+ "loss": 0.2374,
+ "step": 140
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7815586771045535e-05,
+ "loss": 0.1194,
+ "step": 141
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.777723723643014e-05,
+ "loss": 0.1914,
+ "step": 142
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.773859603366626e-05,
+ "loss": 0.0431,
+ "step": 143
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.769966461190707e-05,
+ "loss": 0.081,
+ "step": 144
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.766044443118978e-05,
+ "loss": 0.2162,
+ "step": 145
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.762093696238086e-05,
+ "loss": 0.1151,
+ "step": 146
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7581143687120877e-05,
+ "loss": 0.184,
+ "step": 147
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7541066097768965e-05,
+ "loss": 0.1963,
+ "step": 148
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.750070569734681e-05,
+ "loss": 0.1318,
+ "step": 149
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7460063999482314e-05,
+ "loss": 0.1163,
+ "step": 150
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7419142528352815e-05,
+ "loss": 0.1013,
+ "step": 151
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.737794281862794e-05,
+ "loss": 0.0957,
+ "step": 152
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7336466415412028e-05,
+ "loss": 0.2023,
+ "step": 153
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.729471487418621e-05,
+ "loss": 0.1398,
+ "step": 154
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7252689760750053e-05,
+ "loss": 0.1238,
+ "step": 155
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.721039265116285e-05,
+ "loss": 0.2201,
+ "step": 156
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7167825131684516e-05,
+ "loss": 0.0698,
+ "step": 157
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7124988798716084e-05,
+ "loss": 0.0312,
+ "step": 158
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7081885258739846e-05,
+ "loss": 0.1443,
+ "step": 159
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.7038516128259118e-05,
+ "loss": 0.1349,
+ "step": 160
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6994883033737582e-05,
+ "loss": 0.0751,
+ "step": 161
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.695098761153832e-05,
+ "loss": 0.0543,
+ "step": 162
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6906831507862446e-05,
+ "loss": 0.0533,
+ "step": 163
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.686241637868734e-05,
+ "loss": 0.1328,
+ "step": 164
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6817743889704564e-05,
+ "loss": 0.3057,
+ "step": 165
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6772815716257414e-05,
+ "loss": 0.1642,
+ "step": 166
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.672763354327804e-05,
+ "loss": 0.1479,
+ "step": 167
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6682199065224307e-05,
+ "loss": 0.1163,
+ "step": 168
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6636513986016215e-05,
+ "loss": 0.0395,
+ "step": 169
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6590580018972012e-05,
+ "loss": 0.0456,
+ "step": 170
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6544398886743934e-05,
+ "loss": 0.2018,
+ "step": 171
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.64979723212536e-05,
+ "loss": 0.1655,
+ "step": 172
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6451302063627067e-05,
+ "loss": 0.1805,
+ "step": 173
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6404389864129533e-05,
+ "loss": 0.2445,
+ "step": 174
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6357237482099682e-05,
+ "loss": 0.134,
+ "step": 175
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6309846685883726e-05,
+ "loss": 0.0976,
+ "step": 176
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6262219252769065e-05,
+ "loss": 0.0984,
+ "step": 177
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.621435696891765e-05,
+ "loss": 0.0495,
+ "step": 178
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6166261629298996e-05,
+ "loss": 0.1005,
+ "step": 179
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6117935037622848e-05,
+ "loss": 0.1399,
+ "step": 180
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.606937900627157e-05,
+ "loss": 0.2105,
+ "step": 181
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.6020595356232137e-05,
+ "loss": 0.142,
+ "step": 182
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5971585917027864e-05,
+ "loss": 0.0791,
+ "step": 183
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5922352526649803e-05,
+ "loss": 0.2,
+ "step": 184
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.587289703148779e-05,
+ "loss": 0.1317,
+ "step": 185
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5823221286261217e-05,
+ "loss": 0.1656,
+ "step": 186
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5773327153949465e-05,
+ "loss": 0.3358,
+ "step": 187
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.572321650572205e-05,
+ "loss": 0.2216,
+ "step": 188
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.567289122086843e-05,
+ "loss": 0.0937,
+ "step": 189
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5622353186727542e-05,
+ "loss": 0.0995,
+ "step": 190
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.557160429861702e-05,
+ "loss": 0.2324,
+ "step": 191
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5520646459762102e-05,
+ "loss": 0.2847,
+ "step": 192
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5469481581224274e-05,
+ "loss": 0.1242,
+ "step": 193
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5418111581829575e-05,
+ "loss": 0.1771,
+ "step": 194
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.536653838809667e-05,
+ "loss": 0.2115,
+ "step": 195
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.531476393416456e-05,
+ "loss": 0.074,
+ "step": 196
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5262790161720082e-05,
+ "loss": 0.0893,
+ "step": 197
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5210619019925066e-05,
+ "loss": 0.0644,
+ "step": 198
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5158252465343242e-05,
+ "loss": 0.2146,
+ "step": 199
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5105692461866874e-05,
+ "loss": 0.2579,
+ "step": 200
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.50529409806431e-05,
+ "loss": 0.0806,
+ "step": 201
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5000000000000002e-05,
+ "loss": 0.0806,
+ "step": 202
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4946871505372426e-05,
+ "loss": 0.132,
+ "step": 203
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4893557489227518e-05,
+ "loss": 0.1438,
+ "step": 204
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4840059950989992e-05,
+ "loss": 0.1703,
+ "step": 205
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.478638089696716e-05,
+ "loss": 0.0903,
+ "step": 206
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4732522340273686e-05,
+ "loss": 0.1515,
+ "step": 207
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.467848630075608e-05,
+ "loss": 0.2156,
+ "step": 208
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4624274804916958e-05,
+ "loss": 0.0783,
+ "step": 209
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.456988988583904e-05,
+ "loss": 0.1432,
+ "step": 210
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4515333583108896e-05,
+ "loss": 0.1716,
+ "step": 211
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4460607942740468e-05,
+ "loss": 0.2328,
+ "step": 212
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4405715017098333e-05,
+ "loss": 0.1317,
+ "step": 213
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4350656864820733e-05,
+ "loss": 0.097,
+ "step": 214
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4295435550742372e-05,
+ "loss": 0.1547,
+ "step": 215
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4240053145816968e-05,
+ "loss": 0.0737,
+ "step": 216
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4184511727039612e-05,
+ "loss": 0.0926,
+ "step": 217
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4128813377368851e-05,
+ "loss": 0.0824,
+ "step": 218
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4072960185648576e-05,
+ "loss": 0.1236,
+ "step": 219
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.4016954246529697e-05,
+ "loss": 0.157,
+ "step": 220
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.396079766039157e-05,
+ "loss": 0.1241,
+ "step": 221
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3904492533263243e-05,
+ "loss": 0.1243,
+ "step": 222
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3848040976744459e-05,
+ "loss": 0.1429,
+ "step": 223
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3791445107926478e-05,
+ "loss": 0.0321,
+ "step": 224
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3734707049312674e-05,
+ "loss": 0.0398,
+ "step": 225
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3677828928738934e-05,
+ "loss": 0.2625,
+ "step": 226
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3620812879293864e-05,
+ "loss": 0.0926,
+ "step": 227
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3563661039238785e-05,
+ "loss": 0.06,
+ "step": 228
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3506375551927546e-05,
+ "loss": 0.2397,
+ "step": 229
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3448958565726144e-05,
+ "loss": 0.157,
+ "step": 230
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3391412233932148e-05,
+ "loss": 0.1105,
+ "step": 231
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3333738714693958e-05,
+ "loss": 0.0877,
+ "step": 232
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3275940170929845e-05,
+ "loss": 0.1821,
+ "step": 233
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3218018770246858e-05,
+ "loss": 0.0166,
+ "step": 234
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3159976684859528e-05,
+ "loss": 0.118,
+ "step": 235
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3101816091508389e-05,
+ "loss": 0.2289,
+ "step": 236
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3043539171378362e-05,
+ "loss": 0.0518,
+ "step": 237
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2985148110016947e-05,
+ "loss": 0.1012,
+ "step": 238
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.292664509725226e-05,
+ "loss": 0.2009,
+ "step": 239
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2868032327110904e-05,
+ "loss": 0.252,
+ "step": 240
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2809311997735697e-05,
+ "loss": 0.2044,
+ "step": 241
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2750486311303218e-05,
+ "loss": 0.1908,
+ "step": 242
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2691557473941246e-05,
+ "loss": 0.3064,
+ "step": 243
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2632527695645993e-05,
+ "loss": 0.091,
+ "step": 244
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.257339919019925e-05,
+ "loss": 0.0606,
+ "step": 245
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2514174175085346e-05,
+ "loss": 0.147,
+ "step": 246
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2454854871407993e-05,
+ "loss": 0.2029,
+ "step": 247
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.239544350380699e-05,
+ "loss": 0.0851,
+ "step": 248
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2335942300374788e-05,
+ "loss": 0.0904,
+ "step": 249
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2276353492572937e-05,
+ "loss": 0.0721,
+ "step": 250
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2216679315148388e-05,
+ "loss": 0.1488,
+ "step": 251
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2156922006049703e-05,
+ "loss": 0.1927,
+ "step": 252
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2097083806343104e-05,
+ "loss": 0.029,
+ "step": 253
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2037166960128443e-05,
+ "loss": 0.0301,
+ "step": 254
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1977173714455034e-05,
+ "loss": 0.1231,
+ "step": 255
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1917106319237386e-05,
+ "loss": 0.0348,
+ "step": 256
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1856967027170818e-05,
+ "loss": 0.0869,
+ "step": 257
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1796758093646989e-05,
+ "loss": 0.1164,
+ "step": 258
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1736481776669307e-05,
+ "loss": 0.0388,
+ "step": 259
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1676140336768236e-05,
+ "loss": 0.0433,
+ "step": 260
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.161573603691655e-05,
+ "loss": 0.1996,
+ "step": 261
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1555271142444433e-05,
+ "loss": 0.2182,
+ "step": 262
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1494747920954545e-05,
+ "loss": 0.0509,
+ "step": 263
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1434168642236964e-05,
+ "loss": 0.1078,
+ "step": 264
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1373535578184083e-05,
+ "loss": 0.0412,
+ "step": 265
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1312851002705383e-05,
+ "loss": 0.2425,
+ "step": 266
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1252117191642175e-05,
+ "loss": 0.1119,
+ "step": 267
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1191336422682237e-05,
+ "loss": 0.0455,
+ "step": 268
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1130510975274408e-05,
+ "loss": 0.2613,
+ "step": 269
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1069643130543084e-05,
+ "loss": 0.0651,
+ "step": 270
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1008735171202685e-05,
+ "loss": 0.1155,
+ "step": 271
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.0947789381472035e-05,
+ "loss": 0.0661,
+ "step": 272
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.0886808046988716e-05,
+ "loss": 0.0881,
+ "step": 273
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0825793454723325e-05,
+ "loss": 0.1123,
+ "step": 274
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0764747892893724e-05,
+ "loss": 0.14,
+ "step": 275
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0703673650879219e-05,
+ "loss": 0.0889,
+ "step": 276
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0642573019134703e-05,
+ "loss": 0.1333,
+ "step": 277
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0581448289104759e-05,
+ "loss": 0.0608,
+ "step": 278
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0520301753137725e-05,
+ "loss": 0.2882,
+ "step": 279
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.045913570439972e-05,
+ "loss": 0.0661,
+ "step": 280
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0397952436788643e-05,
+ "loss": 0.107,
+ "step": 281
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0336754244848156e-05,
+ "loss": 0.0499,
+ "step": 282
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0275543423681622e-05,
+ "loss": 0.237,
+ "step": 283
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0214322268866033e-05,
+ "loss": 0.0301,
+ "step": 284
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0153093076365923e-05,
+ "loss": 0.0904,
+ "step": 285
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0091858142447266e-05,
+ "loss": 0.0165,
+ "step": 286
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0030619763591348e-05,
+ "loss": 0.0791,
+ "step": 287
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.969380236408656e-06,
+ "loss": 0.1997,
+ "step": 288
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.908141857552737e-06,
+ "loss": 0.0155,
+ "step": 289
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.846906923634079e-06,
+ "loss": 0.0457,
+ "step": 290
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.785677731133972e-06,
+ "loss": 0.0203,
+ "step": 291
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.724456576318383e-06,
+ "loss": 0.2384,
+ "step": 292
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.663245755151847e-06,
+ "loss": 0.1459,
+ "step": 293
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.602047563211359e-06,
+ "loss": 0.2249,
+ "step": 294
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.540864295600282e-06,
+ "loss": 0.037,
+ "step": 295
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.479698246862277e-06,
+ "loss": 0.145,
+ "step": 296
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.418551710895243e-06,
+ "loss": 0.1501,
+ "step": 297
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.3574269808653e-06,
+ "loss": 0.0727,
+ "step": 298
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.296326349120786e-06,
+ "loss": 0.0992,
+ "step": 299
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.23525210710628e-06,
+ "loss": 0.2516,
+ "step": 300
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.174206545276678e-06,
+ "loss": 0.0628,
+ "step": 301
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.113191953011287e-06,
+ "loss": 0.132,
+ "step": 302
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.052210618527966e-06,
+ "loss": 0.0908,
+ "step": 303
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.991264828797319e-06,
+ "loss": 0.1432,
+ "step": 304
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.93035686945692e-06,
+ "loss": 0.0493,
+ "step": 305
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.869489024725595e-06,
+ "loss": 0.0578,
+ "step": 306
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.808663577317765e-06,
+ "loss": 0.0909,
+ "step": 307
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.747882808357828e-06,
+ "loss": 0.0646,
+ "step": 308
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.687148997294622e-06,
+ "loss": 0.1308,
+ "step": 309
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.626464421815919e-06,
+ "loss": 0.0729,
+ "step": 310
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.565831357763039e-06,
+ "loss": 0.1871,
+ "step": 311
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.505252079045459e-06,
+ "loss": 0.1577,
+ "step": 312
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.444728857555572e-06,
+ "loss": 0.1844,
+ "step": 313
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.384263963083453e-06,
+ "loss": 0.1673,
+ "step": 314
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.323859663231768e-06,
+ "loss": 0.1898,
+ "step": 315
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.263518223330698e-06,
+ "loss": 0.1106,
+ "step": 316
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.203241906353014e-06,
+ "loss": 0.0476,
+ "step": 317
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.143032972829184e-06,
+ "loss": 0.1432,
+ "step": 318
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.082893680762619e-06,
+ "loss": 0.0249,
+ "step": 319
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.022826285544967e-06,
+ "loss": 0.0762,
+ "step": 320
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.962833039871562e-06,
+ "loss": 0.1468,
+ "step": 321
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.902916193656898e-06,
+ "loss": 0.0272,
+ "step": 322
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.843077993950302e-06,
+ "loss": 0.0495,
+ "step": 323
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.783320684851613e-06,
+ "loss": 0.1958,
+ "step": 324
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.72364650742707e-06,
+ "loss": 0.0869,
+ "step": 325
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.664057699625215e-06,
+ "loss": 0.2957,
+ "step": 326
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.604556496193015e-06,
+ "loss": 0.0833,
+ "step": 327
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.545145128592009e-06,
+ "loss": 0.0978,
+ "step": 328
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.485825824914658e-06,
+ "loss": 0.1941,
+ "step": 329
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.426600809800753e-06,
+ "loss": 0.0384,
+ "step": 330
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.367472304354011e-06,
+ "loss": 0.0872,
+ "step": 331
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.308442526058757e-06,
+ "loss": 0.1051,
+ "step": 332
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.249513688696786e-06,
+ "loss": 0.0918,
+ "step": 333
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.190688002264308e-06,
+ "loss": 0.2169,
+ "step": 334
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.131967672889101e-06,
+ "loss": 0.1647,
+ "step": 335
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.073354902747742e-06,
+ "loss": 0.0585,
+ "step": 336
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.014851889983058e-06,
+ "loss": 0.1743,
+ "step": 337
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 6.956460828621641e-06,
+ "loss": 0.3001,
+ "step": 338
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.898183908491617e-06,
+ "loss": 0.0977,
+ "step": 339
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.840023315140476e-06,
+ "loss": 0.0549,
+ "step": 340
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.781981229753145e-06,
+ "loss": 0.0738,
+ "step": 341
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.7240598290701585e-06,
+ "loss": 0.027,
+ "step": 342
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.666261285306048e-06,
+ "loss": 0.0647,
+ "step": 343
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.608587766067853e-06,
+ "loss": 0.0531,
+ "step": 344
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.551041434273862e-06,
+ "loss": 0.0582,
+ "step": 345
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.4936244480724575e-06,
+ "loss": 0.2357,
+ "step": 346
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.4363389607612204e-06,
+ "loss": 0.0614,
+ "step": 347
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.379187120706138e-06,
+ "loss": 0.1516,
+ "step": 348
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.322171071261071e-06,
+ "loss": 0.2906,
+ "step": 349
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.265292950687329e-06,
+ "loss": 0.0402,
+ "step": 350
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.208554892073528e-06,
+ "loss": 0.0895,
+ "step": 351
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.151959023255545e-06,
+ "loss": 0.109,
+ "step": 352
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.095507466736763e-06,
+ "loss": 0.1338,
+ "step": 353
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.039202339608432e-06,
+ "loss": 0.0541,
+ "step": 354
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.983045753470308e-06,
+ "loss": 0.0614,
+ "step": 355
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.927039814351426e-06,
+ "loss": 0.2844,
+ "step": 356
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.871186622631155e-06,
+ "loss": 0.1412,
+ "step": 357
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.815488272960388e-06,
+ "loss": 0.0575,
+ "step": 358
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.759946854183036e-06,
+ "loss": 0.1047,
+ "step": 359
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.704564449257635e-06,
+ "loss": 0.2065,
+ "step": 360
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.649343135179271e-06,
+ "loss": 0.0995,
+ "step": 361
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.59428498290167e-06,
+ "loss": 0.1517,
+ "step": 362
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.539392057259536e-06,
+ "loss": 0.1122,
+ "step": 363
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.484666416891109e-06,
+ "loss": 0.0992,
+ "step": 364
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.430110114160965e-06,
+ "loss": 0.1303,
+ "step": 365
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.375725195083046e-06,
+ "loss": 0.1192,
+ "step": 366
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.321513699243924e-06,
+ "loss": 0.0991,
+ "step": 367
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.267477659726319e-06,
+ "loss": 0.077,
+ "step": 368
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.213619103032845e-06,
+ "loss": 0.1052,
+ "step": 369
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.159940049010015e-06,
+ "loss": 0.2359,
+ "step": 370
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.106442510772489e-06,
+ "loss": 0.0501,
+ "step": 371
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.053128494627578e-06,
+ "loss": 0.0803,
+ "step": 372
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.000000000000003e-06,
+ "loss": 0.2073,
+ "step": 373
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.947059019356904e-06,
+ "loss": 0.0479,
+ "step": 374
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.89430753813313e-06,
+ "loss": 0.125,
+ "step": 375
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.8417475346567635e-06,
+ "loss": 0.0715,
+ "step": 376
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.78938098007494e-06,
+ "loss": 0.0242,
+ "step": 377
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.737209838279923e-06,
+ "loss": 0.1242,
+ "step": 378
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.685236065835443e-06,
+ "loss": 0.1771,
+ "step": 379
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.633461611903336e-06,
+ "loss": 0.1037,
+ "step": 380
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.581888418170429e-06,
+ "loss": 0.0733,
+ "step": 381
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.530518418775734e-06,
+ "loss": 0.0565,
+ "step": 382
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.479353540237903e-06,
+ "loss": 0.1092,
+ "step": 383
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.4283957013829845e-06,
+ "loss": 0.0371,
+ "step": 384
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.3776468132724605e-06,
+ "loss": 0.1105,
+ "step": 385
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.327108779131573e-06,
+ "loss": 0.1856,
+ "step": 386
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.276783494277954e-06,
+ "loss": 0.1237,
+ "step": 387
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.226672846050538e-06,
+ "loss": 0.2521,
+ "step": 388
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.176778713738787e-06,
+ "loss": 0.0565,
+ "step": 389
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.127102968512214e-06,
+ "loss": 0.0518,
+ "step": 390
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.077647473350201e-06,
+ "loss": 0.0735,
+ "step": 391
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.028414082972141e-06,
+ "loss": 0.0786,
+ "step": 392
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.9794046437678705e-06,
+ "loss": 0.025,
+ "step": 393
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.930620993728434e-06,
+ "loss": 0.2235,
+ "step": 394
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.882064962377154e-06,
+ "loss": 0.1307,
+ "step": 395
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.83373837070101e-06,
+ "loss": 0.0224,
+ "step": 396
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.7856430310823546e-06,
+ "loss": 0.1109,
+ "step": 397
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.737780747230941e-06,
+ "loss": 0.0624,
+ "step": 398
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6901533141162804e-06,
+ "loss": 0.055,
+ "step": 399
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6427625179003223e-06,
+ "loss": 0.2079,
+ "step": 400
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.595610135870472e-06,
+ "loss": 0.2215,
+ "step": 401
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.548697936372937e-06,
+ "loss": 0.1016,
+ "step": 402
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.5020276787464058e-06,
+ "loss": 0.1229,
+ "step": 403
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.455601113256073e-06,
+ "loss": 0.0759,
+ "step": 404
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.4094199810279926e-06,
+ "loss": 0.1667,
+ "step": 405
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.3634860139837877e-06,
+ "loss": 0.048,
+ "step": 406
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.317800934775696e-06,
+ "loss": 0.0543,
+ "step": 407
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.2723664567219627e-06,
+ "loss": 0.1656,
+ "step": 408
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.2271842837425917e-06,
+ "loss": 0.0409,
+ "step": 409
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.1822561102954373e-06,
+ "loss": 0.1173,
+ "step": 410
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.1375836213126653e-06,
+ "loss": 0.0964,
+ "step": 411
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0931684921375572e-06,
+ "loss": 0.0432,
+ "step": 412
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0490123884616795e-06,
+ "loss": 0.1451,
+ "step": 413
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 3.0051169662624224e-06,
+ "loss": 0.1226,
+ "step": 414
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.9614838717408866e-06,
+ "loss": 0.096,
+ "step": 415
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.918114741260156e-06,
+ "loss": 0.1152,
+ "step": 416
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.8750112012839215e-06,
+ "loss": 0.0575,
+ "step": 417
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.8321748683154893e-06,
+ "loss": 0.097,
+ "step": 418
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.7896073488371535e-06,
+ "loss": 0.0513,
+ "step": 419
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.7473102392499517e-06,
+ "loss": 0.0566,
+ "step": 420
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.7052851258137936e-06,
+ "loss": 0.0193,
+ "step": 421
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.663533584587974e-06,
+ "loss": 0.1507,
+ "step": 422
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.622057181372063e-06,
+ "loss": 0.0208,
+ "step": 423
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.580857471647186e-06,
+ "loss": 0.0893,
+ "step": 424
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.539936000517689e-06,
+ "loss": 0.0988,
+ "step": 425
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4992943026531935e-06,
+ "loss": 0.0368,
+ "step": 426
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4589339022310386e-06,
+ "loss": 0.0911,
+ "step": 427
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4188563128791255e-06,
+ "loss": 0.1093,
+ "step": 428
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.379063037619146e-06,
+ "loss": 0.0717,
+ "step": 429
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.339555568810221e-06,
+ "loss": 0.1486,
+ "step": 430
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.300335388092929e-06,
+ "loss": 0.1174,
+ "step": 431
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.261403966333742e-06,
+ "loss": 0.2022,
+ "step": 432
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.2227627635698624e-06,
+ "loss": 0.0376,
+ "step": 433
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.1844132289544684e-06,
+ "loss": 0.3022,
+ "step": 434
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.1463568007023706e-06,
+ "loss": 0.0121,
+ "step": 435
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.1085949060360654e-06,
+ "loss": 0.1441,
+ "step": 436
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0711289611322204e-06,
+ "loss": 0.0457,
+ "step": 437
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0339603710685574e-06,
+ "loss": 0.0324,
+ "step": 438
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.9970905297711606e-06,
+ "loss": 0.045,
+ "step": 439
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.9605208199621993e-06,
+ "loss": 0.0644,
+ "step": 440
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.924252613108073e-06,
+ "loss": 0.0743,
+ "step": 441
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8882872693679787e-06,
+ "loss": 0.054,
+ "step": 442
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8526261375428955e-06,
+ "loss": 0.1679,
+ "step": 443
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8172705550250093e-06,
+ "loss": 0.0666,
+ "step": 444
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7822218477475496e-06,
+ "loss": 0.2,
+ "step": 445
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7474813301350668e-06,
+ "loss": 0.1191,
+ "step": 446
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.7130503050541368e-06,
+ "loss": 0.1166,
+ "step": 447
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6789300637645e-06,
+ "loss": 0.0089,
+ "step": 448
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6451218858706374e-06,
+ "loss": 0.0848,
+ "step": 449
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6116270392737753e-06,
+ "loss": 0.1263,
+ "step": 450
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.578446780124344e-06,
+ "loss": 0.1338,
+ "step": 451
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.5455823527748626e-06,
+ "loss": 0.0566,
+ "step": 452
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.5130349897332764e-06,
+ "loss": 0.0618,
+ "step": 453
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4808059116167306e-06,
+ "loss": 0.0259,
+ "step": 454
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4488963271057943e-06,
+ "loss": 0.1682,
+ "step": 455
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4173074328991376e-06,
+ "loss": 0.0967,
+ "step": 456
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.3860404136686411e-06,
+ "loss": 0.0799,
+ "step": 457
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.355096442014977e-06,
+ "loss": 0.1426,
+ "step": 458
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3244766784236307e-06,
+ "loss": 0.1401,
+ "step": 459
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.294182271221377e-06,
+ "loss": 0.0526,
+ "step": 460
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.2642143565332154e-06,
+ "loss": 0.1516,
+ "step": 461
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.2345740582397647e-06,
+ "loss": 0.0326,
+ "step": 462
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.2052624879351105e-06,
+ "loss": 0.0517,
+ "step": 463
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.176280744885121e-06,
+ "loss": 0.094,
+ "step": 464
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1476299159862204e-06,
+ "loss": 0.0684,
+ "step": 465
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1193110757246251e-06,
+ "loss": 0.0845,
+ "step": 466
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.09132528613605e-06,
+ "loss": 0.1105,
+ "step": 467
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.0636735967658785e-06,
+ "loss": 0.0947,
+ "step": 468
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.0363570446297999e-06,
+ "loss": 0.0685,
+ "step": 469
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.0093766541749206e-06,
+ "loss": 0.0902,
+ "step": 470
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.827334372413444e-07,
+ "loss": 0.0257,
+ "step": 471
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.564283930242258e-07,
+ "loss": 0.1048,
+ "step": 472
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.304625080362939e-07,
+ "loss": 0.1365,
+ "step": 473
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 9.048367560708604e-07,
+ "loss": 0.2323,
+ "step": 474
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.79552098165296e-07,
+ "loss": 0.0435,
+ "step": 475
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.546094825649909e-07,
+ "loss": 0.0644,
+ "step": 476
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.300098446877925e-07,
+ "loss": 0.0884,
+ "step": 477
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.057541070889229e-07,
+ "loss": 0.1381,
+ "step": 478
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.818431794263837e-07,
+ "loss": 0.0472,
+ "step": 479
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.582779584268374e-07,
+ "loss": 0.0606,
+ "step": 480
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.350593278519824e-07,
+ "loss": 0.0325,
+ "step": 481
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.121881584654056e-07,
+ "loss": 0.0391,
+ "step": 482
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 6.896653079999249e-07,
+ "loss": 0.0965,
+ "step": 483
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 6.67491621125429e-07,
+ "loss": 0.0288,
+ "step": 484
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.45667929417193e-07,
+ "loss": 0.0608,
+ "step": 485
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.241950513246931e-07,
+ "loss": 0.0619,
+ "step": 486
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.030737921409169e-07,
+ "loss": 0.2691,
+ "step": 487
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 5.823049439721562e-07,
+ "loss": 0.1071,
+ "step": 488
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 5.618892857083069e-07,
+ "loss": 0.1501,
+ "step": 489
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.418275829936537e-07,
+ "loss": 0.0807,
+ "step": 490
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.221205881981594e-07,
+ "loss": 0.0666,
+ "step": 491
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.027690403892461e-07,
+ "loss": 0.0993,
+ "step": 492
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 4.837736653040825e-07,
+ "loss": 0.2467,
+ "step": 493
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 4.6513517532236096e-07,
+ "loss": 0.0563,
+ "step": 494
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 4.468542694395861e-07,
+ "loss": 0.0792,
+ "step": 495
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.2893163324085886e-07,
+ "loss": 0.0648,
+ "step": 496
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.113679388751635e-07,
+ "loss": 0.3011,
+ "step": 497
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.941638450301644e-07,
+ "loss": 0.221,
+ "step": 498
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.773199969074959e-07,
+ "loss": 0.0961,
+ "step": 499
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.608370261985761e-07,
+ "loss": 0.0816,
+ "step": 500
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.4471555106090573e-07,
+ "loss": 0.0565,
+ "step": 501
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.2895617609489337e-07,
+ "loss": 0.0703,
+ "step": 502
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.135594923211771e-07,
+ "loss": 0.0622,
+ "step": 503
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 2.9852607715846194e-07,
+ "loss": 0.1138,
+ "step": 504
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 2.838564944018618e-07,
+ "loss": 0.2741,
+ "step": 505
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.6955129420176193e-07,
+ "loss": 0.06,
+ "step": 506
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.556110130431788e-07,
+ "loss": 0.0322,
+ "step": 507
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.420361737256438e-07,
+ "loss": 0.1867,
+ "step": 508
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.2882728534360131e-07,
+ "loss": 0.2815,
+ "step": 509
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.159848432673084e-07,
+ "loss": 0.133,
+ "step": 510
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.035093291242607e-07,
+ "loss": 0.134,
+ "step": 511
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.914012107811336e-07,
+ "loss": 0.1156,
+ "step": 512
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.7966094232622856e-07,
+ "loss": 0.3407,
+ "step": 513
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.6828896405244988e-07,
+ "loss": 0.0641,
+ "step": 514
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.572857024407881e-07,
+ "loss": 0.0459,
+ "step": 515
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.466515701443294e-07,
+ "loss": 0.1403,
+ "step": 516
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.3638696597277678e-07,
+ "loss": 0.0836,
+ "step": 517
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.264922748774955e-07,
+ "loss": 0.1507,
+ "step": 518
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.1696786793707782e-07,
+ "loss": 0.1091,
+ "step": 519
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.0781410234342093e-07,
+ "loss": 0.0432,
+ "step": 520
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 9.90313213883376e-08,
+ "loss": 0.0166,
+ "step": 521
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 9.061985445067756e-08,
+ "loss": 0.1675,
+ "step": 522
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 8.258001698397744e-08,
+ "loss": 0.0462,
+ "step": 523
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 7.491211050462798e-08,
+ "loss": 0.0219,
+ "step": 524
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 6.761642258056977e-08,
+ "loss": 0.1261,
+ "step": 525
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 6.069322682050516e-08,
+ "loss": 0.1249,
+ "step": 526
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 5.414278286363761e-08,
+ "loss": 0.0674,
+ "step": 527
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 4.796533636993728e-08,
+ "loss": 0.0171,
+ "step": 528
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 4.216111901092501e-08,
+ "loss": 0.0653,
+ "step": 529
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 3.6730348460986e-08,
+ "loss": 0.0292,
+ "step": 530
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 3.167322838920406e-08,
+ "loss": 0.1442,
+ "step": 531
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 2.6989948451726643e-08,
+ "loss": 0.0773,
+ "step": 532
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 2.2680684284650532e-08,
+ "loss": 0.0428,
+ "step": 533
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.8745597497433765e-08,
+ "loss": 0.2392,
+ "step": 534
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.518483566683826e-08,
+ "loss": 0.1413,
+ "step": 535
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.1998532331389812e-08,
+ "loss": 0.0554,
+ "step": 536
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 9.186806986376528e-09,
+ "loss": 0.1174,
+ "step": 537
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 6.749765079363535e-09,
+ "loss": 0.048,
+ "step": 538
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 4.687498006236135e-09,
+ "loss": 0.1818,
+ "step": 539
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 3.0000831077803273e-09,
+ "loss": 0.083,
+ "step": 540
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 1.6875836667729073e-09,
+ "loss": 0.0186,
+ "step": 541
+ },
+ {
+ "epoch": 1.0,
+ "step": 541,
+ "total_flos": 1291092221952.0,
+ "train_loss": 0.14259492732281495,
+ "train_runtime": 4024.6248,
+ "train_samples_per_second": 1.342,
+ "train_steps_per_second": 0.134
+ }
+ ],
+ "logging_steps": 1.0,
+ "max_steps": 541,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 1,
+ "save_steps": 500,
+ "total_flos": 1291092221952.0,
+ "train_batch_size": 10,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/CheckGuard Models/wholeimage/check_no/finetune_lora_llava_mistral.sh b/CheckGuard Models/wholeimage/check_no/finetune_lora_llava_mistral.sh
new file mode 100644
index 0000000000000000000000000000000000000000..5ca5fe49f959f0b968b4d2e38d75e79f56d60a65
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/finetune_lora_llava_mistral.sh
@@ -0,0 +1,43 @@
+#!/bin/bash
+# Use first parameter as GPU IDs, default to "0,1,2,3" if not provided
+GPU_IDS=${1:-0,1,2,3}
+
+
+CUDA_VISIBLE_DEVICES=0,1,2,3 deepspeed --include localhost:"$GPU_IDS" --master_port 29604\
+ llava/train/train_mem.py \
+ --lora_enable True --lora_r 128 --lora_alpha 256 --mm_projector_lr 2e-5 \
+ --deepspeed ./scripts/zero3.json \
+ --model_name_or_path liuhaotian/llava-v1.6-mistral-7b \
+ --version mistral_instruct \
+ --data_path /home/larry5/project/LLaVA-1.6-ft/data/peft/check_no/check_no_dataset.json \
+ --image_folder /home/larry5/project/LLaVA-1.6-ft/data/data/ \
+ --vision_tower openai/clip-vit-large-patch14-336 \
+ --mm_projector_type mlp2x_gelu \
+ --mm_vision_select_layer -2 \
+ --mm_use_im_start_end False \
+ --mm_use_im_patch_token False \
+ --mm_patch_merge_type spatial_unpad \
+ --image_aspect_ratio anyres \
+ --group_by_modality_length False \
+ --bf16 False \
+ --fp16 True \
+ --output_dir /home/larry5/project/LLaVA-1.6-ft/scripts_peft/mistral/lora/llava-lora-mistral-r128a256/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model \
+ --num_train_epochs 1 \
+ --per_device_train_batch_size 10 \
+ --per_device_eval_batch_size 1 \
+ --gradient_accumulation_steps 1 \
+ --evaluation_strategy "no" \
+ --save_strategy "steps" \
+ --save_steps 500 \
+ --save_total_limit 5 \
+ --learning_rate 2e-5 \
+ --weight_decay 0. \
+ --warmup_ratio 0.05 \
+ --lr_scheduler_type "cosine" \
+ --logging_steps 1 \
+ --tf32 True \
+ --model_max_length 4096 \
+ --gradient_checkpointing True \
+ --dataloader_num_workers 4 \
+ --lazy_preprocess True \
+ --report_to wandb \
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/README.md b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..bdb138eee6972419f6d60676388b52fd99ec478e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/README.md
@@ -0,0 +1,202 @@
+---
+library_name: peft
+base_model: liuhaotian/llava-v1.6-mistral-7b
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.10.0
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/adapter_config.json b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..0e163707620676751a77d129283cfb640741b250
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/adapter_config.json
@@ -0,0 +1,34 @@
+{
+ "alpha_pattern": {},
+ "auto_mapping": null,
+ "base_model_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "bias": "none",
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 256,
+ "lora_dropout": 0.05,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "r": 128,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "down_proj",
+ "v_proj",
+ "k_proj",
+ "up_proj",
+ "o_proj",
+ "q_proj",
+ "gate_proj"
+ ],
+ "task_type": "CAUSAL_LM",
+ "use_dora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..793a1a71ca546956a55b6828f4e3ea4e67ae6fb5
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c8f0dbb6856b765a05c7e93391c05550af39b46719ef6459192604c7184f0a89
+size 708924928
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/README.md b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..bdb138eee6972419f6d60676388b52fd99ec478e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/README.md
@@ -0,0 +1,202 @@
+---
+library_name: peft
+base_model: liuhaotian/llava-v1.6-mistral-7b
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.10.0
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_config.json b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..0e163707620676751a77d129283cfb640741b250
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_config.json
@@ -0,0 +1,34 @@
+{
+ "alpha_pattern": {},
+ "auto_mapping": null,
+ "base_model_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "bias": "none",
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 256,
+ "lora_dropout": 0.05,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "r": 128,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "down_proj",
+ "v_proj",
+ "k_proj",
+ "up_proj",
+ "o_proj",
+ "q_proj",
+ "gate_proj"
+ ],
+ "task_type": "CAUSAL_LM",
+ "use_dora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_model.safetensors b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..a2557ce3d3357d955269e2a1a3620e9e96c42728
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:01fdbcbbf2edbb9aa8fef04a478d715b9e9e545bc872424d656678c3fcd77ccb
+size 1417762896
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_model_states.pt b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_model_states.pt
new file mode 100644
index 0000000000000000000000000000000000000000..ac29cf0cf9f5bf1ef167b5d40e9829b5dc1dde1d
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:226fc9e905373774d8912bdb20c78509485e53529d90f992b6b6206046ce75a3
+size 632242
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_optim_states.pt b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000000000000000000000000000000000000..a09c96de2137a53bbe876f666af8f7f10157ed86
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8bbc9332ac931af2f88625870a521e28cc0450e95c0a1599e4837a0696810681
+size 4504787266
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/latest b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/latest
new file mode 100644
index 0000000000000000000000000000000000000000..f0b47ce15fff9a01b2a416a473b2148085048a50
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/latest
@@ -0,0 +1 @@
+global_step500
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/rng_state.pth b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..01582df0d776ac681b70983ebc573290db9da60a
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b39f5a82d2a2a1a7a3a30b0bf2d931224c4be95b6c1b51efcca4ecb335f633f0
+size 14244
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/scheduler.pt b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..b01e1fb1ea70560e1480afa552e4d755c9fc00c8
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e9b700d1b0ce8daefc78edb016ce7385644daea455f48cf6c5e4891f50fab41
+size 1064
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/special_tokens_map.json b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..14761dcf1466dc232bd41de9c21d4c617b15755e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/special_tokens_map.json
@@ -0,0 +1,24 @@
+{
+ "bos_token": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false
+ },
+ "eos_token": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false
+ },
+ "pad_token": "",
+ "unk_token": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false
+ }
+}
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer.model b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer.model
new file mode 100644
index 0000000000000000000000000000000000000000..8b443ef19c2a19acc3ac64fb9c3db4a72921dff6
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer.model
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
+size 493443
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer_config.json b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..d0ea5c3458cd84f0062b47fa0476bb328b3e208a
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer_config.json
@@ -0,0 +1,44 @@
+{
+ "add_bos_token": true,
+ "add_eos_token": false,
+ "added_tokens_decoder": {
+ "0": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "1": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "2": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ }
+ },
+ "additional_special_tokens": [],
+ "bos_token": "",
+ "chat_template": "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if message['role'] == 'user' %}{{ '[INST] ' + message['content'] + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ message['content'] + eos_token}}{% else %}{{ raise_exception('Only user and assistant roles are supported!') }}{% endif %}{% endfor %}",
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "",
+ "legacy": true,
+ "model_max_length": 4096,
+ "pad_token": "",
+ "padding_side": "right",
+ "sp_model_kwargs": {},
+ "spaces_between_special_tokens": false,
+ "tokenizer_class": "LlamaTokenizer",
+ "unk_token": "",
+ "use_default_system_prompt": false
+}
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/trainer_state.json b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..7318989220461e8abd627b9a2bf5ed2a69cbf609
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/trainer_state.json
@@ -0,0 +1,3021 @@
+{
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.8665511265164645,
+ "eval_steps": 500,
+ "global_step": 500,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "epoch": 0.0,
+ "learning_rate": 6.896551724137931e-07,
+ "loss": 0.22,
+ "step": 1
+ },
+ {
+ "epoch": 0.0,
+ "learning_rate": 1.3793103448275862e-06,
+ "loss": 0.3344,
+ "step": 2
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.0689655172413796e-06,
+ "loss": 0.1757,
+ "step": 3
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.7586206896551725e-06,
+ "loss": 0.1164,
+ "step": 4
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 3.448275862068966e-06,
+ "loss": 0.0771,
+ "step": 5
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 4.137931034482759e-06,
+ "loss": 0.0448,
+ "step": 6
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 4.8275862068965525e-06,
+ "loss": 0.0181,
+ "step": 7
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 5.517241379310345e-06,
+ "loss": 0.0101,
+ "step": 8
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 6.206896551724138e-06,
+ "loss": 0.0879,
+ "step": 9
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 6.896551724137932e-06,
+ "loss": 0.0985,
+ "step": 10
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.586206896551724e-06,
+ "loss": 0.0134,
+ "step": 11
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 8.275862068965518e-06,
+ "loss": 0.0125,
+ "step": 12
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 8.965517241379312e-06,
+ "loss": 0.0061,
+ "step": 13
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 9.655172413793105e-06,
+ "loss": 0.0632,
+ "step": 14
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.0344827586206898e-05,
+ "loss": 0.1219,
+ "step": 15
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.103448275862069e-05,
+ "loss": 0.0033,
+ "step": 16
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.1724137931034483e-05,
+ "loss": 0.0014,
+ "step": 17
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.2413793103448277e-05,
+ "loss": 0.0101,
+ "step": 18
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.310344827586207e-05,
+ "loss": 0.0243,
+ "step": 19
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.3793103448275863e-05,
+ "loss": 0.0489,
+ "step": 20
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.4482758620689657e-05,
+ "loss": 0.0759,
+ "step": 21
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.5172413793103448e-05,
+ "loss": 0.0398,
+ "step": 22
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.586206896551724e-05,
+ "loss": 0.0177,
+ "step": 23
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.6551724137931037e-05,
+ "loss": 0.003,
+ "step": 24
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.7241379310344828e-05,
+ "loss": 0.101,
+ "step": 25
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.7931034482758623e-05,
+ "loss": 0.0107,
+ "step": 26
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.8620689655172415e-05,
+ "loss": 0.1171,
+ "step": 27
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.931034482758621e-05,
+ "loss": 0.0034,
+ "step": 28
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 2e-05,
+ "loss": 0.0876,
+ "step": 29
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.9999835673561284e-05,
+ "loss": 0.076,
+ "step": 30
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.9999342699645774e-05,
+ "loss": 0.0034,
+ "step": 31
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9998521094455198e-05,
+ "loss": 0.0078,
+ "step": 32
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9997370884991842e-05,
+ "loss": 0.1035,
+ "step": 33
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9995892109057675e-05,
+ "loss": 0.1977,
+ "step": 34
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.99940848152531e-05,
+ "loss": 0.0091,
+ "step": 35
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.99940848152531e-05,
+ "loss": 0.1202,
+ "step": 36
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9991949062975336e-05,
+ "loss": 0.2529,
+ "step": 37
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9989484922416503e-05,
+ "loss": 0.0083,
+ "step": 38
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9986692474561292e-05,
+ "loss": 0.0756,
+ "step": 39
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9983571811184297e-05,
+ "loss": 0.1869,
+ "step": 40
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9980123034847025e-05,
+ "loss": 0.0357,
+ "step": 41
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9976346258894502e-05,
+ "loss": 0.0935,
+ "step": 42
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9972241607451552e-05,
+ "loss": 0.0436,
+ "step": 43
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.996780921541873e-05,
+ "loss": 0.0582,
+ "step": 44
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9963049228467875e-05,
+ "loss": 0.0388,
+ "step": 45
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9957961803037325e-05,
+ "loss": 0.0256,
+ "step": 46
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9952547106326787e-05,
+ "loss": 0.0561,
+ "step": 47
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9946805316291817e-05,
+ "loss": 0.0654,
+ "step": 48
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9940736621638e-05,
+ "loss": 0.0149,
+ "step": 49
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.993434122181474e-05,
+ "loss": 0.1098,
+ "step": 50
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.992761932700868e-05,
+ "loss": 0.0321,
+ "step": 51
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9920571158136837e-05,
+ "loss": 0.0154,
+ "step": 52
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9913196946839304e-05,
+ "loss": 0.0319,
+ "step": 53
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.990549693547166e-05,
+ "loss": 0.0812,
+ "step": 54
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9897471377096992e-05,
+ "loss": 0.0021,
+ "step": 55
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9889120535477584e-05,
+ "loss": 0.0049,
+ "step": 56
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9880444685066252e-05,
+ "loss": 0.0076,
+ "step": 57
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.987144411099731e-05,
+ "loss": 0.0344,
+ "step": 58
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9862119109077226e-05,
+ "loss": 0.0273,
+ "step": 59
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.985246998577486e-05,
+ "loss": 0.128,
+ "step": 60
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.985246998577486e-05,
+ "loss": 0.1083,
+ "step": 61
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.984249705821143e-05,
+ "loss": 0.0264,
+ "step": 62
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9832200654150077e-05,
+ "loss": 0.0513,
+ "step": 63
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9821581111985072e-05,
+ "loss": 0.0494,
+ "step": 64
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.981063878073073e-05,
+ "loss": 0.0866,
+ "step": 65
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.979937402000991e-05,
+ "loss": 0.0027,
+ "step": 66
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9787787200042224e-05,
+ "loss": 0.0013,
+ "step": 67
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.977587870163184e-05,
+ "loss": 0.0624,
+ "step": 68
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9763648916154982e-05,
+ "loss": 0.0617,
+ "step": 69
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.975109824554707e-05,
+ "loss": 0.0131,
+ "step": 70
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.973822710228951e-05,
+ "loss": 0.0499,
+ "step": 71
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.972503590939612e-05,
+ "loss": 0.0263,
+ "step": 72
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.971152510039926e-05,
+ "loss": 0.1537,
+ "step": 73
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9697695119335547e-05,
+ "loss": 0.0017,
+ "step": 74
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9683546420731292e-05,
+ "loss": 0.0376,
+ "step": 75
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9669079469587548e-05,
+ "loss": 0.0018,
+ "step": 76
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.965429474136482e-05,
+ "loss": 0.0199,
+ "step": 77
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.963919272196746e-05,
+ "loss": 0.0501,
+ "step": 78
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9623773907727682e-05,
+ "loss": 0.0005,
+ "step": 79
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9608038805389253e-05,
+ "loss": 0.1262,
+ "step": 80
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9591987932090836e-05,
+ "loss": 0.0047,
+ "step": 81
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9575621815349e-05,
+ "loss": 0.009,
+ "step": 82
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9558940993040885e-05,
+ "loss": 0.0154,
+ "step": 83
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.954194601338651e-05,
+ "loss": 0.0011,
+ "step": 84
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.952463743493078e-05,
+ "loss": 0.0052,
+ "step": 85
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9507015826525096e-05,
+ "loss": 0.0229,
+ "step": 86
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9489081767308696e-05,
+ "loss": 0.0018,
+ "step": 87
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9470835846689596e-05,
+ "loss": 0.0013,
+ "step": 88
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9452278664325227e-05,
+ "loss": 0.0074,
+ "step": 89
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9433410830102724e-05,
+ "loss": 0.0205,
+ "step": 90
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9414232964118893e-05,
+ "loss": 0.0026,
+ "step": 91
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.939474569665981e-05,
+ "loss": 0.1344,
+ "step": 92
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.937494966818014e-05,
+ "loss": 0.0314,
+ "step": 93
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9354845529282042e-05,
+ "loss": 0.022,
+ "step": 94
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.933443394069383e-05,
+ "loss": 0.0051,
+ "step": 95
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9313715573248238e-05,
+ "loss": 0.0056,
+ "step": 96
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9292691107860374e-05,
+ "loss": 0.0133,
+ "step": 97
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.927136123550534e-05,
+ "loss": 0.0109,
+ "step": 98
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9249726657195534e-05,
+ "loss": 0.0322,
+ "step": 99
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.922778808395759e-05,
+ "loss": 0.0068,
+ "step": 100
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9205546236809037e-05,
+ "loss": 0.0015,
+ "step": 101
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9183001846734573e-05,
+ "loss": 0.0495,
+ "step": 102
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9160155654662075e-05,
+ "loss": 0.0935,
+ "step": 103
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9137008411438213e-05,
+ "loss": 0.0096,
+ "step": 104
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9113560877803798e-05,
+ "loss": 0.004,
+ "step": 105
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9089813824368765e-05,
+ "loss": 0.0376,
+ "step": 106
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.9065768031586864e-05,
+ "loss": 0.0069,
+ "step": 107
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.9041424289729994e-05,
+ "loss": 0.0008,
+ "step": 108
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.901678339886223e-05,
+ "loss": 0.014,
+ "step": 109
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8991846168813547e-05,
+ "loss": 0.0046,
+ "step": 110
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.896661341915318e-05,
+ "loss": 0.0013,
+ "step": 111
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8941085979162714e-05,
+ "loss": 0.1203,
+ "step": 112
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.891526468780881e-05,
+ "loss": 0.0151,
+ "step": 113
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8889150393715627e-05,
+ "loss": 0.0246,
+ "step": 114
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8862743955136966e-05,
+ "loss": 0.0282,
+ "step": 115
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8836046239928025e-05,
+ "loss": 0.0033,
+ "step": 116
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8809058125516894e-05,
+ "loss": 0.0281,
+ "step": 117
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8781780498875727e-05,
+ "loss": 0.0508,
+ "step": 118
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8754214256491564e-05,
+ "loss": 0.0738,
+ "step": 119
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8726360304336896e-05,
+ "loss": 0.0048,
+ "step": 120
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8698219557839875e-05,
+ "loss": 0.0649,
+ "step": 121
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.866979294185423e-05,
+ "loss": 0.0053,
+ "step": 122
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.864108139062888e-05,
+ "loss": 0.0165,
+ "step": 123
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8612085847777215e-05,
+ "loss": 0.0066,
+ "step": 124
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.858280726624609e-05,
+ "loss": 0.0023,
+ "step": 125
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.855324660828452e-05,
+ "loss": 0.0308,
+ "step": 126
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8523404845412028e-05,
+ "loss": 0.224,
+ "step": 127
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.849328295838674e-05,
+ "loss": 0.0128,
+ "step": 128
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8462881937173144e-05,
+ "loss": 0.0362,
+ "step": 129
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8432202780909542e-05,
+ "loss": 0.0699,
+ "step": 130
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8401246497875238e-05,
+ "loss": 0.0157,
+ "step": 131
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8370014105457378e-05,
+ "loss": 0.0355,
+ "step": 132
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8338506630117527e-05,
+ "loss": 0.0003,
+ "step": 133
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8306725107357933e-05,
+ "loss": 0.0747,
+ "step": 134
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.827467058168748e-05,
+ "loss": 0.0029,
+ "step": 135
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.824234410658738e-05,
+ "loss": 0.0355,
+ "step": 136
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8209746744476538e-05,
+ "loss": 0.0194,
+ "step": 137
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.817687956667664e-05,
+ "loss": 0.0548,
+ "step": 138
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8143743653376944e-05,
+ "loss": 0.0087,
+ "step": 139
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.811034009359877e-05,
+ "loss": 0.0089,
+ "step": 140
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8076669985159726e-05,
+ "loss": 0.0073,
+ "step": 141
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8042734434637615e-05,
+ "loss": 0.0252,
+ "step": 142
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8008534557334064e-05,
+ "loss": 0.1149,
+ "step": 143
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7974071477237887e-05,
+ "loss": 0.0008,
+ "step": 144
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7939346326988127e-05,
+ "loss": 0.0276,
+ "step": 145
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7904360247836838e-05,
+ "loss": 0.0032,
+ "step": 146
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7869114389611574e-05,
+ "loss": 0.013,
+ "step": 147
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7833609910677613e-05,
+ "loss": 0.0004,
+ "step": 148
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7797847977899873e-05,
+ "loss": 0.1131,
+ "step": 149
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7761829766604556e-05,
+ "loss": 0.0019,
+ "step": 150
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7725556460540553e-05,
+ "loss": 0.0099,
+ "step": 151
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7689029251840492e-05,
+ "loss": 0.0627,
+ "step": 152
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7652249340981608e-05,
+ "loss": 0.0626,
+ "step": 153
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7615217936746246e-05,
+ "loss": 0.0007,
+ "step": 154
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.757793625618217e-05,
+ "loss": 0.1323,
+ "step": 155
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7540405524562533e-05,
+ "loss": 0.0348,
+ "step": 156
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.750262697534563e-05,
+ "loss": 0.0024,
+ "step": 157
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7464601850134353e-05,
+ "loss": 0.0134,
+ "step": 158
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.742633139863538e-05,
+ "loss": 0.0037,
+ "step": 159
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.738781687861812e-05,
+ "loss": 0.0089,
+ "step": 160
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7349059555873348e-05,
+ "loss": 0.0082,
+ "step": 161
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.731006070417163e-05,
+ "loss": 0.0082,
+ "step": 162
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7270821605221448e-05,
+ "loss": 0.003,
+ "step": 163
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7231343548627085e-05,
+ "loss": 0.0097,
+ "step": 164
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7191627831846226e-05,
+ "loss": 0.0123,
+ "step": 165
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7151675760147325e-05,
+ "loss": 0.0011,
+ "step": 166
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7111488646566728e-05,
+ "loss": 0.1161,
+ "step": 167
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7071067811865477e-05,
+ "loss": 0.0262,
+ "step": 168
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7030414584485938e-05,
+ "loss": 0.0992,
+ "step": 169
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.6989530300508126e-05,
+ "loss": 0.0019,
+ "step": 170
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6948416303605796e-05,
+ "loss": 0.0056,
+ "step": 171
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.690707394500229e-05,
+ "loss": 0.0053,
+ "step": 172
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6865504583426117e-05,
+ "loss": 0.0796,
+ "step": 173
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6823709585066308e-05,
+ "loss": 0.003,
+ "step": 174
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6781690323527512e-05,
+ "loss": 0.0228,
+ "step": 175
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6739448179784846e-05,
+ "loss": 0.0108,
+ "step": 176
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.669698454213852e-05,
+ "loss": 0.0053,
+ "step": 177
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.665430080616821e-05,
+ "loss": 0.0339,
+ "step": 178
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6611398374687172e-05,
+ "loss": 0.0375,
+ "step": 179
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6568278657696166e-05,
+ "loss": 0.0007,
+ "step": 180
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6524943072337094e-05,
+ "loss": 0.002,
+ "step": 181
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6481393042846442e-05,
+ "loss": 0.0152,
+ "step": 182
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6437630000508466e-05,
+ "loss": 0.0039,
+ "step": 183
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6393655383608132e-05,
+ "loss": 0.0065,
+ "step": 184
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.634947063738389e-05,
+ "loss": 0.0028,
+ "step": 185
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.630507721398013e-05,
+ "loss": 0.0037,
+ "step": 186
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6260476572399494e-05,
+ "loss": 0.013,
+ "step": 187
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6215670178454893e-05,
+ "loss": 0.0137,
+ "step": 188
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6170659504721365e-05,
+ "loss": 0.0516,
+ "step": 189
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6125446030487642e-05,
+ "loss": 0.0333,
+ "step": 190
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.608003124170758e-05,
+ "loss": 0.0041,
+ "step": 191
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6034416630951265e-05,
+ "loss": 0.0053,
+ "step": 192
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.598860369735601e-05,
+ "loss": 0.0003,
+ "step": 193
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.594259394657707e-05,
+ "loss": 0.2021,
+ "step": 194
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.589638889073813e-05,
+ "loss": 0.0217,
+ "step": 195
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.584999004838165e-05,
+ "loss": 0.0141,
+ "step": 196
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5803398944418934e-05,
+ "loss": 0.0006,
+ "step": 197
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5756617110080023e-05,
+ "loss": 0.0186,
+ "step": 198
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.570964608286336e-05,
+ "loss": 0.0073,
+ "step": 199
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5662487406485273e-05,
+ "loss": 0.0481,
+ "step": 200
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.561514263082923e-05,
+ "loss": 0.0002,
+ "step": 201
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5567613311894908e-05,
+ "loss": 0.0093,
+ "step": 202
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5519901011747046e-05,
+ "loss": 0.0393,
+ "step": 203
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5472007298464117e-05,
+ "loss": 0.0019,
+ "step": 204
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5423933746086793e-05,
+ "loss": 0.0035,
+ "step": 205
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5375681934566203e-05,
+ "loss": 0.0926,
+ "step": 206
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.532725344971202e-05,
+ "loss": 0.0379,
+ "step": 207
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.527864988314033e-05,
+ "loss": 0.0011,
+ "step": 208
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5229872832221336e-05,
+ "loss": 0.0024,
+ "step": 209
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5180923900026847e-05,
+ "loss": 0.0005,
+ "step": 210
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5131804695277612e-05,
+ "loss": 0.002,
+ "step": 211
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5131804695277612e-05,
+ "loss": 0.195,
+ "step": 212
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5082516832290424e-05,
+ "loss": 0.0007,
+ "step": 213
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5082516832290424e-05,
+ "loss": 0.0333,
+ "step": 214
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5033061930925081e-05,
+ "loss": 0.0148,
+ "step": 215
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.4983441616531152e-05,
+ "loss": 0.1092,
+ "step": 216
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4933657519894542e-05,
+ "loss": 0.0003,
+ "step": 217
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4883711277183917e-05,
+ "loss": 0.0008,
+ "step": 218
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.483360452989691e-05,
+ "loss": 0.0761,
+ "step": 219
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4783338924806191e-05,
+ "loss": 0.0002,
+ "step": 220
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4732916113905336e-05,
+ "loss": 0.0023,
+ "step": 221
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4682337754354534e-05,
+ "loss": 0.0008,
+ "step": 222
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4631605508426124e-05,
+ "loss": 0.0004,
+ "step": 223
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4580721043449968e-05,
+ "loss": 0.0853,
+ "step": 224
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4529686031758642e-05,
+ "loss": 0.0958,
+ "step": 225
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4478502150632503e-05,
+ "loss": 0.0142,
+ "step": 226
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4427171082244523e-05,
+ "loss": 0.038,
+ "step": 227
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4375694513605037e-05,
+ "loss": 0.004,
+ "step": 228
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4324074136506283e-05,
+ "loss": 0.0091,
+ "step": 229
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.427231164746681e-05,
+ "loss": 0.0714,
+ "step": 230
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4220408747675714e-05,
+ "loss": 0.0618,
+ "step": 231
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4168367142936736e-05,
+ "loss": 0.031,
+ "step": 232
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4116188543612182e-05,
+ "loss": 0.0235,
+ "step": 233
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.4063874664566734e-05,
+ "loss": 0.0027,
+ "step": 234
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.4011427225111091e-05,
+ "loss": 0.1164,
+ "step": 235
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3958847948945428e-05,
+ "loss": 0.0296,
+ "step": 236
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3906138564102794e-05,
+ "loss": 0.0319,
+ "step": 237
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3853300802892285e-05,
+ "loss": 0.0744,
+ "step": 238
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.380033640184213e-05,
+ "loss": 0.0016,
+ "step": 239
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3747247101642605e-05,
+ "loss": 0.0307,
+ "step": 240
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.369403464708884e-05,
+ "loss": 0.0102,
+ "step": 241
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3640700787023465e-05,
+ "loss": 0.0709,
+ "step": 242
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.358724727427914e-05,
+ "loss": 0.0292,
+ "step": 243
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3533675865620937e-05,
+ "loss": 0.0308,
+ "step": 244
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3479988321688619e-05,
+ "loss": 0.0331,
+ "step": 245
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3426186406938769e-05,
+ "loss": 0.0022,
+ "step": 246
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.337227188958679e-05,
+ "loss": 0.0527,
+ "step": 247
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3318246541548812e-05,
+ "loss": 0.0625,
+ "step": 248
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3264112138383445e-05,
+ "loss": 0.121,
+ "step": 249
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3209870459233422e-05,
+ "loss": 0.1122,
+ "step": 250
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.315552328676714e-05,
+ "loss": 0.0018,
+ "step": 251
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3101072407120056e-05,
+ "loss": 0.1122,
+ "step": 252
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3046519609836002e-05,
+ "loss": 0.028,
+ "step": 253
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2991866687808355e-05,
+ "loss": 0.004,
+ "step": 254
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2937115437221119e-05,
+ "loss": 0.1273,
+ "step": 255
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2882267657489908e-05,
+ "loss": 0.0723,
+ "step": 256
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2827325151202783e-05,
+ "loss": 0.0252,
+ "step": 257
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2772289724061015e-05,
+ "loss": 0.0202,
+ "step": 258
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2717163184819761e-05,
+ "loss": 0.005,
+ "step": 259
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2661947345228593e-05,
+ "loss": 0.0346,
+ "step": 260
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2606644019971967e-05,
+ "loss": 0.0018,
+ "step": 261
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.255125502660958e-05,
+ "loss": 0.001,
+ "step": 262
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2495782185516638e-05,
+ "loss": 0.0267,
+ "step": 263
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2440227319824024e-05,
+ "loss": 0.0369,
+ "step": 264
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2384592255358385e-05,
+ "loss": 0.085,
+ "step": 265
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2328878820582122e-05,
+ "loss": 0.0776,
+ "step": 266
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2273088846533303e-05,
+ "loss": 0.0086,
+ "step": 267
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2217224166765478e-05,
+ "loss": 0.1672,
+ "step": 268
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.216128661728742e-05,
+ "loss": 0.0571,
+ "step": 269
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2105278036502787e-05,
+ "loss": 0.0663,
+ "step": 270
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.204920026514971e-05,
+ "loss": 0.0057,
+ "step": 271
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1993055146240273e-05,
+ "loss": 0.018,
+ "step": 272
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1936844524999966e-05,
+ "loss": 0.0013,
+ "step": 273
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1880570248807033e-05,
+ "loss": 0.0021,
+ "step": 274
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1824234167131748e-05,
+ "loss": 0.0732,
+ "step": 275
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1767838131475654e-05,
+ "loss": 0.0053,
+ "step": 276
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.171138399531068e-05,
+ "loss": 0.0258,
+ "step": 277
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1654873614018266e-05,
+ "loss": 0.0943,
+ "step": 278
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1598308844828348e-05,
+ "loss": 0.0011,
+ "step": 279
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1541691546758343e-05,
+ "loss": 0.0781,
+ "step": 280
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1485023580552039e-05,
+ "loss": 0.0078,
+ "step": 281
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1428306808618456e-05,
+ "loss": 0.067,
+ "step": 282
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1371543094970624e-05,
+ "loss": 0.0188,
+ "step": 283
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.131473430516432e-05,
+ "loss": 0.0005,
+ "step": 284
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1257882306236776e-05,
+ "loss": 0.0017,
+ "step": 285
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1200988966645286e-05,
+ "loss": 0.0009,
+ "step": 286
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1144056156205834e-05,
+ "loss": 0.0087,
+ "step": 287
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1087085746031612e-05,
+ "loss": 0.0678,
+ "step": 288
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1030079608471544e-05,
+ "loss": 0.002,
+ "step": 289
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.0973039617048748e-05,
+ "loss": 0.0742,
+ "step": 290
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.091596764639895e-05,
+ "loss": 0.001,
+ "step": 291
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0858865572208892e-05,
+ "loss": 0.0016,
+ "step": 292
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.080173527115467e-05,
+ "loss": 0.0317,
+ "step": 293
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0744578620840065e-05,
+ "loss": 0.0461,
+ "step": 294
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0687397499734842e-05,
+ "loss": 0.002,
+ "step": 295
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0630193787112994e-05,
+ "loss": 0.0008,
+ "step": 296
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0572969362991e-05,
+ "loss": 0.0604,
+ "step": 297
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0515726108066025e-05,
+ "loss": 0.0513,
+ "step": 298
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0458465903654107e-05,
+ "loss": 0.0007,
+ "step": 299
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0401190631628348e-05,
+ "loss": 0.0144,
+ "step": 300
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.034390217435704e-05,
+ "loss": 0.0002,
+ "step": 301
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0286602414641818e-05,
+ "loss": 0.007,
+ "step": 302
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0229293235655768e-05,
+ "loss": 0.0012,
+ "step": 303
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0171976520881552e-05,
+ "loss": 0.0118,
+ "step": 304
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.011465415404949e-05,
+ "loss": 0.0163,
+ "step": 305
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.005732801907567e-05,
+ "loss": 0.0012,
+ "step": 306
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1e-05,
+ "loss": 0.0206,
+ "step": 307
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.942671980924336e-06,
+ "loss": 0.0014,
+ "step": 308
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.88534584595051e-06,
+ "loss": 0.008,
+ "step": 309
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.82802347911845e-06,
+ "loss": 0.0016,
+ "step": 310
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.770706764344235e-06,
+ "loss": 0.0019,
+ "step": 311
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.713397585358189e-06,
+ "loss": 0.0082,
+ "step": 312
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.65609782564296e-06,
+ "loss": 0.1033,
+ "step": 313
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.598809368371656e-06,
+ "loss": 0.06,
+ "step": 314
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.541534096345896e-06,
+ "loss": 0.0028,
+ "step": 315
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.484273891933982e-06,
+ "loss": 0.0309,
+ "step": 316
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.427030637009002e-06,
+ "loss": 0.0243,
+ "step": 317
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.369806212887008e-06,
+ "loss": 0.0116,
+ "step": 318
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.312602500265162e-06,
+ "loss": 0.0049,
+ "step": 319
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.255421379159935e-06,
+ "loss": 0.0005,
+ "step": 320
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.198264728845332e-06,
+ "loss": 0.1163,
+ "step": 321
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.14113442779111e-06,
+ "loss": 0.007,
+ "step": 322
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.084032353601053e-06,
+ "loss": 0.0624,
+ "step": 323
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.026960382951253e-06,
+ "loss": 0.0014,
+ "step": 324
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.969920391528459e-06,
+ "loss": 0.0039,
+ "step": 325
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.912914253968391e-06,
+ "loss": 0.0032,
+ "step": 326
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.855943843794171e-06,
+ "loss": 0.0022,
+ "step": 327
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.799011033354716e-06,
+ "loss": 0.036,
+ "step": 328
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.742117693763229e-06,
+ "loss": 0.0109,
+ "step": 329
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.685265694835681e-06,
+ "loss": 0.1677,
+ "step": 330
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.628456905029383e-06,
+ "loss": 0.0719,
+ "step": 331
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.571693191381545e-06,
+ "loss": 0.0012,
+ "step": 332
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.514976419447963e-06,
+ "loss": 0.0172,
+ "step": 333
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.458308453241664e-06,
+ "loss": 0.0033,
+ "step": 334
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.401691155171654e-06,
+ "loss": 0.0215,
+ "step": 335
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.345126385981737e-06,
+ "loss": 0.0965,
+ "step": 336
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.288616004689321e-06,
+ "loss": 0.0044,
+ "step": 337
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.23216186852435e-06,
+ "loss": 0.0018,
+ "step": 338
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.175765832868252e-06,
+ "loss": 0.0035,
+ "step": 339
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.119429751192972e-06,
+ "loss": 0.0413,
+ "step": 340
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.063155475000037e-06,
+ "loss": 0.0011,
+ "step": 341
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.006944853759732e-06,
+ "loss": 0.0175,
+ "step": 342
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.950799734850292e-06,
+ "loss": 0.0034,
+ "step": 343
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.894721963497214e-06,
+ "loss": 0.0622,
+ "step": 344
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.838713382712583e-06,
+ "loss": 0.085,
+ "step": 345
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.782775833234522e-06,
+ "loss": 0.0018,
+ "step": 346
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.726911153466699e-06,
+ "loss": 0.0569,
+ "step": 347
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.67112117941788e-06,
+ "loss": 0.0022,
+ "step": 348
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.615407744641618e-06,
+ "loss": 0.0948,
+ "step": 349
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.559772680175979e-06,
+ "loss": 0.0015,
+ "step": 350
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.504217814483364e-06,
+ "loss": 0.0014,
+ "step": 351
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.448744973390423e-06,
+ "loss": 0.0025,
+ "step": 352
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.393355980028039e-06,
+ "loss": 0.095,
+ "step": 353
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.338052654771407e-06,
+ "loss": 0.019,
+ "step": 354
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.282836815180241e-06,
+ "loss": 0.001,
+ "step": 355
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.227710275938987e-06,
+ "loss": 0.0035,
+ "step": 356
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.172674848797218e-06,
+ "loss": 0.0793,
+ "step": 357
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.117732342510093e-06,
+ "loss": 0.0008,
+ "step": 358
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.062884562778883e-06,
+ "loss": 0.0129,
+ "step": 359
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.008133312191649e-06,
+ "loss": 0.0234,
+ "step": 360
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.953480390164001e-06,
+ "loss": 0.0049,
+ "step": 361
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.898927592879945e-06,
+ "loss": 0.0133,
+ "step": 362
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.844476713232863e-06,
+ "loss": 0.0024,
+ "step": 363
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.790129540766581e-06,
+ "loss": 0.0112,
+ "step": 364
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.735887861616555e-06,
+ "loss": 0.0018,
+ "step": 365
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.68175345845119e-06,
+ "loss": 0.0021,
+ "step": 366
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.627728110413214e-06,
+ "loss": 0.0347,
+ "step": 367
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.5738135930612355e-06,
+ "loss": 0.0193,
+ "step": 368
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.520011678311382e-06,
+ "loss": 0.001,
+ "step": 369
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.466324134379066e-06,
+ "loss": 0.0499,
+ "step": 370
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.412752725720864e-06,
+ "loss": 0.0011,
+ "step": 371
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.359299212976535e-06,
+ "loss": 0.0006,
+ "step": 372
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.305965352911162e-06,
+ "loss": 0.0025,
+ "step": 373
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.252752898357397e-06,
+ "loss": 0.0015,
+ "step": 374
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.1996635981578755e-06,
+ "loss": 0.0019,
+ "step": 375
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.146699197107715e-06,
+ "loss": 0.0344,
+ "step": 376
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.093861435897208e-06,
+ "loss": 0.001,
+ "step": 377
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 6.041152051054575e-06,
+ "loss": 0.0007,
+ "step": 378
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.988572774888913e-06,
+ "loss": 0.0031,
+ "step": 379
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.936125335433265e-06,
+ "loss": 0.0145,
+ "step": 380
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.883811456387821e-06,
+ "loss": 0.0021,
+ "step": 381
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.831632857063271e-06,
+ "loss": 0.0145,
+ "step": 382
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.779591252324286e-06,
+ "loss": 0.0127,
+ "step": 383
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.7276883525331915e-06,
+ "loss": 0.0031,
+ "step": 384
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.675925863493721e-06,
+ "loss": 0.0029,
+ "step": 385
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.6243054863949675e-06,
+ "loss": 0.0011,
+ "step": 386
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.5728289177554805e-06,
+ "loss": 0.0057,
+ "step": 387
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.521497849367501e-06,
+ "loss": 0.0019,
+ "step": 388
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.4703139682413585e-06,
+ "loss": 0.0126,
+ "step": 389
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.419278956550037e-06,
+ "loss": 0.0415,
+ "step": 390
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.368394491573876e-06,
+ "loss": 0.0075,
+ "step": 391
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.31766224564547e-06,
+ "loss": 0.0004,
+ "step": 392
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.267083886094668e-06,
+ "loss": 0.0172,
+ "step": 393
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.216661075193814e-06,
+ "loss": 0.0011,
+ "step": 394
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.166395470103092e-06,
+ "loss": 0.0028,
+ "step": 395
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.116288722816087e-06,
+ "loss": 0.0013,
+ "step": 396
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.06634248010546e-06,
+ "loss": 0.072,
+ "step": 397
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.016558383468851e-06,
+ "loss": 0.0028,
+ "step": 398
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.9669380690749215e-06,
+ "loss": 0.0076,
+ "step": 399
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.91748316770958e-06,
+ "loss": 0.0003,
+ "step": 400
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.868195304722391e-06,
+ "loss": 0.0006,
+ "step": 401
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.819076099973152e-06,
+ "loss": 0.0058,
+ "step": 402
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.77012716777867e-06,
+ "loss": 0.0224,
+ "step": 403
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.721350116859675e-06,
+ "loss": 0.0062,
+ "step": 404
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.672746550287985e-06,
+ "loss": 0.0003,
+ "step": 405
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.6243180654337975e-06,
+ "loss": 0.0108,
+ "step": 406
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.576066253913209e-06,
+ "loss": 0.0757,
+ "step": 407
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.527992701535884e-06,
+ "loss": 0.0041,
+ "step": 408
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.480098988252958e-06,
+ "loss": 0.0003,
+ "step": 409
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.432386688105095e-06,
+ "loss": 0.0002,
+ "step": 410
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.384857369170772e-06,
+ "loss": 0.0172,
+ "step": 411
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.337512593514729e-06,
+ "loss": 0.0877,
+ "step": 412
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.290353917136639e-06,
+ "loss": 0.0002,
+ "step": 413
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.243382889919981e-06,
+ "loss": 0.0003,
+ "step": 414
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.1966010555810696e-06,
+ "loss": 0.0683,
+ "step": 415
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.1500099516183555e-06,
+ "loss": 0.072,
+ "step": 416
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.1036111092618725e-06,
+ "loss": 0.0009,
+ "step": 417
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.057406053422933e-06,
+ "loss": 0.0029,
+ "step": 418
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 4.011396302643989e-06,
+ "loss": 0.0039,
+ "step": 419
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.965583369048737e-06,
+ "loss": 0.0005,
+ "step": 420
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.919968758292425e-06,
+ "loss": 0.0412,
+ "step": 421
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.874553969512358e-06,
+ "loss": 0.0023,
+ "step": 422
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.82934049527864e-06,
+ "loss": 0.0233,
+ "step": 423
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.784329821545105e-06,
+ "loss": 0.0006,
+ "step": 424
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.739523427600509e-06,
+ "loss": 0.0702,
+ "step": 425
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6949227860198712e-06,
+ "loss": 0.0131,
+ "step": 426
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.650529362616113e-06,
+ "loss": 0.0916,
+ "step": 427
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.606344616391867e-06,
+ "loss": 0.0005,
+ "step": 428
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.5623699994915363e-06,
+ "loss": 0.0059,
+ "step": 429
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.5186069571535575e-06,
+ "loss": 0.0003,
+ "step": 430
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.475056927662912e-06,
+ "loss": 0.0022,
+ "step": 431
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.4317213423038386e-06,
+ "loss": 0.1156,
+ "step": 432
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.388601625312833e-06,
+ "loss": 0.0024,
+ "step": 433
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.345699193831795e-06,
+ "loss": 0.0005,
+ "step": 434
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.3030154578614783e-06,
+ "loss": 0.0683,
+ "step": 435
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.2605518202151577e-06,
+ "loss": 0.0005,
+ "step": 436
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.218309676472492e-06,
+ "loss": 0.0051,
+ "step": 437
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.1762904149336947e-06,
+ "loss": 0.051,
+ "step": 438
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.134495416573884e-06,
+ "loss": 0.0281,
+ "step": 439
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0929260549977116e-06,
+ "loss": 0.2115,
+ "step": 440
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0515836963942056e-06,
+ "loss": 0.0008,
+ "step": 441
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 3.01046969949188e-06,
+ "loss": 0.0005,
+ "step": 442
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.9695854155140648e-06,
+ "loss": 0.1599,
+ "step": 443
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.9289321881345257e-06,
+ "loss": 0.0008,
+ "step": 444
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.8885113534332742e-06,
+ "loss": 0.1016,
+ "step": 445
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.8483242398526723e-06,
+ "loss": 0.1585,
+ "step": 446
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.80837216815378e-06,
+ "loss": 0.1403,
+ "step": 447
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.7686564513729198e-06,
+ "loss": 0.0157,
+ "step": 448
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.7291783947785544e-06,
+ "loss": 0.0033,
+ "step": 449
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.689939295828371e-06,
+ "loss": 0.0015,
+ "step": 450
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.650940444126654e-06,
+ "loss": 0.0435,
+ "step": 451
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.6121831213818825e-06,
+ "loss": 0.0008,
+ "step": 452
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.5736686013646226e-06,
+ "loss": 0.0009,
+ "step": 453
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.535398149865651e-06,
+ "loss": 0.1289,
+ "step": 454
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4973730246543736e-06,
+ "loss": 0.0088,
+ "step": 455
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4595944754374723e-06,
+ "loss": 0.0111,
+ "step": 456
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.422063743817832e-06,
+ "loss": 0.0183,
+ "step": 457
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.3847820632537565e-06,
+ "loss": 0.0222,
+ "step": 458
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.347750659018397e-06,
+ "loss": 0.0018,
+ "step": 459
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.3109707481595113e-06,
+ "loss": 0.0021,
+ "step": 460
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.27444353945945e-06,
+ "loss": 0.0016,
+ "step": 461
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.2381702333954436e-06,
+ "loss": 0.0016,
+ "step": 462
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.2021520221001304e-06,
+ "loss": 0.001,
+ "step": 463
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.16639008932239e-06,
+ "loss": 0.0011,
+ "step": 464
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.130885610388428e-06,
+ "loss": 0.0224,
+ "step": 465
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0956397521631666e-06,
+ "loss": 0.0141,
+ "step": 466
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0606536730118767e-06,
+ "loss": 0.0007,
+ "step": 467
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0259285227621152e-06,
+ "loss": 0.0025,
+ "step": 468
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.9914654426659374e-06,
+ "loss": 0.0591,
+ "step": 469
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.9572655653623884e-06,
+ "loss": 0.0012,
+ "step": 470
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.9233300148402767e-06,
+ "loss": 0.0732,
+ "step": 471
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.88965990640123e-06,
+ "loss": 0.005,
+ "step": 472
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8562563466230577e-06,
+ "loss": 0.0015,
+ "step": 473
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.823120433323361e-06,
+ "loss": 0.1624,
+ "step": 474
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7902532555234653e-06,
+ "loss": 0.0064,
+ "step": 475
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.757655893412622e-06,
+ "loss": 0.152,
+ "step": 476
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.7253294183125223e-06,
+ "loss": 0.0008,
+ "step": 477
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6932748926420695e-06,
+ "loss": 0.0414,
+ "step": 478
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.661493369882473e-06,
+ "loss": 0.0009,
+ "step": 479
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6299858945426251e-06,
+ "loss": 0.0938,
+ "step": 480
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.5987535021247668e-06,
+ "loss": 0.001,
+ "step": 481
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.5677972190904623e-06,
+ "loss": 0.0014,
+ "step": 482
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.537118062826859e-06,
+ "loss": 0.0023,
+ "step": 483
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.5067170416132603e-06,
+ "loss": 0.0012,
+ "step": 484
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4765951545879732e-06,
+ "loss": 0.0014,
+ "step": 485
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4467533917154842e-06,
+ "loss": 0.0102,
+ "step": 486
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4171927337539103e-06,
+ "loss": 0.0033,
+ "step": 487
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3879141522227878e-06,
+ "loss": 0.0009,
+ "step": 488
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3589186093711227e-06,
+ "loss": 0.088,
+ "step": 489
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3302070581457716e-06,
+ "loss": 0.0011,
+ "step": 490
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3017804421601298e-06,
+ "loss": 0.001,
+ "step": 491
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.273639695663108e-06,
+ "loss": 0.0536,
+ "step": 492
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.245785743508441e-06,
+ "loss": 0.0178,
+ "step": 493
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.2182195011242747e-06,
+ "loss": 0.0127,
+ "step": 494
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1909418744831048e-06,
+ "loss": 0.0008,
+ "step": 495
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1639537600719764e-06,
+ "loss": 0.0329,
+ "step": 496
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1372560448630377e-06,
+ "loss": 0.001,
+ "step": 497
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1108496062843743e-06,
+ "loss": 0.0007,
+ "step": 498
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.0847353121911952e-06,
+ "loss": 0.0008,
+ "step": 499
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.0589140208372872e-06,
+ "loss": 0.1002,
+ "step": 500
+ }
+ ],
+ "logging_steps": 1.0,
+ "max_steps": 577,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 1,
+ "save_steps": 500,
+ "total_flos": 1001833021440.0,
+ "train_batch_size": 10,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/training_args.bin b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..1a8d9d69f956643ffcb110e550428f53966a2ab7
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6322ce7627903be868d98b9cd110189814e4c85078c9b4b56c507228fa29b95b
+size 6968
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/zero_to_fp32.py b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..c98caae31534368be22b67fc4ae906836c992a8d
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/zero_to_fp32.py
@@ -0,0 +1,587 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+ FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+ FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+ buffers: dict()
+ param_shapes: dict()
+ shared_params: list
+ ds_version: int
+ frozen_param_shapes: dict()
+ frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+ return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+ '''
+ alist.sort(key=natural_keys) sorts in human order
+ http://nedbatchelder.com/blog/200712/human_sorting.html
+ (See Toothy's implementation in the comments)
+ '''
+ return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+ if not os.path.isdir(checkpoint_dir):
+ raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+ # there should be only one file
+ if zero_stage <= 2:
+ file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+ elif zero_stage == 3:
+ file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+ if not os.path.exists(file):
+ raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+ return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+ # XXX: need to test that this simple glob rule works for multi-node setup too
+ ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+ if len(ckpt_files) == 0:
+ raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+ return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+ return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+ return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+ zero_model_states = []
+ for file in files:
+ state_dict = torch.load(file, map_location=device)
+
+ if BUFFER_NAMES not in state_dict:
+ raise ValueError(f"{file} is not a model state checkpoint")
+ buffer_names = state_dict[BUFFER_NAMES]
+ if debug:
+ print("Found buffers:", buffer_names)
+
+ # recover just the buffers while restoring them to fp32 if they were saved in fp16
+ buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+ param_shapes = state_dict[PARAM_SHAPES]
+
+ # collect parameters that are included in param_shapes
+ param_names = []
+ for s in param_shapes:
+ for name in s.keys():
+ param_names.append(name)
+
+ # update with frozen parameters
+ frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+ if frozen_param_shapes is not None:
+ if debug:
+ print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+ param_names += list(frozen_param_shapes.keys())
+
+ # handle shared params
+ shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+ ds_version = state_dict.get(DS_VERSION, None)
+
+ frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+ z_model_state = zero_model_state(buffers=buffers,
+ param_shapes=param_shapes,
+ shared_params=shared_params,
+ ds_version=ds_version,
+ frozen_param_shapes=frozen_param_shapes,
+ frozen_param_fragments=frozen_param_fragments)
+ zero_model_states.append(z_model_state)
+
+ return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+ total_files = len(files)
+ state_dicts = []
+ for f in files:
+ state_dict = torch.load(f, map_location=device)
+ # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+ # and also handle the case where it was already removed by another helper script
+ state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+ state_dicts.append(state_dict)
+
+ if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+ raise ValueError(f"{files[0]} is not a zero checkpoint")
+ zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+ world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+ # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+ # parameters can be different from data parallelism for non-expert parameters. So we can just
+ # use the max of the partition_count to get the dp world_size.
+
+ if type(world_size) is list:
+ world_size = max(world_size)
+
+ if world_size != total_files:
+ raise ValueError(
+ f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+ "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+ )
+
+ # the groups are named differently in each stage
+ if zero_stage <= 2:
+ fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+ elif zero_stage == 3:
+ fp32_groups_key = FP32_FLAT_GROUPS
+ else:
+ raise ValueError(f"unknown zero stage {zero_stage}")
+
+ if zero_stage <= 2:
+ fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+ elif zero_stage == 3:
+ # if there is more than one param group, there will be multiple flattened tensors - one
+ # flattened tensor per group - for simplicity merge them into a single tensor
+ #
+ # XXX: could make the script more memory efficient for when there are multiple groups - it
+ # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+ fp32_flat_groups = [
+ torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+ ]
+
+ return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir):
+ """
+ Returns fp32 state_dict reconstructed from ds checkpoint
+
+ Args:
+ - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+ """
+ print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+ optim_files = get_optim_files(ds_checkpoint_dir)
+ zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+ print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+ model_files = get_model_state_files(ds_checkpoint_dir)
+
+ zero_model_states = parse_model_states(model_files)
+ print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+ if zero_stage <= 2:
+ return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+ elif zero_stage == 3:
+ return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+ if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+ return
+
+ frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+ frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+ if debug:
+ num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+ print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+ wanted_params = len(frozen_param_shapes)
+ wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+ avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+ print(f'Frozen params: Have {avail_numel} numels to process.')
+ print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+ total_params = 0
+ total_numel = 0
+ for name, shape in frozen_param_shapes.items():
+ total_params += 1
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+
+ state_dict[name] = frozen_param_fragments[name]
+
+ if debug:
+ print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+ print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+ param_shapes = zero_model_states[0].param_shapes
+
+ # Reconstruction protocol:
+ #
+ # XXX: document this
+
+ if debug:
+ for i in range(world_size):
+ for j in range(len(fp32_flat_groups[0])):
+ print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+ # XXX: memory usage doubles here (zero2)
+ num_param_groups = len(fp32_flat_groups[0])
+ merged_single_partition_of_fp32_groups = []
+ for i in range(num_param_groups):
+ merged_partitions = [sd[i] for sd in fp32_flat_groups]
+ full_single_fp32_vector = torch.cat(merged_partitions, 0)
+ merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+ avail_numel = sum(
+ [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+ if debug:
+ wanted_params = sum([len(shapes) for shapes in param_shapes])
+ wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+ # not asserting if there is a mismatch due to possible padding
+ print(f"Have {avail_numel} numels to process.")
+ print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+ # params
+ # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+ # out-of-core computing solution
+ total_numel = 0
+ total_params = 0
+ for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+ offset = 0
+ avail_numel = full_single_fp32_vector.numel()
+ for name, shape in shapes.items():
+
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+ total_params += 1
+
+ if debug:
+ print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+ state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+ offset += unpartitioned_numel
+
+ # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+ # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+ # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+ # live optimizer object, so we are checking that the numbers are within the right range
+ align_to = 2 * world_size
+
+ def zero2_align(x):
+ return align_to * math.ceil(x / align_to)
+
+ if debug:
+ print(f"original offset={offset}, avail_numel={avail_numel}")
+
+ offset = zero2_align(offset)
+ avail_numel = zero2_align(avail_numel)
+
+ if debug:
+ print(f"aligned offset={offset}, avail_numel={avail_numel}")
+
+ # Sanity check
+ if offset != avail_numel:
+ raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+ print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+ state_dict = OrderedDict()
+
+ # buffers
+ buffers = zero_model_states[0].buffers
+ state_dict.update(buffers)
+ if debug:
+ print(f"added {len(buffers)} buffers")
+
+ _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+ _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+ # recover shared parameters
+ for pair in zero_model_states[0].shared_params:
+ if pair[1] in state_dict:
+ state_dict[pair[0]] = state_dict[pair[1]]
+
+ return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+ remainder = unpartitioned_numel % world_size
+ padding_numel = (world_size - remainder) if remainder else 0
+ partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+ return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+ if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+ return
+
+ if debug:
+ for i in range(world_size):
+ num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+ print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+ frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+ wanted_params = len(frozen_param_shapes)
+ wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+ avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+ print(f'Frozen params: Have {avail_numel} numels to process.')
+ print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+ total_params = 0
+ total_numel = 0
+ for name, shape in zero_model_states[0].frozen_param_shapes.items():
+ total_params += 1
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+
+ param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+ state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+ partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+ if debug:
+ print(
+ f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+ )
+
+ print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+ param_shapes = zero_model_states[0].param_shapes
+ avail_numel = fp32_flat_groups[0].numel() * world_size
+ # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+ # param, re-consolidating each param, while dealing with padding if any
+
+ # merge list of dicts, preserving order
+ param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+ if debug:
+ for i in range(world_size):
+ print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+ wanted_params = len(param_shapes)
+ wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+ # not asserting if there is a mismatch due to possible padding
+ avail_numel = fp32_flat_groups[0].numel() * world_size
+ print(f"Trainable params: Have {avail_numel} numels to process.")
+ print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+ # params
+ # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+ # out-of-core computing solution
+ offset = 0
+ total_numel = 0
+ total_params = 0
+ for name, shape in param_shapes.items():
+
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+ total_params += 1
+
+ partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+ if debug:
+ print(
+ f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+ )
+
+ # XXX: memory usage doubles here
+ state_dict[name] = torch.cat(
+ tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+ 0).narrow(0, 0, unpartitioned_numel).view(shape)
+ offset += partitioned_numel
+
+ offset *= world_size
+
+ # Sanity check
+ if offset != avail_numel:
+ raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+ print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+ state_dict = OrderedDict()
+
+ # buffers
+ buffers = zero_model_states[0].buffers
+ state_dict.update(buffers)
+ if debug:
+ print(f"added {len(buffers)} buffers")
+
+ _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+ _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+ # recover shared parameters
+ for pair in zero_model_states[0].shared_params:
+ if pair[1] in state_dict:
+ state_dict[pair[0]] = state_dict[pair[1]]
+
+ return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None):
+ """
+ Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+ ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+ via a model hub.
+
+ Args:
+ - ``checkpoint_dir``: path to the desired checkpoint folder
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+
+ Returns:
+ - pytorch ``state_dict``
+
+ Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+ you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+ the checkpoint.
+
+ A typical usage might be ::
+
+ from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+ # do the training and checkpoint saving
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+ model = model.cpu() # move to cpu
+ model.load_state_dict(state_dict)
+ # submit to model hub or save the model to share with others
+
+ In this example the ``model`` will no longer be usable in the deepspeed context of the same
+ application. i.e. you will need to re-initialize the deepspeed engine, since
+ ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+ If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+ """
+ if tag is None:
+ latest_path = os.path.join(checkpoint_dir, 'latest')
+ if os.path.isfile(latest_path):
+ with open(latest_path, 'r') as fd:
+ tag = fd.read().strip()
+ else:
+ raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+ ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+ if not os.path.isdir(ds_checkpoint_dir):
+ raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+ return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None):
+ """
+ Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+ loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+ Args:
+ - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+ - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+ """
+
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+ print(f"Saving fp32 state dict to {output_file}")
+ torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+ """
+ 1. Put the provided model to cpu
+ 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+ 3. Load it into the provided model
+
+ Args:
+ - ``model``: the model object to update
+ - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+ Returns:
+ - ``model`: modified model
+
+ Make sure you have plenty of CPU memory available before you call this function. If you don't
+ have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+ conveniently placed for you in the checkpoint folder.
+
+ A typical usage might be ::
+
+ from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+ model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+ # submit to model hub or save the model to share with others
+
+ Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+ of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+ ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+ """
+ logger.info(f"Extracting fp32 weights")
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+ logger.info(f"Overwriting model with fp32 weights")
+ model = model.cpu()
+ model.load_state_dict(state_dict, strict=False)
+
+ return model
+
+
+if __name__ == "__main__":
+
+ parser = argparse.ArgumentParser()
+ parser.add_argument("checkpoint_dir",
+ type=str,
+ help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+ parser.add_argument(
+ "output_file",
+ type=str,
+ help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+ parser.add_argument("-t",
+ "--tag",
+ type=str,
+ default=None,
+ help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+ parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+ args = parser.parse_args()
+
+ debug = args.debug
+
+ convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, args.output_file, tag=args.tag)
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/config.json b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..93e133af45036a778791b5679a8953a4f6a35a33
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/config.json
@@ -0,0 +1,70 @@
+{
+ "_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "architectures": [
+ "LlavaMistralForCausalLM"
+ ],
+ "attention_dropout": 0.0,
+ "bos_token_id": 1,
+ "eos_token_id": 2,
+ "freeze_mm_mlp_adapter": false,
+ "freeze_mm_vision_resampler": false,
+ "hidden_act": "silu",
+ "hidden_size": 4096,
+ "image_aspect_ratio": "anyres",
+ "image_crop_resolution": 224,
+ "image_grid_pinpoints": [
+ [
+ 336,
+ 672
+ ],
+ [
+ 672,
+ 336
+ ],
+ [
+ 672,
+ 672
+ ],
+ [
+ 1008,
+ 336
+ ],
+ [
+ 336,
+ 1008
+ ]
+ ],
+ "image_split_resolution": 224,
+ "initializer_range": 0.02,
+ "intermediate_size": 14336,
+ "max_position_embeddings": 32768,
+ "mm_hidden_size": 1024,
+ "mm_patch_merge_type": "spatial_unpad",
+ "mm_projector_lr": 2e-05,
+ "mm_projector_type": "mlp2x_gelu",
+ "mm_resampler_type": null,
+ "mm_use_im_patch_token": false,
+ "mm_use_im_start_end": false,
+ "mm_vision_select_feature": "patch",
+ "mm_vision_select_layer": -2,
+ "mm_vision_tower": "openai/clip-vit-large-patch14-336",
+ "mm_vision_tower_lr": 2e-06,
+ "model_type": "llava_mistral",
+ "num_attention_heads": 32,
+ "num_hidden_layers": 32,
+ "num_key_value_heads": 8,
+ "rms_norm_eps": 1e-05,
+ "rope_theta": 1000000.0,
+ "sliding_window": null,
+ "tie_word_embeddings": false,
+ "tokenizer_model_max_length": 4096,
+ "tokenizer_padding_side": "right",
+ "torch_dtype": "bfloat16",
+ "transformers_version": "4.37.2",
+ "tune_mm_mlp_adapter": false,
+ "tune_mm_vision_resampler": false,
+ "unfreeze_mm_vision_tower": true,
+ "use_cache": true,
+ "use_mm_proj": true,
+ "vocab_size": 32000
+}
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin
new file mode 100644
index 0000000000000000000000000000000000000000..a078832f8b614e52aa214a4cd8bf3fa19e896476
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee9f345bc360c7d0d0577b86bb5be9ab7afda7df19aa0456f3946cc9bc4f90a7
+size 41961648
diff --git a/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/trainer_state.json b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..3d5fef17742ba11728748f11ddd8e1095ddf0d33
--- /dev/null
+++ b/CheckGuard Models/wholeimage/check_no/llava-lora-mistral-r128a256-10BS-model/trainer_state.json
@@ -0,0 +1,3492 @@
+{
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.0,
+ "eval_steps": 500,
+ "global_step": 577,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "epoch": 0.0,
+ "learning_rate": 6.896551724137931e-07,
+ "loss": 0.22,
+ "step": 1
+ },
+ {
+ "epoch": 0.0,
+ "learning_rate": 1.3793103448275862e-06,
+ "loss": 0.3344,
+ "step": 2
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.0689655172413796e-06,
+ "loss": 0.1757,
+ "step": 3
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.7586206896551725e-06,
+ "loss": 0.1164,
+ "step": 4
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 3.448275862068966e-06,
+ "loss": 0.0771,
+ "step": 5
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 4.137931034482759e-06,
+ "loss": 0.0448,
+ "step": 6
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 4.8275862068965525e-06,
+ "loss": 0.0181,
+ "step": 7
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 5.517241379310345e-06,
+ "loss": 0.0101,
+ "step": 8
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 6.206896551724138e-06,
+ "loss": 0.0879,
+ "step": 9
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 6.896551724137932e-06,
+ "loss": 0.0985,
+ "step": 10
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.586206896551724e-06,
+ "loss": 0.0134,
+ "step": 11
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 8.275862068965518e-06,
+ "loss": 0.0125,
+ "step": 12
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 8.965517241379312e-06,
+ "loss": 0.0061,
+ "step": 13
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 9.655172413793105e-06,
+ "loss": 0.0632,
+ "step": 14
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.0344827586206898e-05,
+ "loss": 0.1219,
+ "step": 15
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.103448275862069e-05,
+ "loss": 0.0033,
+ "step": 16
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.1724137931034483e-05,
+ "loss": 0.0014,
+ "step": 17
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.2413793103448277e-05,
+ "loss": 0.0101,
+ "step": 18
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.310344827586207e-05,
+ "loss": 0.0243,
+ "step": 19
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.3793103448275863e-05,
+ "loss": 0.0489,
+ "step": 20
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.4482758620689657e-05,
+ "loss": 0.0759,
+ "step": 21
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.5172413793103448e-05,
+ "loss": 0.0398,
+ "step": 22
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.586206896551724e-05,
+ "loss": 0.0177,
+ "step": 23
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.6551724137931037e-05,
+ "loss": 0.003,
+ "step": 24
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.7241379310344828e-05,
+ "loss": 0.101,
+ "step": 25
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.7931034482758623e-05,
+ "loss": 0.0107,
+ "step": 26
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.8620689655172415e-05,
+ "loss": 0.1171,
+ "step": 27
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.931034482758621e-05,
+ "loss": 0.0034,
+ "step": 28
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 2e-05,
+ "loss": 0.0876,
+ "step": 29
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.9999835673561284e-05,
+ "loss": 0.076,
+ "step": 30
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.9999342699645774e-05,
+ "loss": 0.0034,
+ "step": 31
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9998521094455198e-05,
+ "loss": 0.0078,
+ "step": 32
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9997370884991842e-05,
+ "loss": 0.1035,
+ "step": 33
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9995892109057675e-05,
+ "loss": 0.1977,
+ "step": 34
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.99940848152531e-05,
+ "loss": 0.0091,
+ "step": 35
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.99940848152531e-05,
+ "loss": 0.1202,
+ "step": 36
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9991949062975336e-05,
+ "loss": 0.2529,
+ "step": 37
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9989484922416503e-05,
+ "loss": 0.0083,
+ "step": 38
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9986692474561292e-05,
+ "loss": 0.0756,
+ "step": 39
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9983571811184297e-05,
+ "loss": 0.1869,
+ "step": 40
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9980123034847025e-05,
+ "loss": 0.0357,
+ "step": 41
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9976346258894502e-05,
+ "loss": 0.0935,
+ "step": 42
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9972241607451552e-05,
+ "loss": 0.0436,
+ "step": 43
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.996780921541873e-05,
+ "loss": 0.0582,
+ "step": 44
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9963049228467875e-05,
+ "loss": 0.0388,
+ "step": 45
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9957961803037325e-05,
+ "loss": 0.0256,
+ "step": 46
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9952547106326787e-05,
+ "loss": 0.0561,
+ "step": 47
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9946805316291817e-05,
+ "loss": 0.0654,
+ "step": 48
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9940736621638e-05,
+ "loss": 0.0149,
+ "step": 49
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.993434122181474e-05,
+ "loss": 0.1098,
+ "step": 50
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.992761932700868e-05,
+ "loss": 0.0321,
+ "step": 51
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9920571158136837e-05,
+ "loss": 0.0154,
+ "step": 52
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9913196946839304e-05,
+ "loss": 0.0319,
+ "step": 53
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.990549693547166e-05,
+ "loss": 0.0812,
+ "step": 54
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9897471377096992e-05,
+ "loss": 0.0021,
+ "step": 55
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9889120535477584e-05,
+ "loss": 0.0049,
+ "step": 56
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9880444685066252e-05,
+ "loss": 0.0076,
+ "step": 57
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.987144411099731e-05,
+ "loss": 0.0344,
+ "step": 58
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9862119109077226e-05,
+ "loss": 0.0273,
+ "step": 59
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.985246998577486e-05,
+ "loss": 0.128,
+ "step": 60
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.985246998577486e-05,
+ "loss": 0.1083,
+ "step": 61
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.984249705821143e-05,
+ "loss": 0.0264,
+ "step": 62
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9832200654150077e-05,
+ "loss": 0.0513,
+ "step": 63
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9821581111985072e-05,
+ "loss": 0.0494,
+ "step": 64
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.981063878073073e-05,
+ "loss": 0.0866,
+ "step": 65
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.979937402000991e-05,
+ "loss": 0.0027,
+ "step": 66
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9787787200042224e-05,
+ "loss": 0.0013,
+ "step": 67
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.977587870163184e-05,
+ "loss": 0.0624,
+ "step": 68
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9763648916154982e-05,
+ "loss": 0.0617,
+ "step": 69
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.975109824554707e-05,
+ "loss": 0.0131,
+ "step": 70
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.973822710228951e-05,
+ "loss": 0.0499,
+ "step": 71
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.972503590939612e-05,
+ "loss": 0.0263,
+ "step": 72
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.971152510039926e-05,
+ "loss": 0.1537,
+ "step": 73
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9697695119335547e-05,
+ "loss": 0.0017,
+ "step": 74
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9683546420731292e-05,
+ "loss": 0.0376,
+ "step": 75
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9669079469587548e-05,
+ "loss": 0.0018,
+ "step": 76
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.965429474136482e-05,
+ "loss": 0.0199,
+ "step": 77
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.963919272196746e-05,
+ "loss": 0.0501,
+ "step": 78
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9623773907727682e-05,
+ "loss": 0.0005,
+ "step": 79
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9608038805389253e-05,
+ "loss": 0.1262,
+ "step": 80
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9591987932090836e-05,
+ "loss": 0.0047,
+ "step": 81
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9575621815349e-05,
+ "loss": 0.009,
+ "step": 82
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9558940993040885e-05,
+ "loss": 0.0154,
+ "step": 83
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.954194601338651e-05,
+ "loss": 0.0011,
+ "step": 84
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.952463743493078e-05,
+ "loss": 0.0052,
+ "step": 85
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9507015826525096e-05,
+ "loss": 0.0229,
+ "step": 86
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9489081767308696e-05,
+ "loss": 0.0018,
+ "step": 87
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9470835846689596e-05,
+ "loss": 0.0013,
+ "step": 88
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9452278664325227e-05,
+ "loss": 0.0074,
+ "step": 89
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9433410830102724e-05,
+ "loss": 0.0205,
+ "step": 90
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9414232964118893e-05,
+ "loss": 0.0026,
+ "step": 91
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.939474569665981e-05,
+ "loss": 0.1344,
+ "step": 92
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.937494966818014e-05,
+ "loss": 0.0314,
+ "step": 93
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9354845529282042e-05,
+ "loss": 0.022,
+ "step": 94
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.933443394069383e-05,
+ "loss": 0.0051,
+ "step": 95
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9313715573248238e-05,
+ "loss": 0.0056,
+ "step": 96
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9292691107860374e-05,
+ "loss": 0.0133,
+ "step": 97
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.927136123550534e-05,
+ "loss": 0.0109,
+ "step": 98
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9249726657195534e-05,
+ "loss": 0.0322,
+ "step": 99
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.922778808395759e-05,
+ "loss": 0.0068,
+ "step": 100
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9205546236809037e-05,
+ "loss": 0.0015,
+ "step": 101
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9183001846734573e-05,
+ "loss": 0.0495,
+ "step": 102
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9160155654662075e-05,
+ "loss": 0.0935,
+ "step": 103
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9137008411438213e-05,
+ "loss": 0.0096,
+ "step": 104
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9113560877803798e-05,
+ "loss": 0.004,
+ "step": 105
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9089813824368765e-05,
+ "loss": 0.0376,
+ "step": 106
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.9065768031586864e-05,
+ "loss": 0.0069,
+ "step": 107
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.9041424289729994e-05,
+ "loss": 0.0008,
+ "step": 108
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.901678339886223e-05,
+ "loss": 0.014,
+ "step": 109
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8991846168813547e-05,
+ "loss": 0.0046,
+ "step": 110
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.896661341915318e-05,
+ "loss": 0.0013,
+ "step": 111
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8941085979162714e-05,
+ "loss": 0.1203,
+ "step": 112
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.891526468780881e-05,
+ "loss": 0.0151,
+ "step": 113
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8889150393715627e-05,
+ "loss": 0.0246,
+ "step": 114
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8862743955136966e-05,
+ "loss": 0.0282,
+ "step": 115
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8836046239928025e-05,
+ "loss": 0.0033,
+ "step": 116
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8809058125516894e-05,
+ "loss": 0.0281,
+ "step": 117
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8781780498875727e-05,
+ "loss": 0.0508,
+ "step": 118
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8754214256491564e-05,
+ "loss": 0.0738,
+ "step": 119
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8726360304336896e-05,
+ "loss": 0.0048,
+ "step": 120
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8698219557839875e-05,
+ "loss": 0.0649,
+ "step": 121
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.866979294185423e-05,
+ "loss": 0.0053,
+ "step": 122
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.864108139062888e-05,
+ "loss": 0.0165,
+ "step": 123
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8612085847777215e-05,
+ "loss": 0.0066,
+ "step": 124
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.858280726624609e-05,
+ "loss": 0.0023,
+ "step": 125
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.855324660828452e-05,
+ "loss": 0.0308,
+ "step": 126
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8523404845412028e-05,
+ "loss": 0.224,
+ "step": 127
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.849328295838674e-05,
+ "loss": 0.0128,
+ "step": 128
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8462881937173144e-05,
+ "loss": 0.0362,
+ "step": 129
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8432202780909542e-05,
+ "loss": 0.0699,
+ "step": 130
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8401246497875238e-05,
+ "loss": 0.0157,
+ "step": 131
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8370014105457378e-05,
+ "loss": 0.0355,
+ "step": 132
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8338506630117527e-05,
+ "loss": 0.0003,
+ "step": 133
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8306725107357933e-05,
+ "loss": 0.0747,
+ "step": 134
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.827467058168748e-05,
+ "loss": 0.0029,
+ "step": 135
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.824234410658738e-05,
+ "loss": 0.0355,
+ "step": 136
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8209746744476538e-05,
+ "loss": 0.0194,
+ "step": 137
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.817687956667664e-05,
+ "loss": 0.0548,
+ "step": 138
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8143743653376944e-05,
+ "loss": 0.0087,
+ "step": 139
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.811034009359877e-05,
+ "loss": 0.0089,
+ "step": 140
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8076669985159726e-05,
+ "loss": 0.0073,
+ "step": 141
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8042734434637615e-05,
+ "loss": 0.0252,
+ "step": 142
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8008534557334064e-05,
+ "loss": 0.1149,
+ "step": 143
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7974071477237887e-05,
+ "loss": 0.0008,
+ "step": 144
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7939346326988127e-05,
+ "loss": 0.0276,
+ "step": 145
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7904360247836838e-05,
+ "loss": 0.0032,
+ "step": 146
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7869114389611574e-05,
+ "loss": 0.013,
+ "step": 147
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7833609910677613e-05,
+ "loss": 0.0004,
+ "step": 148
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7797847977899873e-05,
+ "loss": 0.1131,
+ "step": 149
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7761829766604556e-05,
+ "loss": 0.0019,
+ "step": 150
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7725556460540553e-05,
+ "loss": 0.0099,
+ "step": 151
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7689029251840492e-05,
+ "loss": 0.0627,
+ "step": 152
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7652249340981608e-05,
+ "loss": 0.0626,
+ "step": 153
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7615217936746246e-05,
+ "loss": 0.0007,
+ "step": 154
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.757793625618217e-05,
+ "loss": 0.1323,
+ "step": 155
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7540405524562533e-05,
+ "loss": 0.0348,
+ "step": 156
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.750262697534563e-05,
+ "loss": 0.0024,
+ "step": 157
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7464601850134353e-05,
+ "loss": 0.0134,
+ "step": 158
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.742633139863538e-05,
+ "loss": 0.0037,
+ "step": 159
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.738781687861812e-05,
+ "loss": 0.0089,
+ "step": 160
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7349059555873348e-05,
+ "loss": 0.0082,
+ "step": 161
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.731006070417163e-05,
+ "loss": 0.0082,
+ "step": 162
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7270821605221448e-05,
+ "loss": 0.003,
+ "step": 163
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7231343548627085e-05,
+ "loss": 0.0097,
+ "step": 164
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7191627831846226e-05,
+ "loss": 0.0123,
+ "step": 165
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7151675760147325e-05,
+ "loss": 0.0011,
+ "step": 166
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7111488646566728e-05,
+ "loss": 0.1161,
+ "step": 167
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7071067811865477e-05,
+ "loss": 0.0262,
+ "step": 168
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7030414584485938e-05,
+ "loss": 0.0992,
+ "step": 169
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.6989530300508126e-05,
+ "loss": 0.0019,
+ "step": 170
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6948416303605796e-05,
+ "loss": 0.0056,
+ "step": 171
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.690707394500229e-05,
+ "loss": 0.0053,
+ "step": 172
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6865504583426117e-05,
+ "loss": 0.0796,
+ "step": 173
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6823709585066308e-05,
+ "loss": 0.003,
+ "step": 174
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6781690323527512e-05,
+ "loss": 0.0228,
+ "step": 175
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6739448179784846e-05,
+ "loss": 0.0108,
+ "step": 176
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.669698454213852e-05,
+ "loss": 0.0053,
+ "step": 177
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.665430080616821e-05,
+ "loss": 0.0339,
+ "step": 178
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6611398374687172e-05,
+ "loss": 0.0375,
+ "step": 179
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6568278657696166e-05,
+ "loss": 0.0007,
+ "step": 180
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6524943072337094e-05,
+ "loss": 0.002,
+ "step": 181
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6481393042846442e-05,
+ "loss": 0.0152,
+ "step": 182
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6437630000508466e-05,
+ "loss": 0.0039,
+ "step": 183
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6393655383608132e-05,
+ "loss": 0.0065,
+ "step": 184
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.634947063738389e-05,
+ "loss": 0.0028,
+ "step": 185
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.630507721398013e-05,
+ "loss": 0.0037,
+ "step": 186
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6260476572399494e-05,
+ "loss": 0.013,
+ "step": 187
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6215670178454893e-05,
+ "loss": 0.0137,
+ "step": 188
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6170659504721365e-05,
+ "loss": 0.0516,
+ "step": 189
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6125446030487642e-05,
+ "loss": 0.0333,
+ "step": 190
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.608003124170758e-05,
+ "loss": 0.0041,
+ "step": 191
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6034416630951265e-05,
+ "loss": 0.0053,
+ "step": 192
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.598860369735601e-05,
+ "loss": 0.0003,
+ "step": 193
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.594259394657707e-05,
+ "loss": 0.2021,
+ "step": 194
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.589638889073813e-05,
+ "loss": 0.0217,
+ "step": 195
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.584999004838165e-05,
+ "loss": 0.0141,
+ "step": 196
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5803398944418934e-05,
+ "loss": 0.0006,
+ "step": 197
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5756617110080023e-05,
+ "loss": 0.0186,
+ "step": 198
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.570964608286336e-05,
+ "loss": 0.0073,
+ "step": 199
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5662487406485273e-05,
+ "loss": 0.0481,
+ "step": 200
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.561514263082923e-05,
+ "loss": 0.0002,
+ "step": 201
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5567613311894908e-05,
+ "loss": 0.0093,
+ "step": 202
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5519901011747046e-05,
+ "loss": 0.0393,
+ "step": 203
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5472007298464117e-05,
+ "loss": 0.0019,
+ "step": 204
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5423933746086793e-05,
+ "loss": 0.0035,
+ "step": 205
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5375681934566203e-05,
+ "loss": 0.0926,
+ "step": 206
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.532725344971202e-05,
+ "loss": 0.0379,
+ "step": 207
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.527864988314033e-05,
+ "loss": 0.0011,
+ "step": 208
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5229872832221336e-05,
+ "loss": 0.0024,
+ "step": 209
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5180923900026847e-05,
+ "loss": 0.0005,
+ "step": 210
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5131804695277612e-05,
+ "loss": 0.002,
+ "step": 211
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5131804695277612e-05,
+ "loss": 0.195,
+ "step": 212
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5082516832290424e-05,
+ "loss": 0.0007,
+ "step": 213
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5082516832290424e-05,
+ "loss": 0.0333,
+ "step": 214
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5033061930925081e-05,
+ "loss": 0.0148,
+ "step": 215
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.4983441616531152e-05,
+ "loss": 0.1092,
+ "step": 216
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4933657519894542e-05,
+ "loss": 0.0003,
+ "step": 217
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4883711277183917e-05,
+ "loss": 0.0008,
+ "step": 218
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.483360452989691e-05,
+ "loss": 0.0761,
+ "step": 219
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4783338924806191e-05,
+ "loss": 0.0002,
+ "step": 220
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4732916113905336e-05,
+ "loss": 0.0023,
+ "step": 221
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4682337754354534e-05,
+ "loss": 0.0008,
+ "step": 222
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4631605508426124e-05,
+ "loss": 0.0004,
+ "step": 223
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4580721043449968e-05,
+ "loss": 0.0853,
+ "step": 224
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4529686031758642e-05,
+ "loss": 0.0958,
+ "step": 225
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4478502150632503e-05,
+ "loss": 0.0142,
+ "step": 226
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4427171082244523e-05,
+ "loss": 0.038,
+ "step": 227
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4375694513605037e-05,
+ "loss": 0.004,
+ "step": 228
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4324074136506283e-05,
+ "loss": 0.0091,
+ "step": 229
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.427231164746681e-05,
+ "loss": 0.0714,
+ "step": 230
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4220408747675714e-05,
+ "loss": 0.0618,
+ "step": 231
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4168367142936736e-05,
+ "loss": 0.031,
+ "step": 232
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4116188543612182e-05,
+ "loss": 0.0235,
+ "step": 233
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.4063874664566734e-05,
+ "loss": 0.0027,
+ "step": 234
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.4011427225111091e-05,
+ "loss": 0.1164,
+ "step": 235
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3958847948945428e-05,
+ "loss": 0.0296,
+ "step": 236
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3906138564102794e-05,
+ "loss": 0.0319,
+ "step": 237
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3853300802892285e-05,
+ "loss": 0.0744,
+ "step": 238
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.380033640184213e-05,
+ "loss": 0.0016,
+ "step": 239
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3747247101642605e-05,
+ "loss": 0.0307,
+ "step": 240
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.369403464708884e-05,
+ "loss": 0.0102,
+ "step": 241
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3640700787023465e-05,
+ "loss": 0.0709,
+ "step": 242
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.358724727427914e-05,
+ "loss": 0.0292,
+ "step": 243
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3533675865620937e-05,
+ "loss": 0.0308,
+ "step": 244
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3479988321688619e-05,
+ "loss": 0.0331,
+ "step": 245
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3426186406938769e-05,
+ "loss": 0.0022,
+ "step": 246
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.337227188958679e-05,
+ "loss": 0.0527,
+ "step": 247
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3318246541548812e-05,
+ "loss": 0.0625,
+ "step": 248
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3264112138383445e-05,
+ "loss": 0.121,
+ "step": 249
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3209870459233422e-05,
+ "loss": 0.1122,
+ "step": 250
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.315552328676714e-05,
+ "loss": 0.0018,
+ "step": 251
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3101072407120056e-05,
+ "loss": 0.1122,
+ "step": 252
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3046519609836002e-05,
+ "loss": 0.028,
+ "step": 253
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2991866687808355e-05,
+ "loss": 0.004,
+ "step": 254
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2937115437221119e-05,
+ "loss": 0.1273,
+ "step": 255
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2882267657489908e-05,
+ "loss": 0.0723,
+ "step": 256
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2827325151202783e-05,
+ "loss": 0.0252,
+ "step": 257
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2772289724061015e-05,
+ "loss": 0.0202,
+ "step": 258
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2717163184819761e-05,
+ "loss": 0.005,
+ "step": 259
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2661947345228593e-05,
+ "loss": 0.0346,
+ "step": 260
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2606644019971967e-05,
+ "loss": 0.0018,
+ "step": 261
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.255125502660958e-05,
+ "loss": 0.001,
+ "step": 262
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2495782185516638e-05,
+ "loss": 0.0267,
+ "step": 263
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2440227319824024e-05,
+ "loss": 0.0369,
+ "step": 264
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2384592255358385e-05,
+ "loss": 0.085,
+ "step": 265
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2328878820582122e-05,
+ "loss": 0.0776,
+ "step": 266
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2273088846533303e-05,
+ "loss": 0.0086,
+ "step": 267
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2217224166765478e-05,
+ "loss": 0.1672,
+ "step": 268
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.216128661728742e-05,
+ "loss": 0.0571,
+ "step": 269
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2105278036502787e-05,
+ "loss": 0.0663,
+ "step": 270
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.204920026514971e-05,
+ "loss": 0.0057,
+ "step": 271
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1993055146240273e-05,
+ "loss": 0.018,
+ "step": 272
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1936844524999966e-05,
+ "loss": 0.0013,
+ "step": 273
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1880570248807033e-05,
+ "loss": 0.0021,
+ "step": 274
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1824234167131748e-05,
+ "loss": 0.0732,
+ "step": 275
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1767838131475654e-05,
+ "loss": 0.0053,
+ "step": 276
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.171138399531068e-05,
+ "loss": 0.0258,
+ "step": 277
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1654873614018266e-05,
+ "loss": 0.0943,
+ "step": 278
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1598308844828348e-05,
+ "loss": 0.0011,
+ "step": 279
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1541691546758343e-05,
+ "loss": 0.0781,
+ "step": 280
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1485023580552039e-05,
+ "loss": 0.0078,
+ "step": 281
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1428306808618456e-05,
+ "loss": 0.067,
+ "step": 282
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1371543094970624e-05,
+ "loss": 0.0188,
+ "step": 283
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.131473430516432e-05,
+ "loss": 0.0005,
+ "step": 284
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1257882306236776e-05,
+ "loss": 0.0017,
+ "step": 285
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1200988966645286e-05,
+ "loss": 0.0009,
+ "step": 286
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1144056156205834e-05,
+ "loss": 0.0087,
+ "step": 287
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1087085746031612e-05,
+ "loss": 0.0678,
+ "step": 288
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1030079608471544e-05,
+ "loss": 0.002,
+ "step": 289
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.0973039617048748e-05,
+ "loss": 0.0742,
+ "step": 290
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.091596764639895e-05,
+ "loss": 0.001,
+ "step": 291
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0858865572208892e-05,
+ "loss": 0.0016,
+ "step": 292
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.080173527115467e-05,
+ "loss": 0.0317,
+ "step": 293
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0744578620840065e-05,
+ "loss": 0.0461,
+ "step": 294
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0687397499734842e-05,
+ "loss": 0.002,
+ "step": 295
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0630193787112994e-05,
+ "loss": 0.0008,
+ "step": 296
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0572969362991e-05,
+ "loss": 0.0604,
+ "step": 297
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0515726108066025e-05,
+ "loss": 0.0513,
+ "step": 298
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0458465903654107e-05,
+ "loss": 0.0007,
+ "step": 299
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0401190631628348e-05,
+ "loss": 0.0144,
+ "step": 300
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.034390217435704e-05,
+ "loss": 0.0002,
+ "step": 301
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0286602414641818e-05,
+ "loss": 0.007,
+ "step": 302
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0229293235655768e-05,
+ "loss": 0.0012,
+ "step": 303
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0171976520881552e-05,
+ "loss": 0.0118,
+ "step": 304
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.011465415404949e-05,
+ "loss": 0.0163,
+ "step": 305
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.005732801907567e-05,
+ "loss": 0.0012,
+ "step": 306
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1e-05,
+ "loss": 0.0206,
+ "step": 307
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.942671980924336e-06,
+ "loss": 0.0014,
+ "step": 308
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.88534584595051e-06,
+ "loss": 0.008,
+ "step": 309
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.82802347911845e-06,
+ "loss": 0.0016,
+ "step": 310
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.770706764344235e-06,
+ "loss": 0.0019,
+ "step": 311
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.713397585358189e-06,
+ "loss": 0.0082,
+ "step": 312
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.65609782564296e-06,
+ "loss": 0.1033,
+ "step": 313
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.598809368371656e-06,
+ "loss": 0.06,
+ "step": 314
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.541534096345896e-06,
+ "loss": 0.0028,
+ "step": 315
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.484273891933982e-06,
+ "loss": 0.0309,
+ "step": 316
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.427030637009002e-06,
+ "loss": 0.0243,
+ "step": 317
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.369806212887008e-06,
+ "loss": 0.0116,
+ "step": 318
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.312602500265162e-06,
+ "loss": 0.0049,
+ "step": 319
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.255421379159935e-06,
+ "loss": 0.0005,
+ "step": 320
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.198264728845332e-06,
+ "loss": 0.1163,
+ "step": 321
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.14113442779111e-06,
+ "loss": 0.007,
+ "step": 322
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.084032353601053e-06,
+ "loss": 0.0624,
+ "step": 323
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.026960382951253e-06,
+ "loss": 0.0014,
+ "step": 324
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.969920391528459e-06,
+ "loss": 0.0039,
+ "step": 325
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.912914253968391e-06,
+ "loss": 0.0032,
+ "step": 326
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.855943843794171e-06,
+ "loss": 0.0022,
+ "step": 327
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.799011033354716e-06,
+ "loss": 0.036,
+ "step": 328
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.742117693763229e-06,
+ "loss": 0.0109,
+ "step": 329
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.685265694835681e-06,
+ "loss": 0.1677,
+ "step": 330
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.628456905029383e-06,
+ "loss": 0.0719,
+ "step": 331
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.571693191381545e-06,
+ "loss": 0.0012,
+ "step": 332
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.514976419447963e-06,
+ "loss": 0.0172,
+ "step": 333
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.458308453241664e-06,
+ "loss": 0.0033,
+ "step": 334
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.401691155171654e-06,
+ "loss": 0.0215,
+ "step": 335
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.345126385981737e-06,
+ "loss": 0.0965,
+ "step": 336
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.288616004689321e-06,
+ "loss": 0.0044,
+ "step": 337
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.23216186852435e-06,
+ "loss": 0.0018,
+ "step": 338
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.175765832868252e-06,
+ "loss": 0.0035,
+ "step": 339
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.119429751192972e-06,
+ "loss": 0.0413,
+ "step": 340
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.063155475000037e-06,
+ "loss": 0.0011,
+ "step": 341
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.006944853759732e-06,
+ "loss": 0.0175,
+ "step": 342
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.950799734850292e-06,
+ "loss": 0.0034,
+ "step": 343
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.894721963497214e-06,
+ "loss": 0.0622,
+ "step": 344
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.838713382712583e-06,
+ "loss": 0.085,
+ "step": 345
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.782775833234522e-06,
+ "loss": 0.0018,
+ "step": 346
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.726911153466699e-06,
+ "loss": 0.0569,
+ "step": 347
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.67112117941788e-06,
+ "loss": 0.0022,
+ "step": 348
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.615407744641618e-06,
+ "loss": 0.0948,
+ "step": 349
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.559772680175979e-06,
+ "loss": 0.0015,
+ "step": 350
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.504217814483364e-06,
+ "loss": 0.0014,
+ "step": 351
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.448744973390423e-06,
+ "loss": 0.0025,
+ "step": 352
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.393355980028039e-06,
+ "loss": 0.095,
+ "step": 353
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.338052654771407e-06,
+ "loss": 0.019,
+ "step": 354
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.282836815180241e-06,
+ "loss": 0.001,
+ "step": 355
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.227710275938987e-06,
+ "loss": 0.0035,
+ "step": 356
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.172674848797218e-06,
+ "loss": 0.0793,
+ "step": 357
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.117732342510093e-06,
+ "loss": 0.0008,
+ "step": 358
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.062884562778883e-06,
+ "loss": 0.0129,
+ "step": 359
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.008133312191649e-06,
+ "loss": 0.0234,
+ "step": 360
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.953480390164001e-06,
+ "loss": 0.0049,
+ "step": 361
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.898927592879945e-06,
+ "loss": 0.0133,
+ "step": 362
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.844476713232863e-06,
+ "loss": 0.0024,
+ "step": 363
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.790129540766581e-06,
+ "loss": 0.0112,
+ "step": 364
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.735887861616555e-06,
+ "loss": 0.0018,
+ "step": 365
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.68175345845119e-06,
+ "loss": 0.0021,
+ "step": 366
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.627728110413214e-06,
+ "loss": 0.0347,
+ "step": 367
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.5738135930612355e-06,
+ "loss": 0.0193,
+ "step": 368
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.520011678311382e-06,
+ "loss": 0.001,
+ "step": 369
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.466324134379066e-06,
+ "loss": 0.0499,
+ "step": 370
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.412752725720864e-06,
+ "loss": 0.0011,
+ "step": 371
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.359299212976535e-06,
+ "loss": 0.0006,
+ "step": 372
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.305965352911162e-06,
+ "loss": 0.0025,
+ "step": 373
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.252752898357397e-06,
+ "loss": 0.0015,
+ "step": 374
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.1996635981578755e-06,
+ "loss": 0.0019,
+ "step": 375
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.146699197107715e-06,
+ "loss": 0.0344,
+ "step": 376
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.093861435897208e-06,
+ "loss": 0.001,
+ "step": 377
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 6.041152051054575e-06,
+ "loss": 0.0007,
+ "step": 378
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.988572774888913e-06,
+ "loss": 0.0031,
+ "step": 379
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.936125335433265e-06,
+ "loss": 0.0145,
+ "step": 380
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.883811456387821e-06,
+ "loss": 0.0021,
+ "step": 381
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.831632857063271e-06,
+ "loss": 0.0145,
+ "step": 382
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.779591252324286e-06,
+ "loss": 0.0127,
+ "step": 383
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.7276883525331915e-06,
+ "loss": 0.0031,
+ "step": 384
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.675925863493721e-06,
+ "loss": 0.0029,
+ "step": 385
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.6243054863949675e-06,
+ "loss": 0.0011,
+ "step": 386
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.5728289177554805e-06,
+ "loss": 0.0057,
+ "step": 387
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.521497849367501e-06,
+ "loss": 0.0019,
+ "step": 388
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.4703139682413585e-06,
+ "loss": 0.0126,
+ "step": 389
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.419278956550037e-06,
+ "loss": 0.0415,
+ "step": 390
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.368394491573876e-06,
+ "loss": 0.0075,
+ "step": 391
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.31766224564547e-06,
+ "loss": 0.0004,
+ "step": 392
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.267083886094668e-06,
+ "loss": 0.0172,
+ "step": 393
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.216661075193814e-06,
+ "loss": 0.0011,
+ "step": 394
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.166395470103092e-06,
+ "loss": 0.0028,
+ "step": 395
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.116288722816087e-06,
+ "loss": 0.0013,
+ "step": 396
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.06634248010546e-06,
+ "loss": 0.072,
+ "step": 397
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.016558383468851e-06,
+ "loss": 0.0028,
+ "step": 398
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.9669380690749215e-06,
+ "loss": 0.0076,
+ "step": 399
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.91748316770958e-06,
+ "loss": 0.0003,
+ "step": 400
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.868195304722391e-06,
+ "loss": 0.0006,
+ "step": 401
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.819076099973152e-06,
+ "loss": 0.0058,
+ "step": 402
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.77012716777867e-06,
+ "loss": 0.0224,
+ "step": 403
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.721350116859675e-06,
+ "loss": 0.0062,
+ "step": 404
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.672746550287985e-06,
+ "loss": 0.0003,
+ "step": 405
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.6243180654337975e-06,
+ "loss": 0.0108,
+ "step": 406
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.576066253913209e-06,
+ "loss": 0.0757,
+ "step": 407
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.527992701535884e-06,
+ "loss": 0.0041,
+ "step": 408
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.480098988252958e-06,
+ "loss": 0.0003,
+ "step": 409
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.432386688105095e-06,
+ "loss": 0.0002,
+ "step": 410
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.384857369170772e-06,
+ "loss": 0.0172,
+ "step": 411
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.337512593514729e-06,
+ "loss": 0.0877,
+ "step": 412
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.290353917136639e-06,
+ "loss": 0.0002,
+ "step": 413
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.243382889919981e-06,
+ "loss": 0.0003,
+ "step": 414
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.1966010555810696e-06,
+ "loss": 0.0683,
+ "step": 415
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.1500099516183555e-06,
+ "loss": 0.072,
+ "step": 416
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.1036111092618725e-06,
+ "loss": 0.0009,
+ "step": 417
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.057406053422933e-06,
+ "loss": 0.0029,
+ "step": 418
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 4.011396302643989e-06,
+ "loss": 0.0039,
+ "step": 419
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.965583369048737e-06,
+ "loss": 0.0005,
+ "step": 420
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.919968758292425e-06,
+ "loss": 0.0412,
+ "step": 421
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.874553969512358e-06,
+ "loss": 0.0023,
+ "step": 422
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.82934049527864e-06,
+ "loss": 0.0233,
+ "step": 423
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.784329821545105e-06,
+ "loss": 0.0006,
+ "step": 424
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.739523427600509e-06,
+ "loss": 0.0702,
+ "step": 425
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6949227860198712e-06,
+ "loss": 0.0131,
+ "step": 426
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.650529362616113e-06,
+ "loss": 0.0916,
+ "step": 427
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.606344616391867e-06,
+ "loss": 0.0005,
+ "step": 428
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.5623699994915363e-06,
+ "loss": 0.0059,
+ "step": 429
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.5186069571535575e-06,
+ "loss": 0.0003,
+ "step": 430
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.475056927662912e-06,
+ "loss": 0.0022,
+ "step": 431
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.4317213423038386e-06,
+ "loss": 0.1156,
+ "step": 432
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.388601625312833e-06,
+ "loss": 0.0024,
+ "step": 433
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.345699193831795e-06,
+ "loss": 0.0005,
+ "step": 434
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.3030154578614783e-06,
+ "loss": 0.0683,
+ "step": 435
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.2605518202151577e-06,
+ "loss": 0.0005,
+ "step": 436
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.218309676472492e-06,
+ "loss": 0.0051,
+ "step": 437
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.1762904149336947e-06,
+ "loss": 0.051,
+ "step": 438
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.134495416573884e-06,
+ "loss": 0.0281,
+ "step": 439
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0929260549977116e-06,
+ "loss": 0.2115,
+ "step": 440
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0515836963942056e-06,
+ "loss": 0.0008,
+ "step": 441
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 3.01046969949188e-06,
+ "loss": 0.0005,
+ "step": 442
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.9695854155140648e-06,
+ "loss": 0.1599,
+ "step": 443
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.9289321881345257e-06,
+ "loss": 0.0008,
+ "step": 444
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.8885113534332742e-06,
+ "loss": 0.1016,
+ "step": 445
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.8483242398526723e-06,
+ "loss": 0.1585,
+ "step": 446
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.80837216815378e-06,
+ "loss": 0.1403,
+ "step": 447
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.7686564513729198e-06,
+ "loss": 0.0157,
+ "step": 448
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.7291783947785544e-06,
+ "loss": 0.0033,
+ "step": 449
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.689939295828371e-06,
+ "loss": 0.0015,
+ "step": 450
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.650940444126654e-06,
+ "loss": 0.0435,
+ "step": 451
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.6121831213818825e-06,
+ "loss": 0.0008,
+ "step": 452
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.5736686013646226e-06,
+ "loss": 0.0009,
+ "step": 453
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.535398149865651e-06,
+ "loss": 0.1289,
+ "step": 454
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4973730246543736e-06,
+ "loss": 0.0088,
+ "step": 455
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4595944754374723e-06,
+ "loss": 0.0111,
+ "step": 456
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.422063743817832e-06,
+ "loss": 0.0183,
+ "step": 457
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.3847820632537565e-06,
+ "loss": 0.0222,
+ "step": 458
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.347750659018397e-06,
+ "loss": 0.0018,
+ "step": 459
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.3109707481595113e-06,
+ "loss": 0.0021,
+ "step": 460
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.27444353945945e-06,
+ "loss": 0.0016,
+ "step": 461
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.2381702333954436e-06,
+ "loss": 0.0016,
+ "step": 462
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.2021520221001304e-06,
+ "loss": 0.001,
+ "step": 463
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.16639008932239e-06,
+ "loss": 0.0011,
+ "step": 464
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.130885610388428e-06,
+ "loss": 0.0224,
+ "step": 465
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0956397521631666e-06,
+ "loss": 0.0141,
+ "step": 466
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0606536730118767e-06,
+ "loss": 0.0007,
+ "step": 467
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0259285227621152e-06,
+ "loss": 0.0025,
+ "step": 468
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.9914654426659374e-06,
+ "loss": 0.0591,
+ "step": 469
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.9572655653623884e-06,
+ "loss": 0.0012,
+ "step": 470
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.9233300148402767e-06,
+ "loss": 0.0732,
+ "step": 471
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.88965990640123e-06,
+ "loss": 0.005,
+ "step": 472
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8562563466230577e-06,
+ "loss": 0.0015,
+ "step": 473
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.823120433323361e-06,
+ "loss": 0.1624,
+ "step": 474
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7902532555234653e-06,
+ "loss": 0.0064,
+ "step": 475
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.757655893412622e-06,
+ "loss": 0.152,
+ "step": 476
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.7253294183125223e-06,
+ "loss": 0.0008,
+ "step": 477
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6932748926420695e-06,
+ "loss": 0.0414,
+ "step": 478
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.661493369882473e-06,
+ "loss": 0.0009,
+ "step": 479
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6299858945426251e-06,
+ "loss": 0.0938,
+ "step": 480
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.5987535021247668e-06,
+ "loss": 0.001,
+ "step": 481
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.5677972190904623e-06,
+ "loss": 0.0014,
+ "step": 482
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.537118062826859e-06,
+ "loss": 0.0023,
+ "step": 483
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.5067170416132603e-06,
+ "loss": 0.0012,
+ "step": 484
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4765951545879732e-06,
+ "loss": 0.0014,
+ "step": 485
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4467533917154842e-06,
+ "loss": 0.0102,
+ "step": 486
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4171927337539103e-06,
+ "loss": 0.0033,
+ "step": 487
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3879141522227878e-06,
+ "loss": 0.0009,
+ "step": 488
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3589186093711227e-06,
+ "loss": 0.088,
+ "step": 489
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3302070581457716e-06,
+ "loss": 0.0011,
+ "step": 490
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3017804421601298e-06,
+ "loss": 0.001,
+ "step": 491
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.273639695663108e-06,
+ "loss": 0.0536,
+ "step": 492
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.245785743508441e-06,
+ "loss": 0.0178,
+ "step": 493
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.2182195011242747e-06,
+ "loss": 0.0127,
+ "step": 494
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1909418744831048e-06,
+ "loss": 0.0008,
+ "step": 495
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1639537600719764e-06,
+ "loss": 0.0329,
+ "step": 496
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1372560448630377e-06,
+ "loss": 0.001,
+ "step": 497
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1108496062843743e-06,
+ "loss": 0.0007,
+ "step": 498
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.0847353121911952e-06,
+ "loss": 0.0008,
+ "step": 499
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.0589140208372872e-06,
+ "loss": 0.1002,
+ "step": 500
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.0333865808468203e-06,
+ "loss": 0.0013,
+ "step": 501
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.008153831186457e-06,
+ "loss": 0.0359,
+ "step": 502
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.83216601137773e-07,
+ "loss": 0.0183,
+ "step": 503
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.58575710270011e-07,
+ "loss": 0.0189,
+ "step": 504
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 9.342319684131396e-07,
+ "loss": 0.0205,
+ "step": 505
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 9.101861756312369e-07,
+ "loss": 0.1422,
+ "step": 506
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.864391221962065e-07,
+ "loss": 0.1253,
+ "step": 507
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.629915885617912e-07,
+ "loss": 0.001,
+ "step": 508
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.398443453379268e-07,
+ "loss": 0.0326,
+ "step": 509
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.169981532654269e-07,
+ "loss": 0.0362,
+ "step": 510
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.944537631909666e-07,
+ "loss": 0.0025,
+ "step": 511
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.722119160424113e-07,
+ "loss": 0.2316,
+ "step": 512
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.502733428044684e-07,
+ "loss": 0.0009,
+ "step": 513
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.286387644946602e-07,
+ "loss": 0.0008,
+ "step": 514
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.073088921396287e-07,
+ "loss": 0.0009,
+ "step": 515
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 6.862844267517643e-07,
+ "loss": 0.0099,
+ "step": 516
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.655660593061719e-07,
+ "loss": 0.0008,
+ "step": 517
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.451544707179635e-07,
+ "loss": 0.0011,
+ "step": 518
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.250503318198664e-07,
+ "loss": 0.0414,
+ "step": 519
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.052543033401892e-07,
+ "loss": 0.0056,
+ "step": 520
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 5.857670358811096e-07,
+ "loss": 0.0592,
+ "step": 521
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 5.665891698972769e-07,
+ "loss": 0.0139,
+ "step": 522
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.477213356747746e-07,
+ "loss": 0.0274,
+ "step": 523
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.291641533104053e-07,
+ "loss": 0.004,
+ "step": 524
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.109182326913053e-07,
+ "loss": 0.0359,
+ "step": 525
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 4.929841734749063e-07,
+ "loss": 0.0168,
+ "step": 526
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 4.7536256506922507e-07,
+ "loss": 0.001,
+ "step": 527
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.580539866134914e-07,
+ "loss": 0.105,
+ "step": 528
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.410590069591192e-07,
+ "loss": 0.0006,
+ "step": 529
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.2437818465100313e-07,
+ "loss": 0.0707,
+ "step": 530
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.0801206790916815e-07,
+ "loss": 0.0013,
+ "step": 531
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.919611946107493e-07,
+ "loss": 0.0008,
+ "step": 532
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.762260922723182e-07,
+ "loss": 0.0008,
+ "step": 533
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.6080727803254003e-07,
+ "loss": 0.001,
+ "step": 534
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.457052586351817e-07,
+ "loss": 0.0042,
+ "step": 535
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.309205304124552e-07,
+ "loss": 0.0155,
+ "step": 536
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.1645357926870957e-07,
+ "loss": 0.0029,
+ "step": 537
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.0230488066445465e-07,
+ "loss": 0.0401,
+ "step": 538
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 2.8847489960074136e-07,
+ "loss": 0.0315,
+ "step": 539
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.7496409060387973e-07,
+ "loss": 0.0436,
+ "step": 540
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.6177289771049274e-07,
+ "loss": 0.0036,
+ "step": 541
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.489017544529315e-07,
+ "loss": 0.0231,
+ "step": 542
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.3635108384502003e-07,
+ "loss": 0.0116,
+ "step": 543
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.2412129836816287e-07,
+ "loss": 0.0011,
+ "step": 544
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.1221279995777833e-07,
+ "loss": 0.0011,
+ "step": 545
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 2.0062597999009114e-07,
+ "loss": 0.001,
+ "step": 546
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.8936121926927508e-07,
+ "loss": 0.0072,
+ "step": 547
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.7841888801493178e-07,
+ "loss": 0.0031,
+ "step": 548
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.677993458499272e-07,
+ "loss": 0.0314,
+ "step": 549
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.5750294178856872e-07,
+ "loss": 0.002,
+ "step": 550
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.4753001422514125e-07,
+ "loss": 0.0012,
+ "step": 551
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.378808909227769e-07,
+ "loss": 0.1963,
+ "step": 552
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.2855588900269057e-07,
+ "loss": 0.0804,
+ "step": 553
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.1955531493375137e-07,
+ "loss": 0.0057,
+ "step": 554
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.1087946452241871e-07,
+ "loss": 0.0025,
+ "step": 555
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.0252862290301092e-07,
+ "loss": 0.0244,
+ "step": 556
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 9.45030645283418e-08,
+ "loss": 0.0488,
+ "step": 557
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 8.68030531606967e-08,
+ "loss": 0.0262,
+ "step": 558
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 7.94288418631639e-08,
+ "loss": 0.0212,
+ "step": 559
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 7.238067299131901e-08,
+ "loss": 0.0257,
+ "step": 560
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 6.565877818526245e-08,
+ "loss": 0.001,
+ "step": 561
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 5.926337836199891e-08,
+ "loss": 0.0038,
+ "step": 562
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 5.319468370818537e-08,
+ "loss": 0.0277,
+ "step": 563
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 4.7452893673216596e-08,
+ "loss": 0.0009,
+ "step": 564
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 4.203819696267486e-08,
+ "loss": 0.007,
+ "step": 565
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 3.6950771532126004e-08,
+ "loss": 0.0007,
+ "step": 566
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 3.2190784581270786e-08,
+ "loss": 0.001,
+ "step": 567
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 2.7758392548449253e-08,
+ "loss": 0.0172,
+ "step": 568
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 2.3653741105499338e-08,
+ "loss": 0.0019,
+ "step": 569
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.9876965152975102e-08,
+ "loss": 0.0148,
+ "step": 570
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.6428188815703627e-08,
+ "loss": 0.0614,
+ "step": 571
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.3307525438711611e-08,
+ "loss": 0.0032,
+ "step": 572
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.0515077583498346e-08,
+ "loss": 0.0093,
+ "step": 573
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 8.050937024666195e-09,
+ "loss": 0.0009,
+ "step": 574
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 5.9151847469041125e-09,
+ "loss": 0.0212,
+ "step": 575
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 4.1078909423253325e-09,
+ "loss": 0.0008,
+ "step": 576
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 2.629115008160321e-09,
+ "loss": 0.0273,
+ "step": 577
+ },
+ {
+ "epoch": 1.0,
+ "step": 577,
+ "total_flos": 1156027152384.0,
+ "train_loss": 0.03200564120091745,
+ "train_runtime": 4242.8648,
+ "train_samples_per_second": 1.358,
+ "train_steps_per_second": 0.136
+ }
+ ],
+ "logging_steps": 1.0,
+ "max_steps": 577,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 1,
+ "save_steps": 500,
+ "total_flos": 1156027152384.0,
+ "train_batch_size": 10,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/CheckGuard Models/wholeimage/date/finetune_lora_llava_mistral.sh b/CheckGuard Models/wholeimage/date/finetune_lora_llava_mistral.sh
new file mode 100644
index 0000000000000000000000000000000000000000..d5e581ee3cdf75b2579fb103b167f02d0d9351b4
--- /dev/null
+++ b/CheckGuard Models/wholeimage/date/finetune_lora_llava_mistral.sh
@@ -0,0 +1,39 @@
+#!/bin/bash
+
+deepspeed llava/train/train_mem.py \
+ --lora_enable True --lora_r 128 --lora_alpha 256 --mm_projector_lr 2e-5 \
+ --deepspeed ./scripts/zero3.json \
+ --model_name_or_path liuhaotian/llava-v1.6-mistral-7b \
+ --version mistral_instruct \
+ --data_path /home/larry5/project/LLaVA-1.6-ft/data/peft/date/csv_gt/date_modified_path_to_train_val_human-gpt-whole-check.json \
+ --image_folder /home/larry5/project/LLaVA-1.6-ft/data/data/ \
+ --vision_tower openai/clip-vit-large-patch14-336 \
+ --mm_projector_type mlp2x_gelu \
+ --mm_vision_select_layer -2 \
+ --mm_use_im_start_end False \
+ --mm_use_im_patch_token False \
+ --mm_patch_merge_type spatial_unpad \
+ --image_aspect_ratio anyres \
+ --group_by_modality_length False \
+ --bf16 False \
+ --fp16 True \
+ --output_dir /home/larry5/project/LLaVA-1.6-ft/scripts_peft/mistral/lora/llava-lora-mistral-r128a256/wholeimage/date/llava-lora-mistral-r128a256-10BS-model \
+ --num_train_epochs 1 \
+ --per_device_train_batch_size 10 \
+ --per_device_eval_batch_size 1 \
+ --gradient_accumulation_steps 1 \
+ --evaluation_strategy "no" \
+ --save_strategy "steps" \
+ --save_steps 500 \
+ --save_total_limit 5 \
+ --learning_rate 2e-5 \
+ --weight_decay 0. \
+ --warmup_ratio 0.05 \
+ --lr_scheduler_type "cosine" \
+ --logging_steps 1 \
+ --tf32 True \
+ --model_max_length 4096 \
+ --gradient_checkpointing True \
+ --dataloader_num_workers 4 \
+ --lazy_preprocess True \
+ --report_to wandb \
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/README.md b/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..bdb138eee6972419f6d60676388b52fd99ec478e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/README.md
@@ -0,0 +1,202 @@
+---
+library_name: peft
+base_model: liuhaotian/llava-v1.6-mistral-7b
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.10.0
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/adapter_config.json b/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..832a63b9a74cc1cea90552535b9699c9b402fec8
--- /dev/null
+++ b/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/adapter_config.json
@@ -0,0 +1,34 @@
+{
+ "alpha_pattern": {},
+ "auto_mapping": null,
+ "base_model_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "bias": "none",
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 256,
+ "lora_dropout": 0.05,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "r": 128,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "up_proj",
+ "gate_proj",
+ "q_proj",
+ "v_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj"
+ ],
+ "task_type": "CAUSAL_LM",
+ "use_dora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors b/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..7d2fe23932ce99dd3bd04992942268785a7f6b4a
--- /dev/null
+++ b/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e79997695f8f7928e94b7798ff1c85f84ad9cdab989c52c8827cab35990915ad
+size 708924928
diff --git a/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/config.json b/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..93e133af45036a778791b5679a8953a4f6a35a33
--- /dev/null
+++ b/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/config.json
@@ -0,0 +1,70 @@
+{
+ "_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "architectures": [
+ "LlavaMistralForCausalLM"
+ ],
+ "attention_dropout": 0.0,
+ "bos_token_id": 1,
+ "eos_token_id": 2,
+ "freeze_mm_mlp_adapter": false,
+ "freeze_mm_vision_resampler": false,
+ "hidden_act": "silu",
+ "hidden_size": 4096,
+ "image_aspect_ratio": "anyres",
+ "image_crop_resolution": 224,
+ "image_grid_pinpoints": [
+ [
+ 336,
+ 672
+ ],
+ [
+ 672,
+ 336
+ ],
+ [
+ 672,
+ 672
+ ],
+ [
+ 1008,
+ 336
+ ],
+ [
+ 336,
+ 1008
+ ]
+ ],
+ "image_split_resolution": 224,
+ "initializer_range": 0.02,
+ "intermediate_size": 14336,
+ "max_position_embeddings": 32768,
+ "mm_hidden_size": 1024,
+ "mm_patch_merge_type": "spatial_unpad",
+ "mm_projector_lr": 2e-05,
+ "mm_projector_type": "mlp2x_gelu",
+ "mm_resampler_type": null,
+ "mm_use_im_patch_token": false,
+ "mm_use_im_start_end": false,
+ "mm_vision_select_feature": "patch",
+ "mm_vision_select_layer": -2,
+ "mm_vision_tower": "openai/clip-vit-large-patch14-336",
+ "mm_vision_tower_lr": 2e-06,
+ "model_type": "llava_mistral",
+ "num_attention_heads": 32,
+ "num_hidden_layers": 32,
+ "num_key_value_heads": 8,
+ "rms_norm_eps": 1e-05,
+ "rope_theta": 1000000.0,
+ "sliding_window": null,
+ "tie_word_embeddings": false,
+ "tokenizer_model_max_length": 4096,
+ "tokenizer_padding_side": "right",
+ "torch_dtype": "bfloat16",
+ "transformers_version": "4.37.2",
+ "tune_mm_mlp_adapter": false,
+ "tune_mm_vision_resampler": false,
+ "unfreeze_mm_vision_tower": true,
+ "use_cache": true,
+ "use_mm_proj": true,
+ "vocab_size": 32000
+}
diff --git a/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin b/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin
new file mode 100644
index 0000000000000000000000000000000000000000..a9b1ac414d504eea0f57a4a74d09c9d7cdf03330
--- /dev/null
+++ b/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1bdcf0c8f51abd2364a9061cb31a1b79f7ce70a5726220e46518bd17984af4b3
+size 41961648
diff --git a/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/trainer_state.json b/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..1b036b21fd448ed907e68fa48c7bb62fb992486a
--- /dev/null
+++ b/CheckGuard Models/wholeimage/date/llava-lora-mistral-r128a256-10BS-model/trainer_state.json
@@ -0,0 +1,1938 @@
+{
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.0,
+ "eval_steps": 500,
+ "global_step": 318,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "epoch": 0.0,
+ "learning_rate": 1.25e-06,
+ "loss": 0.1459,
+ "step": 1
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.5e-06,
+ "loss": 0.2617,
+ "step": 2
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 3.7500000000000005e-06,
+ "loss": 0.1687,
+ "step": 3
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 5e-06,
+ "loss": 0.1222,
+ "step": 4
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 6.25e-06,
+ "loss": 0.0698,
+ "step": 5
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.500000000000001e-06,
+ "loss": 0.0603,
+ "step": 6
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 8.750000000000001e-06,
+ "loss": 0.0467,
+ "step": 7
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1e-05,
+ "loss": 0.0846,
+ "step": 8
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.125e-05,
+ "loss": 0.0627,
+ "step": 9
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.25e-05,
+ "loss": 0.089,
+ "step": 10
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.375e-05,
+ "loss": 0.0668,
+ "step": 11
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.5000000000000002e-05,
+ "loss": 0.0456,
+ "step": 12
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.6250000000000002e-05,
+ "loss": 0.0442,
+ "step": 13
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.7500000000000002e-05,
+ "loss": 0.0451,
+ "step": 14
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.8750000000000002e-05,
+ "loss": 0.0421,
+ "step": 15
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.8750000000000002e-05,
+ "loss": 0.1125,
+ "step": 16
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 2e-05,
+ "loss": 0.0672,
+ "step": 17
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.999945893187807e-05,
+ "loss": 0.0274,
+ "step": 18
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.999783578606323e-05,
+ "loss": 0.0691,
+ "step": 19
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9995130738201966e-05,
+ "loss": 0.0485,
+ "step": 20
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9991344081017312e-05,
+ "loss": 0.1008,
+ "step": 21
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9986476224277167e-05,
+ "loss": 0.0323,
+ "step": 22
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9980527694749952e-05,
+ "loss": 0.0585,
+ "step": 23
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.997349913614761e-05,
+ "loss": 0.091,
+ "step": 24
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.996539130905593e-05,
+ "loss": 0.0351,
+ "step": 25
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.995620509085228e-05,
+ "loss": 0.0776,
+ "step": 26
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9945941475610623e-05,
+ "loss": 0.0447,
+ "step": 27
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.993460157399396e-05,
+ "loss": 0.0317,
+ "step": 28
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9922186613134152e-05,
+ "loss": 0.0532,
+ "step": 29
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9908697936499105e-05,
+ "loss": 0.0589,
+ "step": 30
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9894137003747404e-05,
+ "loss": 0.1019,
+ "step": 31
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.987850539057036e-05,
+ "loss": 0.0542,
+ "step": 32
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.986180478852149e-05,
+ "loss": 0.0688,
+ "step": 33
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.984403700483347e-05,
+ "loss": 0.0529,
+ "step": 34
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9825203962222573e-05,
+ "loss": 0.0781,
+ "step": 35
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9805307698680592e-05,
+ "loss": 0.0445,
+ "step": 36
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9784350367254322e-05,
+ "loss": 0.0474,
+ "step": 37
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.976233423581255e-05,
+ "loss": 0.0281,
+ "step": 38
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9739261686800662e-05,
+ "loss": 0.0791,
+ "step": 39
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.97151352169828e-05,
+ "loss": 0.0267,
+ "step": 40
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.968995743717171e-05,
+ "loss": 0.0292,
+ "step": 41
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9663731071946207e-05,
+ "loss": 0.0924,
+ "step": 42
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.963645895935632e-05,
+ "loss": 0.0382,
+ "step": 43
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9608144050616192e-05,
+ "loss": 0.0407,
+ "step": 44
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9578789409784727e-05,
+ "loss": 0.0279,
+ "step": 45
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.954839821343401e-05,
+ "loss": 0.031,
+ "step": 46
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.951697375030553e-05,
+ "loss": 0.0441,
+ "step": 47
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9484519420954356e-05,
+ "loss": 0.0429,
+ "step": 48
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9451038737381078e-05,
+ "loss": 0.038,
+ "step": 49
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.941653532265182e-05,
+ "loss": 0.0212,
+ "step": 50
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9381012910506146e-05,
+ "loss": 0.0713,
+ "step": 51
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.934447534495301e-05,
+ "loss": 0.0833,
+ "step": 52
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.930692657985482e-05,
+ "loss": 0.0168,
+ "step": 53
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.926837067849953e-05,
+ "loss": 0.0442,
+ "step": 54
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9228811813160972e-05,
+ "loss": 0.0174,
+ "step": 55
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9188254264647338e-05,
+ "loss": 0.0499,
+ "step": 56
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9146702421837952e-05,
+ "loss": 0.0431,
+ "step": 57
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.910416078120832e-05,
+ "loss": 0.023,
+ "step": 58
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.906063394634356e-05,
+ "loss": 0.0238,
+ "step": 59
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.901612662744024e-05,
+ "loss": 0.0139,
+ "step": 60
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8970643640796642e-05,
+ "loss": 0.0808,
+ "step": 61
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.89241899082916e-05,
+ "loss": 0.0155,
+ "step": 62
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.887677045685188e-05,
+ "loss": 0.0236,
+ "step": 63
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.882839041790818e-05,
+ "loss": 0.0387,
+ "step": 64
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.877905502683987e-05,
+ "loss": 0.0338,
+ "step": 65
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8728769622408423e-05,
+ "loss": 0.0755,
+ "step": 66
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8677539646179706e-05,
+ "loss": 0.0364,
+ "step": 67
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.862537064193513e-05,
+ "loss": 0.0475,
+ "step": 68
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8572268255071718e-05,
+ "loss": 0.0162,
+ "step": 69
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.851823823199122e-05,
+ "loss": 0.0373,
+ "step": 70
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.851823823199122e-05,
+ "loss": 0.0331,
+ "step": 71
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8463286419478256e-05,
+ "loss": 0.0221,
+ "step": 72
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8407418764067627e-05,
+ "loss": 0.0437,
+ "step": 73
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8407418764067627e-05,
+ "loss": 0.077,
+ "step": 74
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8350641311400813e-05,
+ "loss": 0.0175,
+ "step": 75
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8292960205571742e-05,
+ "loss": 0.0354,
+ "step": 76
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8234381688461943e-05,
+ "loss": 0.0093,
+ "step": 77
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.817491209906506e-05,
+ "loss": 0.0214,
+ "step": 78
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8114557872800906e-05,
+ "loss": 0.0172,
+ "step": 79
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8053325540819048e-05,
+ "loss": 0.0376,
+ "step": 80
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.799122172929206e-05,
+ "loss": 0.0288,
+ "step": 81
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7928253158698474e-05,
+ "loss": 0.0212,
+ "step": 82
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7864426643095537e-05,
+ "loss": 0.0628,
+ "step": 83
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7799749089381843e-05,
+ "loss": 0.0616,
+ "step": 84
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.773422749654988e-05,
+ "loss": 0.0661,
+ "step": 85
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7667868954928695e-05,
+ "loss": 0.0349,
+ "step": 86
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7600680645416583e-05,
+ "loss": 0.0488,
+ "step": 87
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7532669838704036e-05,
+ "loss": 0.0415,
+ "step": 88
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.746384389448694e-05,
+ "loss": 0.0232,
+ "step": 89
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.739421026067017e-05,
+ "loss": 0.0421,
+ "step": 90
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7323776472561625e-05,
+ "loss": 0.0205,
+ "step": 91
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7252550152056795e-05,
+ "loss": 0.0839,
+ "step": 92
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7180539006813973e-05,
+ "loss": 0.0184,
+ "step": 93
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.7107750829420177e-05,
+ "loss": 0.0409,
+ "step": 94
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.7034193496547903e-05,
+ "loss": 0.0343,
+ "step": 95
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6959874968102736e-05,
+ "loss": 0.0144,
+ "step": 96
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6884803286362e-05,
+ "loss": 0.0377,
+ "step": 97
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6808986575104464e-05,
+ "loss": 0.0562,
+ "step": 98
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6732433038731245e-05,
+ "loss": 0.0438,
+ "step": 99
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.665515096137797e-05,
+ "loss": 0.0106,
+ "step": 100
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.657714870601833e-05,
+ "loss": 0.0294,
+ "step": 101
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.649843471355909e-05,
+ "loss": 0.0284,
+ "step": 102
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.641901750192666e-05,
+ "loss": 0.0144,
+ "step": 103
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6338905665145352e-05,
+ "loss": 0.0094,
+ "step": 104
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6258107872407376e-05,
+ "loss": 0.0353,
+ "step": 105
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6176632867134738e-05,
+ "loss": 0.0529,
+ "step": 106
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.609448946603304e-05,
+ "loss": 0.0392,
+ "step": 107
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.601168655813745e-05,
+ "loss": 0.0646,
+ "step": 108
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.592823310385073e-05,
+ "loss": 0.0207,
+ "step": 109
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.584413813397364e-05,
+ "loss": 0.0133,
+ "step": 110
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5759410748727663e-05,
+ "loss": 0.0229,
+ "step": 111
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5674060116770234e-05,
+ "loss": 0.0408,
+ "step": 112
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5588095474202597e-05,
+ "loss": 0.0145,
+ "step": 113
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5501526123570277e-05,
+ "loss": 0.0389,
+ "step": 114
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5414361432856475e-05,
+ "loss": 0.063,
+ "step": 115
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.532661083446829e-05,
+ "loss": 0.0663,
+ "step": 116
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5238283824216015e-05,
+ "loss": 0.0387,
+ "step": 117
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.514938996028556e-05,
+ "loss": 0.0437,
+ "step": 118
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5059938862204126e-05,
+ "loss": 0.0204,
+ "step": 119
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4969940209799248e-05,
+ "loss": 0.0333,
+ "step": 120
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4879403742151283e-05,
+ "loss": 0.0429,
+ "step": 121
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4788339256539543e-05,
+ "loss": 0.0382,
+ "step": 122
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.469675660738206e-05,
+ "loss": 0.0408,
+ "step": 123
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4604665705169239e-05,
+ "loss": 0.0193,
+ "step": 124
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4512076515391375e-05,
+ "loss": 0.0293,
+ "step": 125
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4418999057460277e-05,
+ "loss": 0.0277,
+ "step": 126
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4325443403625012e-05,
+ "loss": 0.0423,
+ "step": 127
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4231419677881966e-05,
+ "loss": 0.0222,
+ "step": 128
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.4136938054879284e-05,
+ "loss": 0.0238,
+ "step": 129
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.404200875881582e-05,
+ "loss": 0.0364,
+ "step": 130
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3946642062334765e-05,
+ "loss": 0.0313,
+ "step": 131
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3850848285411994e-05,
+ "loss": 0.0484,
+ "step": 132
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3754637794239303e-05,
+ "loss": 0.0325,
+ "step": 133
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3658021000102638e-05,
+ "loss": 0.0396,
+ "step": 134
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.356100835825547e-05,
+ "loss": 0.0268,
+ "step": 135
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3463610366787392e-05,
+ "loss": 0.0143,
+ "step": 136
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3365837565488065e-05,
+ "loss": 0.012,
+ "step": 137
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.326770053470668e-05,
+ "loss": 0.0388,
+ "step": 138
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.316920989420703e-05,
+ "loss": 0.0152,
+ "step": 139
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3070376302018287e-05,
+ "loss": 0.0296,
+ "step": 140
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2971210453281675e-05,
+ "loss": 0.02,
+ "step": 141
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2871723079093101e-05,
+ "loss": 0.007,
+ "step": 142
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2771924945341906e-05,
+ "loss": 0.0375,
+ "step": 143
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2671826851545851e-05,
+ "loss": 0.0237,
+ "step": 144
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.257143962968246e-05,
+ "loss": 0.0162,
+ "step": 145
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2470774143016854e-05,
+ "loss": 0.0575,
+ "step": 146
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.236984128492619e-05,
+ "loss": 0.0142,
+ "step": 147
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2268651977720867e-05,
+ "loss": 0.0405,
+ "step": 148
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2167217171462566e-05,
+ "loss": 0.0531,
+ "step": 149
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.206554784277931e-05,
+ "loss": 0.0157,
+ "step": 150
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1963654993677645e-05,
+ "loss": 0.0332,
+ "step": 151
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1861549650352069e-05,
+ "loss": 0.0092,
+ "step": 152
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1759242861991855e-05,
+ "loss": 0.0278,
+ "step": 153
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1656745699585373e-05,
+ "loss": 0.0445,
+ "step": 154
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.155406925472205e-05,
+ "loss": 0.0546,
+ "step": 155
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.145122463839213e-05,
+ "loss": 0.0293,
+ "step": 156
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1348222979784289e-05,
+ "loss": 0.0078,
+ "step": 157
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1245075425081328e-05,
+ "loss": 0.0327,
+ "step": 158
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1141793136253987e-05,
+ "loss": 0.0459,
+ "step": 159
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1038387289853069e-05,
+ "loss": 0.0583,
+ "step": 160
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.09348690758e-05,
+ "loss": 0.0631,
+ "step": 161
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0831249696175918e-05,
+ "loss": 0.0197,
+ "step": 162
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.072754036400944e-05,
+ "loss": 0.0409,
+ "step": 163
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0623752302063284e-05,
+ "loss": 0.054,
+ "step": 164
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0519896741619803e-05,
+ "loss": 0.0105,
+ "step": 165
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.041598492126561e-05,
+ "loss": 0.0594,
+ "step": 166
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0312028085675393e-05,
+ "loss": 0.0402,
+ "step": 167
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0208037484395114e-05,
+ "loss": 0.0202,
+ "step": 168
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0104024370624644e-05,
+ "loss": 0.0433,
+ "step": 169
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1e-05,
+ "loss": 0.049,
+ "step": 170
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.89597562937536e-06,
+ "loss": 0.0529,
+ "step": 171
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.791962515604887e-06,
+ "loss": 0.02,
+ "step": 172
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.687971914324607e-06,
+ "loss": 0.0358,
+ "step": 173
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.584015078734395e-06,
+ "loss": 0.0268,
+ "step": 174
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.480103258380198e-06,
+ "loss": 0.0211,
+ "step": 175
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.376247697936719e-06,
+ "loss": 0.045,
+ "step": 176
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.272459635990563e-06,
+ "loss": 0.0487,
+ "step": 177
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.168750303824085e-06,
+ "loss": 0.0386,
+ "step": 178
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.065130924199998e-06,
+ "loss": 0.0332,
+ "step": 179
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.961612710146934e-06,
+ "loss": 0.0447,
+ "step": 180
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.858206863746018e-06,
+ "loss": 0.0215,
+ "step": 181
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.754924574918675e-06,
+ "loss": 0.0513,
+ "step": 182
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.651777020215713e-06,
+ "loss": 0.0232,
+ "step": 183
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.548775361607872e-06,
+ "loss": 0.0623,
+ "step": 184
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.445930745277953e-06,
+ "loss": 0.0338,
+ "step": 185
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.343254300414629e-06,
+ "loss": 0.0129,
+ "step": 186
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.240757138008149e-06,
+ "loss": 0.0535,
+ "step": 187
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.138450349647936e-06,
+ "loss": 0.047,
+ "step": 188
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.036345006322358e-06,
+ "loss": 0.0273,
+ "step": 189
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.934452157220693e-06,
+ "loss": 0.0069,
+ "step": 190
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.832782828537437e-06,
+ "loss": 0.0336,
+ "step": 191
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.731348022279135e-06,
+ "loss": 0.031,
+ "step": 192
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.630158715073813e-06,
+ "loss": 0.039,
+ "step": 193
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.529225856983151e-06,
+ "loss": 0.0368,
+ "step": 194
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.428560370317542e-06,
+ "loss": 0.0225,
+ "step": 195
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.328173148454151e-06,
+ "loss": 0.0558,
+ "step": 196
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.228075054658096e-06,
+ "loss": 0.034,
+ "step": 197
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.1282769209069005e-06,
+ "loss": 0.051,
+ "step": 198
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 7.028789546718327e-06,
+ "loss": 0.0714,
+ "step": 199
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.9296236979817175e-06,
+ "loss": 0.0525,
+ "step": 200
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.8307901057929735e-06,
+ "loss": 0.0178,
+ "step": 201
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.732299465293322e-06,
+ "loss": 0.0326,
+ "step": 202
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.634162434511939e-06,
+ "loss": 0.0276,
+ "step": 203
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.53638963321261e-06,
+ "loss": 0.0151,
+ "step": 204
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.438991641744531e-06,
+ "loss": 0.0449,
+ "step": 205
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.3419789998973655e-06,
+ "loss": 0.0273,
+ "step": 206
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.245362205760703e-06,
+ "loss": 0.0417,
+ "step": 207
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.149151714588009e-06,
+ "loss": 0.0112,
+ "step": 208
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 6.053357937665237e-06,
+ "loss": 0.0358,
+ "step": 209
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.957991241184184e-06,
+ "loss": 0.0559,
+ "step": 210
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.863061945120719e-06,
+ "loss": 0.0312,
+ "step": 211
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.768580322118034e-06,
+ "loss": 0.045,
+ "step": 212
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.674556596374993e-06,
+ "loss": 0.0118,
+ "step": 213
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.581000942539729e-06,
+ "loss": 0.0234,
+ "step": 214
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.487923484608629e-06,
+ "loss": 0.0143,
+ "step": 215
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.395334294830766e-06,
+ "loss": 0.016,
+ "step": 216
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.3032433926179395e-06,
+ "loss": 0.0164,
+ "step": 217
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.211660743460458e-06,
+ "loss": 0.0252,
+ "step": 218
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.120596257848716e-06,
+ "loss": 0.0069,
+ "step": 219
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.0300597902007565e-06,
+ "loss": 0.0417,
+ "step": 220
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.940061137795876e-06,
+ "loss": 0.0282,
+ "step": 221
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.850610039714444e-06,
+ "loss": 0.0277,
+ "step": 222
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.7617161757839895e-06,
+ "loss": 0.0161,
+ "step": 223
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.673389165531714e-06,
+ "loss": 0.0377,
+ "step": 224
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.5856385671435285e-06,
+ "loss": 0.0286,
+ "step": 225
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.498473876429727e-06,
+ "loss": 0.025,
+ "step": 226
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.411904525797408e-06,
+ "loss": 0.04,
+ "step": 227
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.3259398832297665e-06,
+ "loss": 0.0051,
+ "step": 228
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.240589251272342e-06,
+ "loss": 0.036,
+ "step": 229
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.155861866026364e-06,
+ "loss": 0.0291,
+ "step": 230
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 4.0717668961492725e-06,
+ "loss": 0.0459,
+ "step": 231
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.9883134418625535e-06,
+ "loss": 0.0438,
+ "step": 232
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.905510533966959e-06,
+ "loss": 0.0104,
+ "step": 233
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.823367132865266e-06,
+ "loss": 0.0972,
+ "step": 234
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.7418921275926245e-06,
+ "loss": 0.013,
+ "step": 235
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6610943348546524e-06,
+ "loss": 0.0647,
+ "step": 236
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.5809824980733445e-06,
+ "loss": 0.0306,
+ "step": 237
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.5015652864409142e-06,
+ "loss": 0.0271,
+ "step": 238
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.422851293981676e-06,
+ "loss": 0.0238,
+ "step": 239
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.3448490386220355e-06,
+ "loss": 0.0246,
+ "step": 240
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.2675669612687565e-06,
+ "loss": 0.0139,
+ "step": 241
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.191013424895536e-06,
+ "loss": 0.0159,
+ "step": 242
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.115196713638e-06,
+ "loss": 0.0308,
+ "step": 243
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 3.0401250318972643e-06,
+ "loss": 0.0287,
+ "step": 244
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.965806503452098e-06,
+ "loss": 0.0511,
+ "step": 245
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.892249170579826e-06,
+ "loss": 0.0481,
+ "step": 246
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.819460993186032e-06,
+ "loss": 0.0355,
+ "step": 247
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.7474498479432087e-06,
+ "loss": 0.0177,
+ "step": 248
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.6762235274383775e-06,
+ "loss": 0.0495,
+ "step": 249
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.6057897393298328e-06,
+ "loss": 0.0096,
+ "step": 250
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.5361561055130625e-06,
+ "loss": 0.0073,
+ "step": 251
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4673301612959653e-06,
+ "loss": 0.0309,
+ "step": 252
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.3993193545834182e-06,
+ "loss": 0.0416,
+ "step": 253
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.3321310450713066e-06,
+ "loss": 0.0327,
+ "step": 254
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.265772503450122e-06,
+ "loss": 0.0551,
+ "step": 255
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.2002509106181625e-06,
+ "loss": 0.0214,
+ "step": 256
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.1355733569044633e-06,
+ "loss": 0.0226,
+ "step": 257
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0717468413015285e-06,
+ "loss": 0.0623,
+ "step": 258
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.008778270707944e-06,
+ "loss": 0.0428,
+ "step": 259
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.946674459180955e-06,
+ "loss": 0.0054,
+ "step": 260
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8854421271990964e-06,
+ "loss": 0.0274,
+ "step": 261
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8250879009349398e-06,
+ "loss": 0.0495,
+ "step": 262
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.7656183115380577e-06,
+ "loss": 0.0432,
+ "step": 263
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.707039794428259e-06,
+ "loss": 0.01,
+ "step": 264
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6493586885991908e-06,
+ "loss": 0.0189,
+ "step": 265
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.5925812359323745e-06,
+ "loss": 0.0153,
+ "step": 266
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.536713580521746e-06,
+ "loss": 0.0123,
+ "step": 267
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4817617680087826e-06,
+ "loss": 0.0317,
+ "step": 268
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.4277317449282834e-06,
+ "loss": 0.0206,
+ "step": 269
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3746293580648718e-06,
+ "loss": 0.0647,
+ "step": 270
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3224603538202929e-06,
+ "loss": 0.0289,
+ "step": 271
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.2712303775915803e-06,
+ "loss": 0.0391,
+ "step": 272
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.220944973160133e-06,
+ "loss": 0.0286,
+ "step": 273
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1716095820918217e-06,
+ "loss": 0.0037,
+ "step": 274
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1232295431481222e-06,
+ "loss": 0.0317,
+ "step": 275
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.075810091708399e-06,
+ "loss": 0.0212,
+ "step": 276
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.0293563592033595e-06,
+ "loss": 0.0146,
+ "step": 277
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.838733725597615e-07,
+ "loss": 0.0321,
+ "step": 278
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 9.393660536564408e-07,
+ "loss": 0.0099,
+ "step": 279
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.958392187916842e-07,
+ "loss": 0.0283,
+ "step": 280
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.532975781620511e-07,
+ "loss": 0.0274,
+ "step": 281
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 8.117457353526626e-07,
+ "loss": 0.0246,
+ "step": 282
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.711881868390292e-07,
+ "loss": 0.0173,
+ "step": 283
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.316293215004689e-07,
+ "loss": 0.0156,
+ "step": 284
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.930734201451817e-07,
+ "loss": 0.0209,
+ "step": 285
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.555246550469907e-07,
+ "loss": 0.0054,
+ "step": 286
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.189870894938587e-07,
+ "loss": 0.0249,
+ "step": 287
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.834646773481811e-07,
+ "loss": 0.0084,
+ "step": 288
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.489612626189245e-07,
+ "loss": 0.0085,
+ "step": 289
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.154805790456486e-07,
+ "loss": 0.0282,
+ "step": 290
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.830262496944693e-07,
+ "loss": 0.0207,
+ "step": 291
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.5160178656599495e-07,
+ "loss": 0.0277,
+ "step": 292
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.21210590215273e-07,
+ "loss": 0.0155,
+ "step": 293
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.918559493838114e-07,
+ "loss": 0.0233,
+ "step": 294
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.635410406436857e-07,
+ "loss": 0.0204,
+ "step": 295
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.3626892805379565e-07,
+ "loss": 0.031,
+ "step": 296
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.100425628282899e-07,
+ "loss": 0.0142,
+ "step": 297
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.8486478301720246e-07,
+ "loss": 0.0369,
+ "step": 298
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.607383131993424e-07,
+ "loss": 0.0035,
+ "step": 299
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.3766576418745024e-07,
+ "loss": 0.0038,
+ "step": 300
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 2.1564963274568028e-07,
+ "loss": 0.0224,
+ "step": 301
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.9469230131940907e-07,
+ "loss": 0.0196,
+ "step": 302
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.7479603777742937e-07,
+ "loss": 0.0045,
+ "step": 303
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.559629951665298e-07,
+ "loss": 0.0312,
+ "step": 304
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.3819521147851122e-07,
+ "loss": 0.0204,
+ "step": 305
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.2149460942964097e-07,
+ "loss": 0.0438,
+ "step": 306
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 1.0586299625259699e-07,
+ "loss": 0.0589,
+ "step": 307
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 9.130206350089765e-08,
+ "loss": 0.0348,
+ "step": 308
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 7.781338686584928e-08,
+ "loss": 0.0103,
+ "step": 309
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 6.539842600603918e-08,
+ "loss": 0.0441,
+ "step": 310
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 5.405852438937764e-08,
+ "loss": 0.0365,
+ "step": 311
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 4.3794909147720776e-08,
+ "loss": 0.0303,
+ "step": 312
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 3.460869094407127e-08,
+ "loss": 0.0067,
+ "step": 313
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 2.6500863852395585e-08,
+ "loss": 0.0295,
+ "step": 314
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.947230525005006e-08,
+ "loss": 0.0609,
+ "step": 315
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.3523775722834586e-08,
+ "loss": 0.0322,
+ "step": 316
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 8.655918982689582e-09,
+ "loss": 0.0128,
+ "step": 317
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 4.869261798035041e-09,
+ "loss": 0.0368,
+ "step": 318
+ },
+ {
+ "epoch": 1.0,
+ "step": 318,
+ "total_flos": 3024316907520.0,
+ "train_loss": 0.03735237892383253,
+ "train_runtime": 2488.3642,
+ "train_samples_per_second": 2.55,
+ "train_steps_per_second": 0.128
+ }
+ ],
+ "logging_steps": 1.0,
+ "max_steps": 318,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 1,
+ "save_steps": 500,
+ "total_flos": 3024316907520.0,
+ "train_batch_size": 10,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/CheckGuard Models/wholeimage/drawer/finetune_lora_llava_mistral.sh b/CheckGuard Models/wholeimage/drawer/finetune_lora_llava_mistral.sh
new file mode 100644
index 0000000000000000000000000000000000000000..888d2b5cda5d3112b078b17ea22614940aedb94c
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/finetune_lora_llava_mistral.sh
@@ -0,0 +1,43 @@
+#!/bin/bash
+# Use first parameter as GPU IDs, default to "0,1,2,3" if not provided
+GPU_IDS=${1:-0,1,2,3}
+
+
+CUDA_VISIBLE_DEVICES=0,1,2,3 deepspeed --include localhost:"$GPU_IDS" --master_port 29606\
+ llava/train/train_mem.py \
+ --lora_enable True --lora_r 128 --lora_alpha 256 --mm_projector_lr 2e-5 \
+ --deepspeed ./scripts/zero3.json \
+ --model_name_or_path liuhaotian/llava-v1.6-mistral-7b \
+ --version mistral_instruct \
+ --data_path /home/larry5/project/LLaVA-1.6-ft/data/peft/drawer/drawer_dataset.json \
+ --image_folder /home/larry5/project/LLaVA-1.6-ft/data/data/ \
+ --vision_tower openai/clip-vit-large-patch14-336 \
+ --mm_projector_type mlp2x_gelu \
+ --mm_vision_select_layer -2 \
+ --mm_use_im_start_end False \
+ --mm_use_im_patch_token False \
+ --mm_patch_merge_type spatial_unpad \
+ --image_aspect_ratio anyres \
+ --group_by_modality_length False \
+ --bf16 False \
+ --fp16 True \
+ --output_dir /home/larry5/project/LLaVA-1.6-ft/scripts_peft/mistral/lora/llava-lora-mistral-r128a256/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model \
+ --num_train_epochs 1 \
+ --per_device_train_batch_size 10 \
+ --per_device_eval_batch_size 1 \
+ --gradient_accumulation_steps 1 \
+ --evaluation_strategy "no" \
+ --save_strategy "steps" \
+ --save_steps 500 \
+ --save_total_limit 5 \
+ --learning_rate 2e-5 \
+ --weight_decay 0. \
+ --warmup_ratio 0.05 \
+ --lr_scheduler_type "cosine" \
+ --logging_steps 1 \
+ --tf32 True \
+ --model_max_length 4096 \
+ --gradient_checkpointing True \
+ --dataloader_num_workers 4 \
+ --lazy_preprocess True \
+ --report_to wandb \
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/README.md b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..bdb138eee6972419f6d60676388b52fd99ec478e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/README.md
@@ -0,0 +1,202 @@
+---
+library_name: peft
+base_model: liuhaotian/llava-v1.6-mistral-7b
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.10.0
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/adapter_config.json b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..ef5bf065d7583dfd5a0290e6e9f081e6403f2d4b
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/adapter_config.json
@@ -0,0 +1,34 @@
+{
+ "alpha_pattern": {},
+ "auto_mapping": null,
+ "base_model_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "bias": "none",
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 256,
+ "lora_dropout": 0.05,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "r": 128,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "k_proj",
+ "v_proj",
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "up_proj",
+ "o_proj"
+ ],
+ "task_type": "CAUSAL_LM",
+ "use_dora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..25497c4816a2174676682a8129f97b8c8a90231a
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dcb7baa92bad870249687f6d37c2bf0e0e5528f3562bd85db4126f10d29e1f9d
+size 708924928
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/README.md b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..bdb138eee6972419f6d60676388b52fd99ec478e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/README.md
@@ -0,0 +1,202 @@
+---
+library_name: peft
+base_model: liuhaotian/llava-v1.6-mistral-7b
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.10.0
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_config.json b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..ef5bf065d7583dfd5a0290e6e9f081e6403f2d4b
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_config.json
@@ -0,0 +1,34 @@
+{
+ "alpha_pattern": {},
+ "auto_mapping": null,
+ "base_model_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "bias": "none",
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 256,
+ "lora_dropout": 0.05,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "r": 128,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "k_proj",
+ "v_proj",
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "up_proj",
+ "o_proj"
+ ],
+ "task_type": "CAUSAL_LM",
+ "use_dora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_model.safetensors b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..25270a0909f9f03f04ce33fc3df40a29ec431828
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d20cade4c1bb14c7fe64c27a174ee58185d85e1ebe2fa7e7b2d4e66c68d7535
+size 1417762896
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_model_states.pt b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_model_states.pt
new file mode 100644
index 0000000000000000000000000000000000000000..a9e77c89cd2c3e63da4ade9d98e52d10a0f0bcfc
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15855d27b1c9447e6de51958e65550ef87c450448a95db12ed37af1ce46b87ba
+size 632242
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_optim_states.pt b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f7072ef63ef2064f092621c6509723acfb4ff658
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/global_step500/zero_pp_rank_0_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa7042b6a63589c371d7b5f5ee99a3beda47557aed8679ad9a1449546b2ac6ef
+size 4504787266
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/latest b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/latest
new file mode 100644
index 0000000000000000000000000000000000000000..f0b47ce15fff9a01b2a416a473b2148085048a50
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/latest
@@ -0,0 +1 @@
+global_step500
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/rng_state.pth b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..30f630c64f92aad6ee0cf192e35f06b61fa08947
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5ab840621b891b2fa4886e04a2c52941f09c0bd67bac84d3289b54102d26c75
+size 14244
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/scheduler.pt b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..0245a633ad4dd2a8707d70d61e1b1f17491ae6cb
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:905d66842f5634a044a31e5cfed71d9d2c3ea3227bd786ae6077edd1f4d03a9d
+size 1064
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/special_tokens_map.json b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..14761dcf1466dc232bd41de9c21d4c617b15755e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/special_tokens_map.json
@@ -0,0 +1,24 @@
+{
+ "bos_token": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false
+ },
+ "eos_token": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false
+ },
+ "pad_token": "",
+ "unk_token": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false
+ }
+}
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer.model b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer.model
new file mode 100644
index 0000000000000000000000000000000000000000..8b443ef19c2a19acc3ac64fb9c3db4a72921dff6
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer.model
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
+size 493443
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer_config.json b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..d0ea5c3458cd84f0062b47fa0476bb328b3e208a
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/tokenizer_config.json
@@ -0,0 +1,44 @@
+{
+ "add_bos_token": true,
+ "add_eos_token": false,
+ "added_tokens_decoder": {
+ "0": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "1": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "2": {
+ "content": "",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ }
+ },
+ "additional_special_tokens": [],
+ "bos_token": "",
+ "chat_template": "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if message['role'] == 'user' %}{{ '[INST] ' + message['content'] + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ message['content'] + eos_token}}{% else %}{{ raise_exception('Only user and assistant roles are supported!') }}{% endif %}{% endfor %}",
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "",
+ "legacy": true,
+ "model_max_length": 4096,
+ "pad_token": "",
+ "padding_side": "right",
+ "sp_model_kwargs": {},
+ "spaces_between_special_tokens": false,
+ "tokenizer_class": "LlamaTokenizer",
+ "unk_token": "",
+ "use_default_system_prompt": false
+}
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/trainer_state.json b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..11ae466714107c5717841008a81d3575231b21dd
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/trainer_state.json
@@ -0,0 +1,3021 @@
+{
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.8445945945945946,
+ "eval_steps": 500,
+ "global_step": 500,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "epoch": 0.0,
+ "learning_rate": 6.666666666666667e-07,
+ "loss": 0.6778,
+ "step": 1
+ },
+ {
+ "epoch": 0.0,
+ "learning_rate": 1.3333333333333334e-06,
+ "loss": 0.7858,
+ "step": 2
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.0000000000000003e-06,
+ "loss": 0.637,
+ "step": 3
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.666666666666667e-06,
+ "loss": 0.8891,
+ "step": 4
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 3.3333333333333333e-06,
+ "loss": 0.5252,
+ "step": 5
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 4.000000000000001e-06,
+ "loss": 0.5716,
+ "step": 6
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 4.666666666666667e-06,
+ "loss": 0.405,
+ "step": 7
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 5.333333333333334e-06,
+ "loss": 0.3647,
+ "step": 8
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 6e-06,
+ "loss": 0.3804,
+ "step": 9
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 6.666666666666667e-06,
+ "loss": 0.3187,
+ "step": 10
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.333333333333333e-06,
+ "loss": 0.3995,
+ "step": 11
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 8.000000000000001e-06,
+ "loss": 0.1845,
+ "step": 12
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 8.666666666666668e-06,
+ "loss": 0.3313,
+ "step": 13
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 9.333333333333334e-06,
+ "loss": 0.3947,
+ "step": 14
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1e-05,
+ "loss": 0.2065,
+ "step": 15
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.0666666666666667e-05,
+ "loss": 0.3842,
+ "step": 16
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.1333333333333334e-05,
+ "loss": 0.4008,
+ "step": 17
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.2e-05,
+ "loss": 0.2834,
+ "step": 18
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.2666666666666667e-05,
+ "loss": 0.3042,
+ "step": 19
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.3333333333333333e-05,
+ "loss": 0.4071,
+ "step": 20
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.4e-05,
+ "loss": 0.2516,
+ "step": 21
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.4666666666666666e-05,
+ "loss": 0.3165,
+ "step": 22
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.5333333333333334e-05,
+ "loss": 0.2704,
+ "step": 23
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.6000000000000003e-05,
+ "loss": 0.3171,
+ "step": 24
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.6666666666666667e-05,
+ "loss": 0.5139,
+ "step": 25
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.7333333333333336e-05,
+ "loss": 0.3724,
+ "step": 26
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.8e-05,
+ "loss": 0.2179,
+ "step": 27
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.866666666666667e-05,
+ "loss": 0.4084,
+ "step": 28
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.9333333333333333e-05,
+ "loss": 0.3582,
+ "step": 29
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 2e-05,
+ "loss": 0.2471,
+ "step": 30
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.9999843758648253e-05,
+ "loss": 0.254,
+ "step": 31
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.9999375039475278e-05,
+ "loss": 0.3107,
+ "step": 32
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9998593857127736e-05,
+ "loss": 0.1689,
+ "step": 33
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9997500236016233e-05,
+ "loss": 0.3537,
+ "step": 34
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.999609421031453e-05,
+ "loss": 0.215,
+ "step": 35
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9994375823958504e-05,
+ "loss": 0.1335,
+ "step": 36
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.999234513064475e-05,
+ "loss": 0.6074,
+ "step": 37
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9990002193828923e-05,
+ "loss": 0.2102,
+ "step": 38
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.998734708672375e-05,
+ "loss": 0.1328,
+ "step": 39
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.998437989229673e-05,
+ "loss": 0.2783,
+ "step": 40
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9981100703267567e-05,
+ "loss": 0.1648,
+ "step": 41
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9977509622105233e-05,
+ "loss": 0.2885,
+ "step": 42
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9973606761024813e-05,
+ "loss": 0.241,
+ "step": 43
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9969392241983957e-05,
+ "loss": 0.2298,
+ "step": 44
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9964866196679105e-05,
+ "loss": 0.1629,
+ "step": 45
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9960028766541336e-05,
+ "loss": 0.1911,
+ "step": 46
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.995488010273198e-05,
+ "loss": 0.3749,
+ "step": 47
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.994942036613787e-05,
+ "loss": 0.4347,
+ "step": 48
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9943649727366335e-05,
+ "loss": 0.1651,
+ "step": 49
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9937568366739858e-05,
+ "loss": 0.476,
+ "step": 50
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9931176474290438e-05,
+ "loss": 0.269,
+ "step": 51
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9924474249753656e-05,
+ "loss": 0.6341,
+ "step": 52
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9917461902562435e-05,
+ "loss": 0.2207,
+ "step": 53
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9910139651840497e-05,
+ "loss": 0.26,
+ "step": 54
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.990250772639552e-05,
+ "loss": 0.1328,
+ "step": 55
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9894566364711965e-05,
+ "loss": 0.4922,
+ "step": 56
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.988631581494365e-05,
+ "loss": 0.2979,
+ "step": 57
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9877756334905983e-05,
+ "loss": 0.2875,
+ "step": 58
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9868888192067915e-05,
+ "loss": 0.2682,
+ "step": 59
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9859711663543573e-05,
+ "loss": 0.1769,
+ "step": 60
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9850227036083592e-05,
+ "loss": 0.2952,
+ "step": 61
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9840434606066182e-05,
+ "loss": 0.2048,
+ "step": 62
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.983033467948784e-05,
+ "loss": 0.2215,
+ "step": 63
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9819927571953804e-05,
+ "loss": 0.3016,
+ "step": 64
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9809213608668188e-05,
+ "loss": 0.4735,
+ "step": 65
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9798193124423804e-05,
+ "loss": 0.3743,
+ "step": 66
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.978686646359173e-05,
+ "loss": 0.229,
+ "step": 67
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9775233980110524e-05,
+ "loss": 0.2431,
+ "step": 68
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9763296037475174e-05,
+ "loss": 0.2257,
+ "step": 69
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9751053008725736e-05,
+ "loss": 0.1851,
+ "step": 70
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9738505276435692e-05,
+ "loss": 0.179,
+ "step": 71
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9725653232699962e-05,
+ "loss": 0.1604,
+ "step": 72
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9712497279122692e-05,
+ "loss": 0.3912,
+ "step": 73
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.969903782680467e-05,
+ "loss": 0.5239,
+ "step": 74
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.96852752963305e-05,
+ "loss": 0.284,
+ "step": 75
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.967121011775546e-05,
+ "loss": 0.2228,
+ "step": 76
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9656842730592046e-05,
+ "loss": 0.4633,
+ "step": 77
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9642173583796265e-05,
+ "loss": 0.4491,
+ "step": 78
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.962720313575358e-05,
+ "loss": 0.3252,
+ "step": 79
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.961193185426459e-05,
+ "loss": 0.175,
+ "step": 80
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9596360216530436e-05,
+ "loss": 0.2405,
+ "step": 81
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.958048870913786e-05,
+ "loss": 0.1445,
+ "step": 82
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9564317828044022e-05,
+ "loss": 0.1549,
+ "step": 83
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9547848078560975e-05,
+ "loss": 0.2074,
+ "step": 84
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9531079975339912e-05,
+ "loss": 0.375,
+ "step": 85
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9514014042355057e-05,
+ "loss": 0.2914,
+ "step": 86
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9496650812887293e-05,
+ "loss": 0.2202,
+ "step": 87
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9478990829507507e-05,
+ "loss": 0.1598,
+ "step": 88
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.946103464405964e-05,
+ "loss": 0.5804,
+ "step": 89
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9442782817643425e-05,
+ "loss": 0.125,
+ "step": 90
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9424235920596866e-05,
+ "loss": 0.338,
+ "step": 91
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9405394532478422e-05,
+ "loss": 0.3918,
+ "step": 92
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9386259242048883e-05,
+ "loss": 0.302,
+ "step": 93
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9366830647252974e-05,
+ "loss": 0.1556,
+ "step": 94
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9347109355200672e-05,
+ "loss": 0.2169,
+ "step": 95
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9327095982148258e-05,
+ "loss": 0.0737,
+ "step": 96
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9306791153479007e-05,
+ "loss": 0.2776,
+ "step": 97
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.928619550368371e-05,
+ "loss": 0.2158,
+ "step": 98
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9265309676340787e-05,
+ "loss": 0.1743,
+ "step": 99
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9244134324096223e-05,
+ "loss": 0.4233,
+ "step": 100
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9222670108643152e-05,
+ "loss": 0.1737,
+ "step": 101
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9200917700701176e-05,
+ "loss": 0.3509,
+ "step": 102
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9178877779995423e-05,
+ "loss": 0.1843,
+ "step": 103
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.915655103523529e-05,
+ "loss": 0.3164,
+ "step": 104
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9133938164092942e-05,
+ "loss": 0.3705,
+ "step": 105
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9111039873181478e-05,
+ "loss": 0.1795,
+ "step": 106
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.908785687803289e-05,
+ "loss": 0.2387,
+ "step": 107
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9064389903075676e-05,
+ "loss": 0.2459,
+ "step": 108
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.904063968161222e-05,
+ "loss": 0.3093,
+ "step": 109
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.901660695579585e-05,
+ "loss": 0.282,
+ "step": 110
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.899229247660769e-05,
+ "loss": 0.3662,
+ "step": 111
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8967697003833156e-05,
+ "loss": 0.2212,
+ "step": 112
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.894282130603823e-05,
+ "loss": 0.1693,
+ "step": 113
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8917666160545446e-05,
+ "loss": 0.2523,
+ "step": 114
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8892232353409582e-05,
+ "loss": 0.2582,
+ "step": 115
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8892232353409582e-05,
+ "loss": 0.227,
+ "step": 116
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8866520679393127e-05,
+ "loss": 0.1532,
+ "step": 117
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.884053194194142e-05,
+ "loss": 0.2189,
+ "step": 118
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8814266953157557e-05,
+ "loss": 0.147,
+ "step": 119
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8787726533777003e-05,
+ "loss": 0.2196,
+ "step": 120
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.876091151314196e-05,
+ "loss": 0.2105,
+ "step": 121
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8733822729175452e-05,
+ "loss": 0.118,
+ "step": 122
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8706461028355107e-05,
+ "loss": 0.2145,
+ "step": 123
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.867882726568676e-05,
+ "loss": 0.2689,
+ "step": 124
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.865092230467769e-05,
+ "loss": 0.1862,
+ "step": 125
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8622747017309676e-05,
+ "loss": 0.2517,
+ "step": 126
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8594302284011704e-05,
+ "loss": 0.2234,
+ "step": 127
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8565588993632488e-05,
+ "loss": 0.416,
+ "step": 128
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.85366080434127e-05,
+ "loss": 0.2848,
+ "step": 129
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8507360338956896e-05,
+ "loss": 0.2564,
+ "step": 130
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8477846794205258e-05,
+ "loss": 0.1887,
+ "step": 131
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.844806833140501e-05,
+ "loss": 0.2172,
+ "step": 132
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8418025881081612e-05,
+ "loss": 0.2342,
+ "step": 133
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8387720382009665e-05,
+ "loss": 0.3647,
+ "step": 134
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8357152781183606e-05,
+ "loss": 0.4555,
+ "step": 135
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.832632403378808e-05,
+ "loss": 0.7154,
+ "step": 136
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.829523510316813e-05,
+ "loss": 0.2239,
+ "step": 137
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8263886960799062e-05,
+ "loss": 0.2482,
+ "step": 138
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.82322805862561e-05,
+ "loss": 0.314,
+ "step": 139
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8200416967183785e-05,
+ "loss": 0.2708,
+ "step": 140
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8168297099265094e-05,
+ "loss": 0.2582,
+ "step": 141
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.813592198619035e-05,
+ "loss": 0.2136,
+ "step": 142
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.810329263962584e-05,
+ "loss": 0.2046,
+ "step": 143
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8070410079182198e-05,
+ "loss": 0.1413,
+ "step": 144
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.803727533238257e-05,
+ "loss": 0.254,
+ "step": 145
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8003889434630473e-05,
+ "loss": 0.3183,
+ "step": 146
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7970253429177477e-05,
+ "loss": 0.1788,
+ "step": 147
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.793636836709057e-05,
+ "loss": 0.1193,
+ "step": 148
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7902235307219333e-05,
+ "loss": 0.1632,
+ "step": 149
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7867855316162846e-05,
+ "loss": 0.2055,
+ "step": 150
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7833229468236367e-05,
+ "loss": 0.2053,
+ "step": 151
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7798358845437754e-05,
+ "loss": 0.1196,
+ "step": 152
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.776324453741365e-05,
+ "loss": 0.1903,
+ "step": 153
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.772788764142545e-05,
+ "loss": 0.35,
+ "step": 154
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7692289262315e-05,
+ "loss": 0.2117,
+ "step": 155
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.765645051247007e-05,
+ "loss": 0.2519,
+ "step": 156
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7620372511789607e-05,
+ "loss": 0.2019,
+ "step": 157
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7584056387648727e-05,
+ "loss": 0.16,
+ "step": 158
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.75475032748635e-05,
+ "loss": 0.1916,
+ "step": 159
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.751071431565547e-05,
+ "loss": 0.3202,
+ "step": 160
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.747369065961599e-05,
+ "loss": 0.3153,
+ "step": 161
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7436433463670262e-05,
+ "loss": 0.2454,
+ "step": 162
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7398943892041223e-05,
+ "loss": 0.1146,
+ "step": 163
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7361223116213143e-05,
+ "loss": 0.2135,
+ "step": 164
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7323272314895022e-05,
+ "loss": 0.2555,
+ "step": 165
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.728509267398376e-05,
+ "loss": 0.1648,
+ "step": 166
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7246685386527098e-05,
+ "loss": 0.2556,
+ "step": 167
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7208051652686335e-05,
+ "loss": 0.0675,
+ "step": 168
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7169192679698837e-05,
+ "loss": 0.1801,
+ "step": 169
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.713010968184029e-05,
+ "loss": 0.1639,
+ "step": 170
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7090803880386784e-05,
+ "loss": 0.1983,
+ "step": 171
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7051276503576623e-05,
+ "loss": 0.2065,
+ "step": 172
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.701152878657197e-05,
+ "loss": 0.386,
+ "step": 173
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.6971561971420225e-05,
+ "loss": 0.1026,
+ "step": 174
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.693137730701524e-05,
+ "loss": 0.141,
+ "step": 175
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6890976049058267e-05,
+ "loss": 0.3519,
+ "step": 176
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6850359460018737e-05,
+ "loss": 0.1873,
+ "step": 177
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6809528809094808e-05,
+ "loss": 0.2236,
+ "step": 178
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6768485372173696e-05,
+ "loss": 0.1955,
+ "step": 179
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6727230431791816e-05,
+ "loss": 0.2819,
+ "step": 180
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6685765277094702e-05,
+ "loss": 0.1513,
+ "step": 181
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6644091203796707e-05,
+ "loss": 0.1258,
+ "step": 182
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6602209514140552e-05,
+ "loss": 0.3084,
+ "step": 183
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.656012151685659e-05,
+ "loss": 0.1943,
+ "step": 184
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6517828527121942e-05,
+ "loss": 0.1087,
+ "step": 185
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6475331866519387e-05,
+ "loss": 0.3218,
+ "step": 186
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6432632862996056e-05,
+ "loss": 0.2016,
+ "step": 187
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6389732850821967e-05,
+ "loss": 0.2355,
+ "step": 188
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.634663317054829e-05,
+ "loss": 0.3003,
+ "step": 189
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6303335168965484e-05,
+ "loss": 0.2318,
+ "step": 190
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6259840199061215e-05,
+ "loss": 0.1513,
+ "step": 191
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6216149619978064e-05,
+ "loss": 0.1612,
+ "step": 192
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.617226479697105e-05,
+ "loss": 0.1565,
+ "step": 193
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.612818710136499e-05,
+ "loss": 0.3011,
+ "step": 194
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.608391791051163e-05,
+ "loss": 0.2011,
+ "step": 195
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6039458607746614e-05,
+ "loss": 0.1629,
+ "step": 196
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.599481058234626e-05,
+ "loss": 0.1287,
+ "step": 197
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.5949975229484132e-05,
+ "loss": 0.2725,
+ "step": 198
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5904953950187458e-05,
+ "loss": 0.4143,
+ "step": 199
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5859748151293347e-05,
+ "loss": 0.1918,
+ "step": 200
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5814359245404818e-05,
+ "loss": 0.2486,
+ "step": 201
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.576878865084668e-05,
+ "loss": 0.4387,
+ "step": 202
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5723037791621193e-05,
+ "loss": 0.2768,
+ "step": 203
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5677108097363565e-05,
+ "loss": 0.1003,
+ "step": 204
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.563100100329731e-05,
+ "loss": 0.4326,
+ "step": 205
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.558471795018936e-05,
+ "loss": 0.0864,
+ "step": 206
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5538260384305076e-05,
+ "loss": 0.3033,
+ "step": 207
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5491629757363033e-05,
+ "loss": 0.1332,
+ "step": 208
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5444827526489675e-05,
+ "loss": 0.193,
+ "step": 209
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.539785515417376e-05,
+ "loss": 0.2706,
+ "step": 210
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5350714108220673e-05,
+ "loss": 0.1861,
+ "step": 211
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5303405861706574e-05,
+ "loss": 0.3058,
+ "step": 212
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5255931892932333e-05,
+ "loss": 0.1898,
+ "step": 213
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5208293685377357e-05,
+ "loss": 0.2667,
+ "step": 214
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5160492727653241e-05,
+ "loss": 0.1723,
+ "step": 215
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5112530513457236e-05,
+ "loss": 0.2885,
+ "step": 216
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5064408541525573e-05,
+ "loss": 0.214,
+ "step": 217
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.501612831558664e-05,
+ "loss": 0.2457,
+ "step": 218
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.4967691344313995e-05,
+ "loss": 0.1488,
+ "step": 219
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.4919099141279203e-05,
+ "loss": 0.1468,
+ "step": 220
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.4870353224904572e-05,
+ "loss": 0.1331,
+ "step": 221
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4821455118415669e-05,
+ "loss": 0.3833,
+ "step": 222
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4772406349793744e-05,
+ "loss": 0.3083,
+ "step": 223
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4723208451727983e-05,
+ "loss": 0.5103,
+ "step": 224
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4673862961567602e-05,
+ "loss": 0.2315,
+ "step": 225
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4624371421273823e-05,
+ "loss": 0.5056,
+ "step": 226
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.457473537737167e-05,
+ "loss": 0.345,
+ "step": 227
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4524956380901669e-05,
+ "loss": 0.2486,
+ "step": 228
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4475035987371355e-05,
+ "loss": 0.1334,
+ "step": 229
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.442497575670668e-05,
+ "loss": 0.2026,
+ "step": 230
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4374777253203273e-05,
+ "loss": 0.2028,
+ "step": 231
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4324442045477536e-05,
+ "loss": 0.1024,
+ "step": 232
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4273971706417648e-05,
+ "loss": 0.1336,
+ "step": 233
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4223367813134412e-05,
+ "loss": 0.2927,
+ "step": 234
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4172631946911964e-05,
+ "loss": 0.3204,
+ "step": 235
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4121765693158364e-05,
+ "loss": 0.1417,
+ "step": 236
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.407077064135607e-05,
+ "loss": 0.2767,
+ "step": 237
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4019648385012243e-05,
+ "loss": 0.1877,
+ "step": 238
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.3968400521608969e-05,
+ "loss": 0.1902,
+ "step": 239
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3917028652553338e-05,
+ "loss": 0.1976,
+ "step": 240
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3865534383127406e-05,
+ "loss": 0.1658,
+ "step": 241
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3813919322438021e-05,
+ "loss": 0.4002,
+ "step": 242
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3762185083366557e-05,
+ "loss": 0.2596,
+ "step": 243
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3710333282518504e-05,
+ "loss": 0.0957,
+ "step": 244
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3658365540172948e-05,
+ "loss": 0.242,
+ "step": 245
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3606283480231957e-05,
+ "loss": 0.4266,
+ "step": 246
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3554088730169814e-05,
+ "loss": 0.3886,
+ "step": 247
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3501782920982185e-05,
+ "loss": 0.1089,
+ "step": 248
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3449367687135134e-05,
+ "loss": 0.3312,
+ "step": 249
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.339684466651406e-05,
+ "loss": 0.2238,
+ "step": 250
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.334421550037251e-05,
+ "loss": 0.1046,
+ "step": 251
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.334421550037251e-05,
+ "loss": 0.1617,
+ "step": 252
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3291481833280897e-05,
+ "loss": 0.0849,
+ "step": 253
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3238645313075104e-05,
+ "loss": 0.2548,
+ "step": 254
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3185707590805004e-05,
+ "loss": 0.1738,
+ "step": 255
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.313267032068285e-05,
+ "loss": 0.1744,
+ "step": 256
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3079535160031598e-05,
+ "loss": 0.3275,
+ "step": 257
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3026303769233112e-05,
+ "loss": 0.2187,
+ "step": 258
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2972977811676286e-05,
+ "loss": 0.13,
+ "step": 259
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2919558953705055e-05,
+ "loss": 0.1644,
+ "step": 260
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2866048864566338e-05,
+ "loss": 0.3441,
+ "step": 261
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2812449216357863e-05,
+ "loss": 0.1805,
+ "step": 262
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.275876168397593e-05,
+ "loss": 0.1578,
+ "step": 263
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.270498794506307e-05,
+ "loss": 0.4781,
+ "step": 264
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2651129679955604e-05,
+ "loss": 0.1001,
+ "step": 265
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.259718857163117e-05,
+ "loss": 0.225,
+ "step": 266
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2543166305656099e-05,
+ "loss": 0.216,
+ "step": 267
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2489064570132764e-05,
+ "loss": 0.3636,
+ "step": 268
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2434885055646823e-05,
+ "loss": 0.3873,
+ "step": 269
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2380629455214392e-05,
+ "loss": 0.1739,
+ "step": 270
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2326299464229143e-05,
+ "loss": 0.2209,
+ "step": 271
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2271896780409321e-05,
+ "loss": 0.1852,
+ "step": 272
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2217423103744694e-05,
+ "loss": 0.206,
+ "step": 273
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2162880136443447e-05,
+ "loss": 0.073,
+ "step": 274
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2108269582878957e-05,
+ "loss": 0.3641,
+ "step": 275
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2053593149536576e-05,
+ "loss": 0.1036,
+ "step": 276
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1998852544960266e-05,
+ "loss": 0.1654,
+ "step": 277
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1944049479699244e-05,
+ "loss": 0.4466,
+ "step": 278
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1889185666254505e-05,
+ "loss": 0.1723,
+ "step": 279
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1834262819025326e-05,
+ "loss": 0.11,
+ "step": 280
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1779282654255685e-05,
+ "loss": 0.1551,
+ "step": 281
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1724246889980638e-05,
+ "loss": 0.3681,
+ "step": 282
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.166915724597262e-05,
+ "loss": 0.1615,
+ "step": 283
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1614015443687723e-05,
+ "loss": 0.1501,
+ "step": 284
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1558823206211894e-05,
+ "loss": 0.1206,
+ "step": 285
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.150358225820709e-05,
+ "loss": 0.195,
+ "step": 286
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1448294325857387e-05,
+ "loss": 0.0672,
+ "step": 287
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1392961136815046e-05,
+ "loss": 0.1577,
+ "step": 288
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.133758442014651e-05,
+ "loss": 0.4435,
+ "step": 289
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1282165906278402e-05,
+ "loss": 0.249,
+ "step": 290
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.122670732694342e-05,
+ "loss": 0.2221,
+ "step": 291
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1171210415126248e-05,
+ "loss": 0.1312,
+ "step": 292
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1115676905009385e-05,
+ "loss": 0.15,
+ "step": 293
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1060108531918972e-05,
+ "loss": 0.1346,
+ "step": 294
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1004507032270553e-05,
+ "loss": 0.2224,
+ "step": 295
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.094887414351482e-05,
+ "loss": 0.2183,
+ "step": 296
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.0893211604083325e-05,
+ "loss": 0.154,
+ "step": 297
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.0837521153334143e-05,
+ "loss": 0.1895,
+ "step": 298
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.078180453149754e-05,
+ "loss": 0.3471,
+ "step": 299
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0726063479621574e-05,
+ "loss": 0.3142,
+ "step": 300
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.067029973951771e-05,
+ "loss": 0.1031,
+ "step": 301
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0614515053706367e-05,
+ "loss": 0.1477,
+ "step": 302
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0558711165362491e-05,
+ "loss": 0.1898,
+ "step": 303
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0502889818261075e-05,
+ "loss": 0.3198,
+ "step": 304
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.044705275672266e-05,
+ "loss": 0.3473,
+ "step": 305
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0391201725558842e-05,
+ "loss": 0.164,
+ "step": 306
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0335338470017742e-05,
+ "loss": 0.2804,
+ "step": 307
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0279464735729472e-05,
+ "loss": 0.1787,
+ "step": 308
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0223582268651585e-05,
+ "loss": 0.174,
+ "step": 309
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0167692815014527e-05,
+ "loss": 0.204,
+ "step": 310
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0111798121267047e-05,
+ "loss": 0.1659,
+ "step": 311
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0055899934021649e-05,
+ "loss": 0.0851,
+ "step": 312
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1e-05,
+ "loss": 0.125,
+ "step": 313
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.944100065978351e-06,
+ "loss": 0.1399,
+ "step": 314
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.888201878732956e-06,
+ "loss": 0.1191,
+ "step": 315
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.832307184985475e-06,
+ "loss": 0.2573,
+ "step": 316
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.776417731348416e-06,
+ "loss": 0.1156,
+ "step": 317
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.720535264270529e-06,
+ "loss": 0.2918,
+ "step": 318
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.664661529982261e-06,
+ "loss": 0.5064,
+ "step": 319
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.60879827444116e-06,
+ "loss": 0.1789,
+ "step": 320
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.552947243277346e-06,
+ "loss": 0.2524,
+ "step": 321
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.497110181738928e-06,
+ "loss": 0.1238,
+ "step": 322
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.44128883463751e-06,
+ "loss": 0.3283,
+ "step": 323
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.385484946293636e-06,
+ "loss": 0.2177,
+ "step": 324
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.329700260482292e-06,
+ "loss": 0.2896,
+ "step": 325
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.273936520378428e-06,
+ "loss": 0.4432,
+ "step": 326
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.218195468502462e-06,
+ "loss": 0.1969,
+ "step": 327
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.16247884666586e-06,
+ "loss": 0.1486,
+ "step": 328
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.106788395916679e-06,
+ "loss": 0.3046,
+ "step": 329
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.051125856485183e-06,
+ "loss": 0.1931,
+ "step": 330
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.99549296772945e-06,
+ "loss": 0.1927,
+ "step": 331
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.939891468081033e-06,
+ "loss": 0.2417,
+ "step": 332
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.884323094990619e-06,
+ "loss": 0.2002,
+ "step": 333
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.828789584873754e-06,
+ "loss": 0.1437,
+ "step": 334
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.773292673056582e-06,
+ "loss": 0.2163,
+ "step": 335
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.717834093721598e-06,
+ "loss": 0.0948,
+ "step": 336
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.662415579853492e-06,
+ "loss": 0.1959,
+ "step": 337
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.607038863184957e-06,
+ "loss": 0.1448,
+ "step": 338
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.551705674142618e-06,
+ "loss": 0.1835,
+ "step": 339
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.496417741792912e-06,
+ "loss": 0.1655,
+ "step": 340
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.44117679378811e-06,
+ "loss": 0.1666,
+ "step": 341
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.385984556312282e-06,
+ "loss": 0.1771,
+ "step": 342
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.330842754027383e-06,
+ "loss": 0.1141,
+ "step": 343
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.275753110019367e-06,
+ "loss": 0.3248,
+ "step": 344
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.220717345744316e-06,
+ "loss": 0.2598,
+ "step": 345
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.165737180974678e-06,
+ "loss": 0.1587,
+ "step": 346
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.110814333745496e-06,
+ "loss": 0.1711,
+ "step": 347
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.05595052030076e-06,
+ "loss": 0.1615,
+ "step": 348
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.001147455039735e-06,
+ "loss": 0.1857,
+ "step": 349
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.94640685046343e-06,
+ "loss": 0.1025,
+ "step": 350
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.891730417121044e-06,
+ "loss": 0.1696,
+ "step": 351
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.837119863556554e-06,
+ "loss": 0.1765,
+ "step": 352
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.782576896255307e-06,
+ "loss": 0.3,
+ "step": 353
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.72810321959068e-06,
+ "loss": 0.1344,
+ "step": 354
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.673700535770859e-06,
+ "loss": 0.3158,
+ "step": 355
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.619370544785608e-06,
+ "loss": 0.1455,
+ "step": 356
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.56511494435318e-06,
+ "loss": 0.2933,
+ "step": 357
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.510935429867237e-06,
+ "loss": 0.0901,
+ "step": 358
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.4568336943439055e-06,
+ "loss": 0.2275,
+ "step": 359
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.402811428368832e-06,
+ "loss": 0.1897,
+ "step": 360
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.348870320044399e-06,
+ "loss": 0.2391,
+ "step": 361
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.295012054936934e-06,
+ "loss": 0.1899,
+ "step": 362
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.241238316024069e-06,
+ "loss": 0.1385,
+ "step": 363
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.187550783642141e-06,
+ "loss": 0.0762,
+ "step": 364
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.133951135433666e-06,
+ "loss": 0.2305,
+ "step": 365
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.080441046294948e-06,
+ "loss": 0.1229,
+ "step": 366
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.027022188323716e-06,
+ "loss": 0.1246,
+ "step": 367
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 6.973696230766891e-06,
+ "loss": 0.2491,
+ "step": 368
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 6.920464839968405e-06,
+ "loss": 0.1749,
+ "step": 369
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 6.8673296793171555e-06,
+ "loss": 0.1952,
+ "step": 370
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.814292409194998e-06,
+ "loss": 0.115,
+ "step": 371
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.761354686924895e-06,
+ "loss": 0.1391,
+ "step": 372
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.708518166719107e-06,
+ "loss": 0.209,
+ "step": 373
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.655784499627491e-06,
+ "loss": 0.4016,
+ "step": 374
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.603155333485945e-06,
+ "loss": 0.1919,
+ "step": 375
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.550632312864869e-06,
+ "loss": 0.2142,
+ "step": 376
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.498217079017818e-06,
+ "loss": 0.1622,
+ "step": 377
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.445911269830189e-06,
+ "loss": 0.1669,
+ "step": 378
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.393716519768047e-06,
+ "loss": 0.3306,
+ "step": 379
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.341634459827053e-06,
+ "loss": 0.0774,
+ "step": 380
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.289666717481497e-06,
+ "loss": 0.2361,
+ "step": 381
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.237814916633444e-06,
+ "loss": 0.2916,
+ "step": 382
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.1860806775619785e-06,
+ "loss": 0.2664,
+ "step": 383
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.134465616872598e-06,
+ "loss": 0.4905,
+ "step": 384
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.082971347446662e-06,
+ "loss": 0.1202,
+ "step": 385
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.0315994783910345e-06,
+ "loss": 0.2228,
+ "step": 386
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 5.980351614987759e-06,
+ "loss": 0.1403,
+ "step": 387
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.929229358643932e-06,
+ "loss": 0.3658,
+ "step": 388
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.878234306841637e-06,
+ "loss": 0.121,
+ "step": 389
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.827368053088043e-06,
+ "loss": 0.2419,
+ "step": 390
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.7766321868655935e-06,
+ "loss": 0.1211,
+ "step": 391
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.726028293582355e-06,
+ "loss": 0.419,
+ "step": 392
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.67555795452247e-06,
+ "loss": 0.2181,
+ "step": 393
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.62522274679673e-06,
+ "loss": 0.1513,
+ "step": 394
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.575024243293319e-06,
+ "loss": 0.1522,
+ "step": 395
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.524964012628648e-06,
+ "loss": 0.1583,
+ "step": 396
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.475043619098334e-06,
+ "loss": 0.1475,
+ "step": 397
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.42526462262833e-06,
+ "loss": 0.1265,
+ "step": 398
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.375628578726181e-06,
+ "loss": 0.0715,
+ "step": 399
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.326137038432399e-06,
+ "loss": 0.1164,
+ "step": 400
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.276791548272018e-06,
+ "loss": 0.3881,
+ "step": 401
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.227593650206258e-06,
+ "loss": 0.1464,
+ "step": 402
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.1785448815843334e-06,
+ "loss": 0.2286,
+ "step": 403
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.129646775095432e-06,
+ "loss": 0.1454,
+ "step": 404
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.0809008587207965e-06,
+ "loss": 0.1155,
+ "step": 405
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.032308655686011e-06,
+ "loss": 0.1199,
+ "step": 406
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.983871684413363e-06,
+ "loss": 0.3385,
+ "step": 407
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.935591458474433e-06,
+ "loss": 0.2083,
+ "step": 408
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.8874694865427676e-06,
+ "loss": 0.1057,
+ "step": 409
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.8395072723467585e-06,
+ "loss": 0.1584,
+ "step": 410
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.791706314622645e-06,
+ "loss": 0.2643,
+ "step": 411
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.74406810706767e-06,
+ "loss": 0.1302,
+ "step": 412
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.69659413829343e-06,
+ "loss": 0.2024,
+ "step": 413
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.649285891779327e-06,
+ "loss": 0.1177,
+ "step": 414
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.602144845826246e-06,
+ "loss": 0.1947,
+ "step": 415
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.5551724735103285e-06,
+ "loss": 0.1209,
+ "step": 416
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.508370242636968e-06,
+ "loss": 0.2273,
+ "step": 417
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.461739615694929e-06,
+ "loss": 0.3131,
+ "step": 418
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.415282049810644e-06,
+ "loss": 0.41,
+ "step": 419
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.368998996702694e-06,
+ "loss": 0.218,
+ "step": 420
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.3228919026364345e-06,
+ "loss": 0.1857,
+ "step": 421
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.276962208378811e-06,
+ "loss": 0.1159,
+ "step": 422
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.231211349153319e-06,
+ "loss": 0.1981,
+ "step": 423
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.185640754595183e-06,
+ "loss": 0.2907,
+ "step": 424
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.140251848706656e-06,
+ "loss": 0.1582,
+ "step": 425
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.095046049812545e-06,
+ "loss": 0.1264,
+ "step": 426
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.050024770515869e-06,
+ "loss": 0.1817,
+ "step": 427
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.005189417653743e-06,
+ "loss": 0.1073,
+ "step": 428
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 3.960541392253387e-06,
+ "loss": 0.3221,
+ "step": 429
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.916082089488372e-06,
+ "loss": 0.2237,
+ "step": 430
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.8718128986350154e-06,
+ "loss": 0.3927,
+ "step": 431
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.827735203028953e-06,
+ "loss": 0.1443,
+ "step": 432
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.7838503800219393e-06,
+ "loss": 0.1289,
+ "step": 433
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.740159800938784e-06,
+ "loss": 0.1407,
+ "step": 434
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.696664831034519e-06,
+ "loss": 0.4103,
+ "step": 435
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6533668294517154e-06,
+ "loss": 0.1538,
+ "step": 436
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6102671491780393e-06,
+ "loss": 0.4277,
+ "step": 437
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.5673671370039464e-06,
+ "loss": 0.1458,
+ "step": 438
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.5246681334806177e-06,
+ "loss": 0.1699,
+ "step": 439
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.482171472878062e-06,
+ "loss": 0.2724,
+ "step": 440
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.4398784831434127e-06,
+ "loss": 0.2037,
+ "step": 441
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.39779048585945e-06,
+ "loss": 0.3123,
+ "step": 442
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.3559087962032956e-06,
+ "loss": 0.2008,
+ "step": 443
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.314234722905302e-06,
+ "loss": 0.253,
+ "step": 444
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.272769568208183e-06,
+ "loss": 0.1709,
+ "step": 445
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.2315146278263053e-06,
+ "loss": 0.2399,
+ "step": 446
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.1904711909051933e-06,
+ "loss": 0.1039,
+ "step": 447
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.149640539981267e-06,
+ "loss": 0.2212,
+ "step": 448
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.1090239509417364e-06,
+ "loss": 0.2057,
+ "step": 449
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0686226929847617e-06,
+ "loss": 0.1692,
+ "step": 450
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0284380285797767e-06,
+ "loss": 0.1658,
+ "step": 451
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 2.9884712134280324e-06,
+ "loss": 0.1372,
+ "step": 452
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.948723496423379e-06,
+ "loss": 0.1991,
+ "step": 453
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.909196119613218e-06,
+ "loss": 0.1983,
+ "step": 454
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.869890318159713e-06,
+ "loss": 0.4125,
+ "step": 455
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.8308073203011667e-06,
+ "loss": 0.4987,
+ "step": 456
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.7919483473136678e-06,
+ "loss": 0.1212,
+ "step": 457
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.753314613472906e-06,
+ "loss": 0.2399,
+ "step": 458
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.7149073260162416e-06,
+ "loss": 0.1534,
+ "step": 459
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.6767276851049818e-06,
+ "loss": 0.2789,
+ "step": 460
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.63877688378686e-06,
+ "loss": 0.1404,
+ "step": 461
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.6010561079587817e-06,
+ "loss": 0.1757,
+ "step": 462
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.5635665363297424e-06,
+ "loss": 0.1194,
+ "step": 463
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.5263093403840145e-06,
+ "loss": 0.1268,
+ "step": 464
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.489285684344532e-06,
+ "loss": 0.1493,
+ "step": 465
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.452496725136503e-06,
+ "loss": 0.2272,
+ "step": 466
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4159436123512737e-06,
+ "loss": 0.1863,
+ "step": 467
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.3796274882103964e-06,
+ "loss": 0.1783,
+ "step": 468
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.3435494875299315e-06,
+ "loss": 0.1379,
+ "step": 469
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.3077107376850005e-06,
+ "loss": 0.4607,
+ "step": 470
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.272112358574551e-06,
+ "loss": 0.078,
+ "step": 471
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.2367554625863496e-06,
+ "loss": 0.2215,
+ "step": 472
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.2016411545622497e-06,
+ "loss": 0.1333,
+ "step": 473
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.1667705317636333e-06,
+ "loss": 0.1098,
+ "step": 474
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.132144683837155e-06,
+ "loss": 0.1947,
+ "step": 475
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.0977646927806682e-06,
+ "loss": 0.1148,
+ "step": 476
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0636316329094317e-06,
+ "loss": 0.2097,
+ "step": 477
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.029746570822524e-06,
+ "loss": 0.127,
+ "step": 478
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.996110565369527e-06,
+ "loss": 0.2963,
+ "step": 479
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.9627246676174363e-06,
+ "loss": 0.172,
+ "step": 480
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.929589920817806e-06,
+ "loss": 0.2083,
+ "step": 481
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.896707360374167e-06,
+ "loss": 0.1132,
+ "step": 482
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8640780138096515e-06,
+ "loss": 0.2204,
+ "step": 483
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8317029007349086e-06,
+ "loss": 0.2599,
+ "step": 484
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.799583032816219e-06,
+ "loss": 0.2668,
+ "step": 485
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7677194137439036e-06,
+ "loss": 0.1575,
+ "step": 486
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7361130392009407e-06,
+ "loss": 0.2251,
+ "step": 487
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7047648968318697e-06,
+ "loss": 0.2407,
+ "step": 488
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6736759662119183e-06,
+ "loss": 0.1042,
+ "step": 489
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.642847218816398e-06,
+ "loss": 0.2138,
+ "step": 490
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6122796179903355e-06,
+ "loss": 0.185,
+ "step": 491
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.5819741189183902e-06,
+ "loss": 0.1251,
+ "step": 492
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.5519316685949903e-06,
+ "loss": 0.1887,
+ "step": 493
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.522153205794742e-06,
+ "loss": 0.1527,
+ "step": 494
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.492639661043106e-06,
+ "loss": 0.3192,
+ "step": 495
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4633919565873033e-06,
+ "loss": 0.2173,
+ "step": 496
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4344110063675143e-06,
+ "loss": 0.1351,
+ "step": 497
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4056977159883011e-06,
+ "loss": 0.1096,
+ "step": 498
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.377252982690327e-06,
+ "loss": 0.2903,
+ "step": 499
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.3490776953223107e-06,
+ "loss": 0.1286,
+ "step": 500
+ }
+ ],
+ "logging_steps": 1.0,
+ "max_steps": 592,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 1,
+ "save_steps": 500,
+ "total_flos": 1569118617600.0,
+ "train_batch_size": 10,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/training_args.bin b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..e2c71c8e66f03018064e6d1844aa66a28e44832d
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac20dc8b29eda10237fe52b517b0a3b28eb3b5986c5cee88e9da095ea6b8c41f
+size 6968
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/zero_to_fp32.py b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..c98caae31534368be22b67fc4ae906836c992a8d
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/checkpoint-500/zero_to_fp32.py
@@ -0,0 +1,587 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+ FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+ FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+ buffers: dict()
+ param_shapes: dict()
+ shared_params: list
+ ds_version: int
+ frozen_param_shapes: dict()
+ frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+ return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+ '''
+ alist.sort(key=natural_keys) sorts in human order
+ http://nedbatchelder.com/blog/200712/human_sorting.html
+ (See Toothy's implementation in the comments)
+ '''
+ return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+ if not os.path.isdir(checkpoint_dir):
+ raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+ # there should be only one file
+ if zero_stage <= 2:
+ file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+ elif zero_stage == 3:
+ file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+ if not os.path.exists(file):
+ raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+ return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+ # XXX: need to test that this simple glob rule works for multi-node setup too
+ ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+ if len(ckpt_files) == 0:
+ raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+ return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+ return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+ return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+ zero_model_states = []
+ for file in files:
+ state_dict = torch.load(file, map_location=device)
+
+ if BUFFER_NAMES not in state_dict:
+ raise ValueError(f"{file} is not a model state checkpoint")
+ buffer_names = state_dict[BUFFER_NAMES]
+ if debug:
+ print("Found buffers:", buffer_names)
+
+ # recover just the buffers while restoring them to fp32 if they were saved in fp16
+ buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+ param_shapes = state_dict[PARAM_SHAPES]
+
+ # collect parameters that are included in param_shapes
+ param_names = []
+ for s in param_shapes:
+ for name in s.keys():
+ param_names.append(name)
+
+ # update with frozen parameters
+ frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+ if frozen_param_shapes is not None:
+ if debug:
+ print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+ param_names += list(frozen_param_shapes.keys())
+
+ # handle shared params
+ shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+ ds_version = state_dict.get(DS_VERSION, None)
+
+ frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+ z_model_state = zero_model_state(buffers=buffers,
+ param_shapes=param_shapes,
+ shared_params=shared_params,
+ ds_version=ds_version,
+ frozen_param_shapes=frozen_param_shapes,
+ frozen_param_fragments=frozen_param_fragments)
+ zero_model_states.append(z_model_state)
+
+ return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+ total_files = len(files)
+ state_dicts = []
+ for f in files:
+ state_dict = torch.load(f, map_location=device)
+ # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+ # and also handle the case where it was already removed by another helper script
+ state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+ state_dicts.append(state_dict)
+
+ if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+ raise ValueError(f"{files[0]} is not a zero checkpoint")
+ zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+ world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+ # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+ # parameters can be different from data parallelism for non-expert parameters. So we can just
+ # use the max of the partition_count to get the dp world_size.
+
+ if type(world_size) is list:
+ world_size = max(world_size)
+
+ if world_size != total_files:
+ raise ValueError(
+ f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+ "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+ )
+
+ # the groups are named differently in each stage
+ if zero_stage <= 2:
+ fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+ elif zero_stage == 3:
+ fp32_groups_key = FP32_FLAT_GROUPS
+ else:
+ raise ValueError(f"unknown zero stage {zero_stage}")
+
+ if zero_stage <= 2:
+ fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+ elif zero_stage == 3:
+ # if there is more than one param group, there will be multiple flattened tensors - one
+ # flattened tensor per group - for simplicity merge them into a single tensor
+ #
+ # XXX: could make the script more memory efficient for when there are multiple groups - it
+ # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+ fp32_flat_groups = [
+ torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+ ]
+
+ return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir):
+ """
+ Returns fp32 state_dict reconstructed from ds checkpoint
+
+ Args:
+ - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+ """
+ print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+ optim_files = get_optim_files(ds_checkpoint_dir)
+ zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+ print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+ model_files = get_model_state_files(ds_checkpoint_dir)
+
+ zero_model_states = parse_model_states(model_files)
+ print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+ if zero_stage <= 2:
+ return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+ elif zero_stage == 3:
+ return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+ if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+ return
+
+ frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+ frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+ if debug:
+ num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+ print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+ wanted_params = len(frozen_param_shapes)
+ wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+ avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+ print(f'Frozen params: Have {avail_numel} numels to process.')
+ print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+ total_params = 0
+ total_numel = 0
+ for name, shape in frozen_param_shapes.items():
+ total_params += 1
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+
+ state_dict[name] = frozen_param_fragments[name]
+
+ if debug:
+ print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+ print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+ param_shapes = zero_model_states[0].param_shapes
+
+ # Reconstruction protocol:
+ #
+ # XXX: document this
+
+ if debug:
+ for i in range(world_size):
+ for j in range(len(fp32_flat_groups[0])):
+ print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+ # XXX: memory usage doubles here (zero2)
+ num_param_groups = len(fp32_flat_groups[0])
+ merged_single_partition_of_fp32_groups = []
+ for i in range(num_param_groups):
+ merged_partitions = [sd[i] for sd in fp32_flat_groups]
+ full_single_fp32_vector = torch.cat(merged_partitions, 0)
+ merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+ avail_numel = sum(
+ [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+ if debug:
+ wanted_params = sum([len(shapes) for shapes in param_shapes])
+ wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+ # not asserting if there is a mismatch due to possible padding
+ print(f"Have {avail_numel} numels to process.")
+ print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+ # params
+ # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+ # out-of-core computing solution
+ total_numel = 0
+ total_params = 0
+ for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+ offset = 0
+ avail_numel = full_single_fp32_vector.numel()
+ for name, shape in shapes.items():
+
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+ total_params += 1
+
+ if debug:
+ print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+ state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+ offset += unpartitioned_numel
+
+ # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+ # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+ # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+ # live optimizer object, so we are checking that the numbers are within the right range
+ align_to = 2 * world_size
+
+ def zero2_align(x):
+ return align_to * math.ceil(x / align_to)
+
+ if debug:
+ print(f"original offset={offset}, avail_numel={avail_numel}")
+
+ offset = zero2_align(offset)
+ avail_numel = zero2_align(avail_numel)
+
+ if debug:
+ print(f"aligned offset={offset}, avail_numel={avail_numel}")
+
+ # Sanity check
+ if offset != avail_numel:
+ raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+ print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+ state_dict = OrderedDict()
+
+ # buffers
+ buffers = zero_model_states[0].buffers
+ state_dict.update(buffers)
+ if debug:
+ print(f"added {len(buffers)} buffers")
+
+ _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+ _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+ # recover shared parameters
+ for pair in zero_model_states[0].shared_params:
+ if pair[1] in state_dict:
+ state_dict[pair[0]] = state_dict[pair[1]]
+
+ return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+ remainder = unpartitioned_numel % world_size
+ padding_numel = (world_size - remainder) if remainder else 0
+ partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+ return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+ if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+ return
+
+ if debug:
+ for i in range(world_size):
+ num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+ print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+ frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+ wanted_params = len(frozen_param_shapes)
+ wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+ avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+ print(f'Frozen params: Have {avail_numel} numels to process.')
+ print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+ total_params = 0
+ total_numel = 0
+ for name, shape in zero_model_states[0].frozen_param_shapes.items():
+ total_params += 1
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+
+ param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+ state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+ partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+ if debug:
+ print(
+ f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+ )
+
+ print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+ param_shapes = zero_model_states[0].param_shapes
+ avail_numel = fp32_flat_groups[0].numel() * world_size
+ # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+ # param, re-consolidating each param, while dealing with padding if any
+
+ # merge list of dicts, preserving order
+ param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+ if debug:
+ for i in range(world_size):
+ print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+ wanted_params = len(param_shapes)
+ wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+ # not asserting if there is a mismatch due to possible padding
+ avail_numel = fp32_flat_groups[0].numel() * world_size
+ print(f"Trainable params: Have {avail_numel} numels to process.")
+ print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+ # params
+ # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+ # out-of-core computing solution
+ offset = 0
+ total_numel = 0
+ total_params = 0
+ for name, shape in param_shapes.items():
+
+ unpartitioned_numel = shape.numel()
+ total_numel += unpartitioned_numel
+ total_params += 1
+
+ partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+ if debug:
+ print(
+ f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+ )
+
+ # XXX: memory usage doubles here
+ state_dict[name] = torch.cat(
+ tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+ 0).narrow(0, 0, unpartitioned_numel).view(shape)
+ offset += partitioned_numel
+
+ offset *= world_size
+
+ # Sanity check
+ if offset != avail_numel:
+ raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+ print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+ state_dict = OrderedDict()
+
+ # buffers
+ buffers = zero_model_states[0].buffers
+ state_dict.update(buffers)
+ if debug:
+ print(f"added {len(buffers)} buffers")
+
+ _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+ _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+ # recover shared parameters
+ for pair in zero_model_states[0].shared_params:
+ if pair[1] in state_dict:
+ state_dict[pair[0]] = state_dict[pair[1]]
+
+ return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None):
+ """
+ Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+ ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+ via a model hub.
+
+ Args:
+ - ``checkpoint_dir``: path to the desired checkpoint folder
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+
+ Returns:
+ - pytorch ``state_dict``
+
+ Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+ you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+ the checkpoint.
+
+ A typical usage might be ::
+
+ from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+ # do the training and checkpoint saving
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+ model = model.cpu() # move to cpu
+ model.load_state_dict(state_dict)
+ # submit to model hub or save the model to share with others
+
+ In this example the ``model`` will no longer be usable in the deepspeed context of the same
+ application. i.e. you will need to re-initialize the deepspeed engine, since
+ ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+ If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+ """
+ if tag is None:
+ latest_path = os.path.join(checkpoint_dir, 'latest')
+ if os.path.isfile(latest_path):
+ with open(latest_path, 'r') as fd:
+ tag = fd.read().strip()
+ else:
+ raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+ ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+ if not os.path.isdir(ds_checkpoint_dir):
+ raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+ return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None):
+ """
+ Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+ loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+ Args:
+ - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+ - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+ """
+
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+ print(f"Saving fp32 state dict to {output_file}")
+ torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+ """
+ 1. Put the provided model to cpu
+ 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+ 3. Load it into the provided model
+
+ Args:
+ - ``model``: the model object to update
+ - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+ Returns:
+ - ``model`: modified model
+
+ Make sure you have plenty of CPU memory available before you call this function. If you don't
+ have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+ conveniently placed for you in the checkpoint folder.
+
+ A typical usage might be ::
+
+ from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+ model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+ # submit to model hub or save the model to share with others
+
+ Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+ of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+ ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+ """
+ logger.info(f"Extracting fp32 weights")
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+ logger.info(f"Overwriting model with fp32 weights")
+ model = model.cpu()
+ model.load_state_dict(state_dict, strict=False)
+
+ return model
+
+
+if __name__ == "__main__":
+
+ parser = argparse.ArgumentParser()
+ parser.add_argument("checkpoint_dir",
+ type=str,
+ help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+ parser.add_argument(
+ "output_file",
+ type=str,
+ help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+ parser.add_argument("-t",
+ "--tag",
+ type=str,
+ default=None,
+ help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+ parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+ args = parser.parse_args()
+
+ debug = args.debug
+
+ convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, args.output_file, tag=args.tag)
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/config.json b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..93e133af45036a778791b5679a8953a4f6a35a33
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/config.json
@@ -0,0 +1,70 @@
+{
+ "_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "architectures": [
+ "LlavaMistralForCausalLM"
+ ],
+ "attention_dropout": 0.0,
+ "bos_token_id": 1,
+ "eos_token_id": 2,
+ "freeze_mm_mlp_adapter": false,
+ "freeze_mm_vision_resampler": false,
+ "hidden_act": "silu",
+ "hidden_size": 4096,
+ "image_aspect_ratio": "anyres",
+ "image_crop_resolution": 224,
+ "image_grid_pinpoints": [
+ [
+ 336,
+ 672
+ ],
+ [
+ 672,
+ 336
+ ],
+ [
+ 672,
+ 672
+ ],
+ [
+ 1008,
+ 336
+ ],
+ [
+ 336,
+ 1008
+ ]
+ ],
+ "image_split_resolution": 224,
+ "initializer_range": 0.02,
+ "intermediate_size": 14336,
+ "max_position_embeddings": 32768,
+ "mm_hidden_size": 1024,
+ "mm_patch_merge_type": "spatial_unpad",
+ "mm_projector_lr": 2e-05,
+ "mm_projector_type": "mlp2x_gelu",
+ "mm_resampler_type": null,
+ "mm_use_im_patch_token": false,
+ "mm_use_im_start_end": false,
+ "mm_vision_select_feature": "patch",
+ "mm_vision_select_layer": -2,
+ "mm_vision_tower": "openai/clip-vit-large-patch14-336",
+ "mm_vision_tower_lr": 2e-06,
+ "model_type": "llava_mistral",
+ "num_attention_heads": 32,
+ "num_hidden_layers": 32,
+ "num_key_value_heads": 8,
+ "rms_norm_eps": 1e-05,
+ "rope_theta": 1000000.0,
+ "sliding_window": null,
+ "tie_word_embeddings": false,
+ "tokenizer_model_max_length": 4096,
+ "tokenizer_padding_side": "right",
+ "torch_dtype": "bfloat16",
+ "transformers_version": "4.37.2",
+ "tune_mm_mlp_adapter": false,
+ "tune_mm_vision_resampler": false,
+ "unfreeze_mm_vision_tower": true,
+ "use_cache": true,
+ "use_mm_proj": true,
+ "vocab_size": 32000
+}
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin
new file mode 100644
index 0000000000000000000000000000000000000000..1ddf0ac9d3c1f362fdfdb32c168b26967a7c3e38
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50e91c6e79b848a7c5e717c26bd45fad3beda6f3c54739ee912dd5aa12b5b123
+size 41961648
diff --git a/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/trainer_state.json b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..d37c1828403b320337f5f2d3bcf018cf1cbfcd73
--- /dev/null
+++ b/CheckGuard Models/wholeimage/drawer/llava-lora-mistral-r128a256-10BS-model/trainer_state.json
@@ -0,0 +1,3582 @@
+{
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.0,
+ "eval_steps": 500,
+ "global_step": 592,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "epoch": 0.0,
+ "learning_rate": 6.666666666666667e-07,
+ "loss": 0.6778,
+ "step": 1
+ },
+ {
+ "epoch": 0.0,
+ "learning_rate": 1.3333333333333334e-06,
+ "loss": 0.7858,
+ "step": 2
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.0000000000000003e-06,
+ "loss": 0.637,
+ "step": 3
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 2.666666666666667e-06,
+ "loss": 0.8891,
+ "step": 4
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 3.3333333333333333e-06,
+ "loss": 0.5252,
+ "step": 5
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 4.000000000000001e-06,
+ "loss": 0.5716,
+ "step": 6
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 4.666666666666667e-06,
+ "loss": 0.405,
+ "step": 7
+ },
+ {
+ "epoch": 0.01,
+ "learning_rate": 5.333333333333334e-06,
+ "loss": 0.3647,
+ "step": 8
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 6e-06,
+ "loss": 0.3804,
+ "step": 9
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 6.666666666666667e-06,
+ "loss": 0.3187,
+ "step": 10
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 7.333333333333333e-06,
+ "loss": 0.3995,
+ "step": 11
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 8.000000000000001e-06,
+ "loss": 0.1845,
+ "step": 12
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 8.666666666666668e-06,
+ "loss": 0.3313,
+ "step": 13
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 9.333333333333334e-06,
+ "loss": 0.3947,
+ "step": 14
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1e-05,
+ "loss": 0.2065,
+ "step": 15
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.0666666666666667e-05,
+ "loss": 0.3842,
+ "step": 16
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.1333333333333334e-05,
+ "loss": 0.4008,
+ "step": 17
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.2e-05,
+ "loss": 0.2834,
+ "step": 18
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.2666666666666667e-05,
+ "loss": 0.3042,
+ "step": 19
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1.3333333333333333e-05,
+ "loss": 0.4071,
+ "step": 20
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.4e-05,
+ "loss": 0.2516,
+ "step": 21
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.4666666666666666e-05,
+ "loss": 0.3165,
+ "step": 22
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.5333333333333334e-05,
+ "loss": 0.2704,
+ "step": 23
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.6000000000000003e-05,
+ "loss": 0.3171,
+ "step": 24
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.6666666666666667e-05,
+ "loss": 0.5139,
+ "step": 25
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.7333333333333336e-05,
+ "loss": 0.3724,
+ "step": 26
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.8e-05,
+ "loss": 0.2179,
+ "step": 27
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.866666666666667e-05,
+ "loss": 0.4084,
+ "step": 28
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.9333333333333333e-05,
+ "loss": 0.3582,
+ "step": 29
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 2e-05,
+ "loss": 0.2471,
+ "step": 30
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.9999843758648253e-05,
+ "loss": 0.254,
+ "step": 31
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.9999375039475278e-05,
+ "loss": 0.3107,
+ "step": 32
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9998593857127736e-05,
+ "loss": 0.1689,
+ "step": 33
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9997500236016233e-05,
+ "loss": 0.3537,
+ "step": 34
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.999609421031453e-05,
+ "loss": 0.215,
+ "step": 35
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9994375823958504e-05,
+ "loss": 0.1335,
+ "step": 36
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.999234513064475e-05,
+ "loss": 0.6074,
+ "step": 37
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9990002193828923e-05,
+ "loss": 0.2102,
+ "step": 38
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.998734708672375e-05,
+ "loss": 0.1328,
+ "step": 39
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.998437989229673e-05,
+ "loss": 0.2783,
+ "step": 40
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9981100703267567e-05,
+ "loss": 0.1648,
+ "step": 41
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9977509622105233e-05,
+ "loss": 0.2885,
+ "step": 42
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9973606761024813e-05,
+ "loss": 0.241,
+ "step": 43
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.9969392241983957e-05,
+ "loss": 0.2298,
+ "step": 44
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9964866196679105e-05,
+ "loss": 0.1629,
+ "step": 45
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9960028766541336e-05,
+ "loss": 0.1911,
+ "step": 46
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.995488010273198e-05,
+ "loss": 0.3749,
+ "step": 47
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.994942036613787e-05,
+ "loss": 0.4347,
+ "step": 48
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9943649727366335e-05,
+ "loss": 0.1651,
+ "step": 49
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9937568366739858e-05,
+ "loss": 0.476,
+ "step": 50
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9931176474290438e-05,
+ "loss": 0.269,
+ "step": 51
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9924474249753656e-05,
+ "loss": 0.6341,
+ "step": 52
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9917461902562435e-05,
+ "loss": 0.2207,
+ "step": 53
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9910139651840497e-05,
+ "loss": 0.26,
+ "step": 54
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.990250772639552e-05,
+ "loss": 0.1328,
+ "step": 55
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.9894566364711965e-05,
+ "loss": 0.4922,
+ "step": 56
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.988631581494365e-05,
+ "loss": 0.2979,
+ "step": 57
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9877756334905983e-05,
+ "loss": 0.2875,
+ "step": 58
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9868888192067915e-05,
+ "loss": 0.2682,
+ "step": 59
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9859711663543573e-05,
+ "loss": 0.1769,
+ "step": 60
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9850227036083592e-05,
+ "loss": 0.2952,
+ "step": 61
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9840434606066182e-05,
+ "loss": 0.2048,
+ "step": 62
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.983033467948784e-05,
+ "loss": 0.2215,
+ "step": 63
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9819927571953804e-05,
+ "loss": 0.3016,
+ "step": 64
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9809213608668188e-05,
+ "loss": 0.4735,
+ "step": 65
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9798193124423804e-05,
+ "loss": 0.3743,
+ "step": 66
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.978686646359173e-05,
+ "loss": 0.229,
+ "step": 67
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9775233980110524e-05,
+ "loss": 0.2431,
+ "step": 68
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9763296037475174e-05,
+ "loss": 0.2257,
+ "step": 69
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9751053008725736e-05,
+ "loss": 0.1851,
+ "step": 70
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9738505276435692e-05,
+ "loss": 0.179,
+ "step": 71
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9725653232699962e-05,
+ "loss": 0.1604,
+ "step": 72
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9712497279122692e-05,
+ "loss": 0.3912,
+ "step": 73
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.969903782680467e-05,
+ "loss": 0.5239,
+ "step": 74
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.96852752963305e-05,
+ "loss": 0.284,
+ "step": 75
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.967121011775546e-05,
+ "loss": 0.2228,
+ "step": 76
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9656842730592046e-05,
+ "loss": 0.4633,
+ "step": 77
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9642173583796265e-05,
+ "loss": 0.4491,
+ "step": 78
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.962720313575358e-05,
+ "loss": 0.3252,
+ "step": 79
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.961193185426459e-05,
+ "loss": 0.175,
+ "step": 80
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9596360216530436e-05,
+ "loss": 0.2405,
+ "step": 81
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.958048870913786e-05,
+ "loss": 0.1445,
+ "step": 82
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9564317828044022e-05,
+ "loss": 0.1549,
+ "step": 83
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9547848078560975e-05,
+ "loss": 0.2074,
+ "step": 84
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.9531079975339912e-05,
+ "loss": 0.375,
+ "step": 85
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9514014042355057e-05,
+ "loss": 0.2914,
+ "step": 86
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9496650812887293e-05,
+ "loss": 0.2202,
+ "step": 87
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9478990829507507e-05,
+ "loss": 0.1598,
+ "step": 88
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.946103464405964e-05,
+ "loss": 0.5804,
+ "step": 89
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9442782817643425e-05,
+ "loss": 0.125,
+ "step": 90
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.9424235920596866e-05,
+ "loss": 0.338,
+ "step": 91
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9405394532478422e-05,
+ "loss": 0.3918,
+ "step": 92
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9386259242048883e-05,
+ "loss": 0.302,
+ "step": 93
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9366830647252974e-05,
+ "loss": 0.1556,
+ "step": 94
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9347109355200672e-05,
+ "loss": 0.2169,
+ "step": 95
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9327095982148258e-05,
+ "loss": 0.0737,
+ "step": 96
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9306791153479007e-05,
+ "loss": 0.2776,
+ "step": 97
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.928619550368371e-05,
+ "loss": 0.2158,
+ "step": 98
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9265309676340787e-05,
+ "loss": 0.1743,
+ "step": 99
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9244134324096223e-05,
+ "loss": 0.4233,
+ "step": 100
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9222670108643152e-05,
+ "loss": 0.1737,
+ "step": 101
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9200917700701176e-05,
+ "loss": 0.3509,
+ "step": 102
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9178877779995423e-05,
+ "loss": 0.1843,
+ "step": 103
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.915655103523529e-05,
+ "loss": 0.3164,
+ "step": 104
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9133938164092942e-05,
+ "loss": 0.3705,
+ "step": 105
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9111039873181478e-05,
+ "loss": 0.1795,
+ "step": 106
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.908785687803289e-05,
+ "loss": 0.2387,
+ "step": 107
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.9064389903075676e-05,
+ "loss": 0.2459,
+ "step": 108
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.904063968161222e-05,
+ "loss": 0.3093,
+ "step": 109
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.901660695579585e-05,
+ "loss": 0.282,
+ "step": 110
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.899229247660769e-05,
+ "loss": 0.3662,
+ "step": 111
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8967697003833156e-05,
+ "loss": 0.2212,
+ "step": 112
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.894282130603823e-05,
+ "loss": 0.1693,
+ "step": 113
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8917666160545446e-05,
+ "loss": 0.2523,
+ "step": 114
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8892232353409582e-05,
+ "loss": 0.2582,
+ "step": 115
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8892232353409582e-05,
+ "loss": 0.227,
+ "step": 116
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8866520679393127e-05,
+ "loss": 0.1532,
+ "step": 117
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.884053194194142e-05,
+ "loss": 0.2189,
+ "step": 118
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8814266953157557e-05,
+ "loss": 0.147,
+ "step": 119
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8787726533777003e-05,
+ "loss": 0.2196,
+ "step": 120
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.876091151314196e-05,
+ "loss": 0.2105,
+ "step": 121
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8733822729175452e-05,
+ "loss": 0.118,
+ "step": 122
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8706461028355107e-05,
+ "loss": 0.2145,
+ "step": 123
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.867882726568676e-05,
+ "loss": 0.2689,
+ "step": 124
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.865092230467769e-05,
+ "loss": 0.1862,
+ "step": 125
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8622747017309676e-05,
+ "loss": 0.2517,
+ "step": 126
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.8594302284011704e-05,
+ "loss": 0.2234,
+ "step": 127
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8565588993632488e-05,
+ "loss": 0.416,
+ "step": 128
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.85366080434127e-05,
+ "loss": 0.2848,
+ "step": 129
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8507360338956896e-05,
+ "loss": 0.2564,
+ "step": 130
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8477846794205258e-05,
+ "loss": 0.1887,
+ "step": 131
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.844806833140501e-05,
+ "loss": 0.2172,
+ "step": 132
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.8418025881081612e-05,
+ "loss": 0.2342,
+ "step": 133
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8387720382009665e-05,
+ "loss": 0.3647,
+ "step": 134
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8357152781183606e-05,
+ "loss": 0.4555,
+ "step": 135
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.832632403378808e-05,
+ "loss": 0.7154,
+ "step": 136
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.829523510316813e-05,
+ "loss": 0.2239,
+ "step": 137
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8263886960799062e-05,
+ "loss": 0.2482,
+ "step": 138
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.82322805862561e-05,
+ "loss": 0.314,
+ "step": 139
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8200416967183785e-05,
+ "loss": 0.2708,
+ "step": 140
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8168297099265094e-05,
+ "loss": 0.2582,
+ "step": 141
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.813592198619035e-05,
+ "loss": 0.2136,
+ "step": 142
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.810329263962584e-05,
+ "loss": 0.2046,
+ "step": 143
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8070410079182198e-05,
+ "loss": 0.1413,
+ "step": 144
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.803727533238257e-05,
+ "loss": 0.254,
+ "step": 145
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8003889434630473e-05,
+ "loss": 0.3183,
+ "step": 146
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7970253429177477e-05,
+ "loss": 0.1788,
+ "step": 147
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.793636836709057e-05,
+ "loss": 0.1193,
+ "step": 148
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7902235307219333e-05,
+ "loss": 0.1632,
+ "step": 149
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.7867855316162846e-05,
+ "loss": 0.2055,
+ "step": 150
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7833229468236367e-05,
+ "loss": 0.2053,
+ "step": 151
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7798358845437754e-05,
+ "loss": 0.1196,
+ "step": 152
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.776324453741365e-05,
+ "loss": 0.1903,
+ "step": 153
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.772788764142545e-05,
+ "loss": 0.35,
+ "step": 154
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.7692289262315e-05,
+ "loss": 0.2117,
+ "step": 155
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.765645051247007e-05,
+ "loss": 0.2519,
+ "step": 156
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7620372511789607e-05,
+ "loss": 0.2019,
+ "step": 157
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7584056387648727e-05,
+ "loss": 0.16,
+ "step": 158
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.75475032748635e-05,
+ "loss": 0.1916,
+ "step": 159
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.751071431565547e-05,
+ "loss": 0.3202,
+ "step": 160
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.747369065961599e-05,
+ "loss": 0.3153,
+ "step": 161
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.7436433463670262e-05,
+ "loss": 0.2454,
+ "step": 162
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7398943892041223e-05,
+ "loss": 0.1146,
+ "step": 163
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7361223116213143e-05,
+ "loss": 0.2135,
+ "step": 164
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7323272314895022e-05,
+ "loss": 0.2555,
+ "step": 165
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.728509267398376e-05,
+ "loss": 0.1648,
+ "step": 166
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7246685386527098e-05,
+ "loss": 0.2556,
+ "step": 167
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7208051652686335e-05,
+ "loss": 0.0675,
+ "step": 168
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7169192679698837e-05,
+ "loss": 0.1801,
+ "step": 169
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.713010968184029e-05,
+ "loss": 0.1639,
+ "step": 170
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7090803880386784e-05,
+ "loss": 0.1983,
+ "step": 171
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7051276503576623e-05,
+ "loss": 0.2065,
+ "step": 172
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.701152878657197e-05,
+ "loss": 0.386,
+ "step": 173
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.6971561971420225e-05,
+ "loss": 0.1026,
+ "step": 174
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.693137730701524e-05,
+ "loss": 0.141,
+ "step": 175
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6890976049058267e-05,
+ "loss": 0.3519,
+ "step": 176
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6850359460018737e-05,
+ "loss": 0.1873,
+ "step": 177
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6809528809094808e-05,
+ "loss": 0.2236,
+ "step": 178
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6768485372173696e-05,
+ "loss": 0.1955,
+ "step": 179
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6727230431791816e-05,
+ "loss": 0.2819,
+ "step": 180
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6685765277094702e-05,
+ "loss": 0.1513,
+ "step": 181
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6644091203796707e-05,
+ "loss": 0.1258,
+ "step": 182
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6602209514140552e-05,
+ "loss": 0.3084,
+ "step": 183
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.656012151685659e-05,
+ "loss": 0.1943,
+ "step": 184
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6517828527121942e-05,
+ "loss": 0.1087,
+ "step": 185
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6475331866519387e-05,
+ "loss": 0.3218,
+ "step": 186
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6432632862996056e-05,
+ "loss": 0.2016,
+ "step": 187
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6389732850821967e-05,
+ "loss": 0.2355,
+ "step": 188
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.634663317054829e-05,
+ "loss": 0.3003,
+ "step": 189
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6303335168965484e-05,
+ "loss": 0.2318,
+ "step": 190
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6259840199061215e-05,
+ "loss": 0.1513,
+ "step": 191
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6216149619978064e-05,
+ "loss": 0.1612,
+ "step": 192
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.617226479697105e-05,
+ "loss": 0.1565,
+ "step": 193
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.612818710136499e-05,
+ "loss": 0.3011,
+ "step": 194
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.608391791051163e-05,
+ "loss": 0.2011,
+ "step": 195
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6039458607746614e-05,
+ "loss": 0.1629,
+ "step": 196
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.599481058234626e-05,
+ "loss": 0.1287,
+ "step": 197
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.5949975229484132e-05,
+ "loss": 0.2725,
+ "step": 198
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5904953950187458e-05,
+ "loss": 0.4143,
+ "step": 199
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5859748151293347e-05,
+ "loss": 0.1918,
+ "step": 200
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5814359245404818e-05,
+ "loss": 0.2486,
+ "step": 201
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.576878865084668e-05,
+ "loss": 0.4387,
+ "step": 202
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5723037791621193e-05,
+ "loss": 0.2768,
+ "step": 203
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.5677108097363565e-05,
+ "loss": 0.1003,
+ "step": 204
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.563100100329731e-05,
+ "loss": 0.4326,
+ "step": 205
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.558471795018936e-05,
+ "loss": 0.0864,
+ "step": 206
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5538260384305076e-05,
+ "loss": 0.3033,
+ "step": 207
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5491629757363033e-05,
+ "loss": 0.1332,
+ "step": 208
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5444827526489675e-05,
+ "loss": 0.193,
+ "step": 209
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.539785515417376e-05,
+ "loss": 0.2706,
+ "step": 210
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5350714108220673e-05,
+ "loss": 0.1861,
+ "step": 211
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5303405861706574e-05,
+ "loss": 0.3058,
+ "step": 212
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5255931892932333e-05,
+ "loss": 0.1898,
+ "step": 213
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5208293685377357e-05,
+ "loss": 0.2667,
+ "step": 214
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5160492727653241e-05,
+ "loss": 0.1723,
+ "step": 215
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5112530513457236e-05,
+ "loss": 0.2885,
+ "step": 216
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5064408541525573e-05,
+ "loss": 0.214,
+ "step": 217
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.501612831558664e-05,
+ "loss": 0.2457,
+ "step": 218
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.4967691344313995e-05,
+ "loss": 0.1488,
+ "step": 219
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.4919099141279203e-05,
+ "loss": 0.1468,
+ "step": 220
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.4870353224904572e-05,
+ "loss": 0.1331,
+ "step": 221
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4821455118415669e-05,
+ "loss": 0.3833,
+ "step": 222
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4772406349793744e-05,
+ "loss": 0.3083,
+ "step": 223
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4723208451727983e-05,
+ "loss": 0.5103,
+ "step": 224
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4673862961567602e-05,
+ "loss": 0.2315,
+ "step": 225
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4624371421273823e-05,
+ "loss": 0.5056,
+ "step": 226
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.457473537737167e-05,
+ "loss": 0.345,
+ "step": 227
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4524956380901669e-05,
+ "loss": 0.2486,
+ "step": 228
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4475035987371355e-05,
+ "loss": 0.1334,
+ "step": 229
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.442497575670668e-05,
+ "loss": 0.2026,
+ "step": 230
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4374777253203273e-05,
+ "loss": 0.2028,
+ "step": 231
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4324442045477536e-05,
+ "loss": 0.1024,
+ "step": 232
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4273971706417648e-05,
+ "loss": 0.1336,
+ "step": 233
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4223367813134412e-05,
+ "loss": 0.2927,
+ "step": 234
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4172631946911964e-05,
+ "loss": 0.3204,
+ "step": 235
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4121765693158364e-05,
+ "loss": 0.1417,
+ "step": 236
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.407077064135607e-05,
+ "loss": 0.2767,
+ "step": 237
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.4019648385012243e-05,
+ "loss": 0.1877,
+ "step": 238
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.3968400521608969e-05,
+ "loss": 0.1902,
+ "step": 239
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3917028652553338e-05,
+ "loss": 0.1976,
+ "step": 240
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3865534383127406e-05,
+ "loss": 0.1658,
+ "step": 241
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3813919322438021e-05,
+ "loss": 0.4002,
+ "step": 242
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3762185083366557e-05,
+ "loss": 0.2596,
+ "step": 243
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3710333282518504e-05,
+ "loss": 0.0957,
+ "step": 244
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3658365540172948e-05,
+ "loss": 0.242,
+ "step": 245
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3606283480231957e-05,
+ "loss": 0.4266,
+ "step": 246
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3554088730169814e-05,
+ "loss": 0.3886,
+ "step": 247
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3501782920982185e-05,
+ "loss": 0.1089,
+ "step": 248
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3449367687135134e-05,
+ "loss": 0.3312,
+ "step": 249
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.339684466651406e-05,
+ "loss": 0.2238,
+ "step": 250
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.334421550037251e-05,
+ "loss": 0.1046,
+ "step": 251
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.334421550037251e-05,
+ "loss": 0.1617,
+ "step": 252
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3291481833280897e-05,
+ "loss": 0.0849,
+ "step": 253
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3238645313075104e-05,
+ "loss": 0.2548,
+ "step": 254
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3185707590805004e-05,
+ "loss": 0.1738,
+ "step": 255
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.313267032068285e-05,
+ "loss": 0.1744,
+ "step": 256
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3079535160031598e-05,
+ "loss": 0.3275,
+ "step": 257
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.3026303769233112e-05,
+ "loss": 0.2187,
+ "step": 258
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2972977811676286e-05,
+ "loss": 0.13,
+ "step": 259
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2919558953705055e-05,
+ "loss": 0.1644,
+ "step": 260
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2866048864566338e-05,
+ "loss": 0.3441,
+ "step": 261
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.2812449216357863e-05,
+ "loss": 0.1805,
+ "step": 262
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.275876168397593e-05,
+ "loss": 0.1578,
+ "step": 263
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.270498794506307e-05,
+ "loss": 0.4781,
+ "step": 264
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2651129679955604e-05,
+ "loss": 0.1001,
+ "step": 265
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.259718857163117e-05,
+ "loss": 0.225,
+ "step": 266
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2543166305656099e-05,
+ "loss": 0.216,
+ "step": 267
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2489064570132764e-05,
+ "loss": 0.3636,
+ "step": 268
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.2434885055646823e-05,
+ "loss": 0.3873,
+ "step": 269
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2380629455214392e-05,
+ "loss": 0.1739,
+ "step": 270
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2326299464229143e-05,
+ "loss": 0.2209,
+ "step": 271
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2271896780409321e-05,
+ "loss": 0.1852,
+ "step": 272
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2217423103744694e-05,
+ "loss": 0.206,
+ "step": 273
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2162880136443447e-05,
+ "loss": 0.073,
+ "step": 274
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2108269582878957e-05,
+ "loss": 0.3641,
+ "step": 275
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2053593149536576e-05,
+ "loss": 0.1036,
+ "step": 276
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1998852544960266e-05,
+ "loss": 0.1654,
+ "step": 277
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1944049479699244e-05,
+ "loss": 0.4466,
+ "step": 278
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1889185666254505e-05,
+ "loss": 0.1723,
+ "step": 279
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1834262819025326e-05,
+ "loss": 0.11,
+ "step": 280
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.1779282654255685e-05,
+ "loss": 0.1551,
+ "step": 281
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1724246889980638e-05,
+ "loss": 0.3681,
+ "step": 282
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.166915724597262e-05,
+ "loss": 0.1615,
+ "step": 283
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1614015443687723e-05,
+ "loss": 0.1501,
+ "step": 284
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1558823206211894e-05,
+ "loss": 0.1206,
+ "step": 285
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.150358225820709e-05,
+ "loss": 0.195,
+ "step": 286
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.1448294325857387e-05,
+ "loss": 0.0672,
+ "step": 287
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1392961136815046e-05,
+ "loss": 0.1577,
+ "step": 288
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.133758442014651e-05,
+ "loss": 0.4435,
+ "step": 289
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1282165906278402e-05,
+ "loss": 0.249,
+ "step": 290
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.122670732694342e-05,
+ "loss": 0.2221,
+ "step": 291
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1171210415126248e-05,
+ "loss": 0.1312,
+ "step": 292
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1115676905009385e-05,
+ "loss": 0.15,
+ "step": 293
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1060108531918972e-05,
+ "loss": 0.1346,
+ "step": 294
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1004507032270553e-05,
+ "loss": 0.2224,
+ "step": 295
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.094887414351482e-05,
+ "loss": 0.2183,
+ "step": 296
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.0893211604083325e-05,
+ "loss": 0.154,
+ "step": 297
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.0837521153334143e-05,
+ "loss": 0.1895,
+ "step": 298
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.078180453149754e-05,
+ "loss": 0.3471,
+ "step": 299
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0726063479621574e-05,
+ "loss": 0.3142,
+ "step": 300
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.067029973951771e-05,
+ "loss": 0.1031,
+ "step": 301
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0614515053706367e-05,
+ "loss": 0.1477,
+ "step": 302
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0558711165362491e-05,
+ "loss": 0.1898,
+ "step": 303
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.0502889818261075e-05,
+ "loss": 0.3198,
+ "step": 304
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.044705275672266e-05,
+ "loss": 0.3473,
+ "step": 305
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0391201725558842e-05,
+ "loss": 0.164,
+ "step": 306
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0335338470017742e-05,
+ "loss": 0.2804,
+ "step": 307
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0279464735729472e-05,
+ "loss": 0.1787,
+ "step": 308
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0223582268651585e-05,
+ "loss": 0.174,
+ "step": 309
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0167692815014527e-05,
+ "loss": 0.204,
+ "step": 310
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0111798121267047e-05,
+ "loss": 0.1659,
+ "step": 311
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0055899934021649e-05,
+ "loss": 0.0851,
+ "step": 312
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1e-05,
+ "loss": 0.125,
+ "step": 313
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.944100065978351e-06,
+ "loss": 0.1399,
+ "step": 314
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.888201878732956e-06,
+ "loss": 0.1191,
+ "step": 315
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 9.832307184985475e-06,
+ "loss": 0.2573,
+ "step": 316
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.776417731348416e-06,
+ "loss": 0.1156,
+ "step": 317
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.720535264270529e-06,
+ "loss": 0.2918,
+ "step": 318
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.664661529982261e-06,
+ "loss": 0.5064,
+ "step": 319
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.60879827444116e-06,
+ "loss": 0.1789,
+ "step": 320
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.552947243277346e-06,
+ "loss": 0.2524,
+ "step": 321
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 9.497110181738928e-06,
+ "loss": 0.1238,
+ "step": 322
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.44128883463751e-06,
+ "loss": 0.3283,
+ "step": 323
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.385484946293636e-06,
+ "loss": 0.2177,
+ "step": 324
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.329700260482292e-06,
+ "loss": 0.2896,
+ "step": 325
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.273936520378428e-06,
+ "loss": 0.4432,
+ "step": 326
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.218195468502462e-06,
+ "loss": 0.1969,
+ "step": 327
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.16247884666586e-06,
+ "loss": 0.1486,
+ "step": 328
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.106788395916679e-06,
+ "loss": 0.3046,
+ "step": 329
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.051125856485183e-06,
+ "loss": 0.1931,
+ "step": 330
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.99549296772945e-06,
+ "loss": 0.1927,
+ "step": 331
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.939891468081033e-06,
+ "loss": 0.2417,
+ "step": 332
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.884323094990619e-06,
+ "loss": 0.2002,
+ "step": 333
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 8.828789584873754e-06,
+ "loss": 0.1437,
+ "step": 334
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.773292673056582e-06,
+ "loss": 0.2163,
+ "step": 335
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.717834093721598e-06,
+ "loss": 0.0948,
+ "step": 336
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.662415579853492e-06,
+ "loss": 0.1959,
+ "step": 337
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.607038863184957e-06,
+ "loss": 0.1448,
+ "step": 338
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.551705674142618e-06,
+ "loss": 0.1835,
+ "step": 339
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 8.496417741792912e-06,
+ "loss": 0.1655,
+ "step": 340
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.44117679378811e-06,
+ "loss": 0.1666,
+ "step": 341
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.385984556312282e-06,
+ "loss": 0.1771,
+ "step": 342
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.330842754027383e-06,
+ "loss": 0.1141,
+ "step": 343
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.275753110019367e-06,
+ "loss": 0.3248,
+ "step": 344
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.220717345744316e-06,
+ "loss": 0.2598,
+ "step": 345
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.165737180974678e-06,
+ "loss": 0.1587,
+ "step": 346
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.110814333745496e-06,
+ "loss": 0.1711,
+ "step": 347
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.05595052030076e-06,
+ "loss": 0.1615,
+ "step": 348
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.001147455039735e-06,
+ "loss": 0.1857,
+ "step": 349
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.94640685046343e-06,
+ "loss": 0.1025,
+ "step": 350
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.891730417121044e-06,
+ "loss": 0.1696,
+ "step": 351
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 7.837119863556554e-06,
+ "loss": 0.1765,
+ "step": 352
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.782576896255307e-06,
+ "loss": 0.3,
+ "step": 353
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.72810321959068e-06,
+ "loss": 0.1344,
+ "step": 354
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.673700535770859e-06,
+ "loss": 0.3158,
+ "step": 355
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.619370544785608e-06,
+ "loss": 0.1455,
+ "step": 356
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.56511494435318e-06,
+ "loss": 0.2933,
+ "step": 357
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 7.510935429867237e-06,
+ "loss": 0.0901,
+ "step": 358
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.4568336943439055e-06,
+ "loss": 0.2275,
+ "step": 359
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.402811428368832e-06,
+ "loss": 0.1897,
+ "step": 360
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.348870320044399e-06,
+ "loss": 0.2391,
+ "step": 361
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.295012054936934e-06,
+ "loss": 0.1899,
+ "step": 362
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.241238316024069e-06,
+ "loss": 0.1385,
+ "step": 363
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.187550783642141e-06,
+ "loss": 0.0762,
+ "step": 364
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.133951135433666e-06,
+ "loss": 0.2305,
+ "step": 365
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.080441046294948e-06,
+ "loss": 0.1229,
+ "step": 366
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.027022188323716e-06,
+ "loss": 0.1246,
+ "step": 367
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 6.973696230766891e-06,
+ "loss": 0.2491,
+ "step": 368
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 6.920464839968405e-06,
+ "loss": 0.1749,
+ "step": 369
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 6.8673296793171555e-06,
+ "loss": 0.1952,
+ "step": 370
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.814292409194998e-06,
+ "loss": 0.115,
+ "step": 371
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.761354686924895e-06,
+ "loss": 0.1391,
+ "step": 372
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.708518166719107e-06,
+ "loss": 0.209,
+ "step": 373
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.655784499627491e-06,
+ "loss": 0.4016,
+ "step": 374
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 6.603155333485945e-06,
+ "loss": 0.1919,
+ "step": 375
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.550632312864869e-06,
+ "loss": 0.2142,
+ "step": 376
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.498217079017818e-06,
+ "loss": 0.1622,
+ "step": 377
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.445911269830189e-06,
+ "loss": 0.1669,
+ "step": 378
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.393716519768047e-06,
+ "loss": 0.3306,
+ "step": 379
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.341634459827053e-06,
+ "loss": 0.0774,
+ "step": 380
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.289666717481497e-06,
+ "loss": 0.2361,
+ "step": 381
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.237814916633444e-06,
+ "loss": 0.2916,
+ "step": 382
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.1860806775619785e-06,
+ "loss": 0.2664,
+ "step": 383
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.134465616872598e-06,
+ "loss": 0.4905,
+ "step": 384
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.082971347446662e-06,
+ "loss": 0.1202,
+ "step": 385
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.0315994783910345e-06,
+ "loss": 0.2228,
+ "step": 386
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 5.980351614987759e-06,
+ "loss": 0.1403,
+ "step": 387
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.929229358643932e-06,
+ "loss": 0.3658,
+ "step": 388
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.878234306841637e-06,
+ "loss": 0.121,
+ "step": 389
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.827368053088043e-06,
+ "loss": 0.2419,
+ "step": 390
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.7766321868655935e-06,
+ "loss": 0.1211,
+ "step": 391
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.726028293582355e-06,
+ "loss": 0.419,
+ "step": 392
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 5.67555795452247e-06,
+ "loss": 0.2181,
+ "step": 393
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.62522274679673e-06,
+ "loss": 0.1513,
+ "step": 394
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.575024243293319e-06,
+ "loss": 0.1522,
+ "step": 395
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.524964012628648e-06,
+ "loss": 0.1583,
+ "step": 396
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.475043619098334e-06,
+ "loss": 0.1475,
+ "step": 397
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.42526462262833e-06,
+ "loss": 0.1265,
+ "step": 398
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.375628578726181e-06,
+ "loss": 0.0715,
+ "step": 399
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.326137038432399e-06,
+ "loss": 0.1164,
+ "step": 400
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.276791548272018e-06,
+ "loss": 0.3881,
+ "step": 401
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.227593650206258e-06,
+ "loss": 0.1464,
+ "step": 402
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.1785448815843334e-06,
+ "loss": 0.2286,
+ "step": 403
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.129646775095432e-06,
+ "loss": 0.1454,
+ "step": 404
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.0809008587207965e-06,
+ "loss": 0.1155,
+ "step": 405
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.032308655686011e-06,
+ "loss": 0.1199,
+ "step": 406
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.983871684413363e-06,
+ "loss": 0.3385,
+ "step": 407
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.935591458474433e-06,
+ "loss": 0.2083,
+ "step": 408
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.8874694865427676e-06,
+ "loss": 0.1057,
+ "step": 409
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.8395072723467585e-06,
+ "loss": 0.1584,
+ "step": 410
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 4.791706314622645e-06,
+ "loss": 0.2643,
+ "step": 411
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.74406810706767e-06,
+ "loss": 0.1302,
+ "step": 412
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.69659413829343e-06,
+ "loss": 0.2024,
+ "step": 413
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.649285891779327e-06,
+ "loss": 0.1177,
+ "step": 414
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.602144845826246e-06,
+ "loss": 0.1947,
+ "step": 415
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.5551724735103285e-06,
+ "loss": 0.1209,
+ "step": 416
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 4.508370242636968e-06,
+ "loss": 0.2273,
+ "step": 417
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.461739615694929e-06,
+ "loss": 0.3131,
+ "step": 418
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.415282049810644e-06,
+ "loss": 0.41,
+ "step": 419
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.368998996702694e-06,
+ "loss": 0.218,
+ "step": 420
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.3228919026364345e-06,
+ "loss": 0.1857,
+ "step": 421
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.276962208378811e-06,
+ "loss": 0.1159,
+ "step": 422
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.231211349153319e-06,
+ "loss": 0.1981,
+ "step": 423
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.185640754595183e-06,
+ "loss": 0.2907,
+ "step": 424
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.140251848706656e-06,
+ "loss": 0.1582,
+ "step": 425
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.095046049812545e-06,
+ "loss": 0.1264,
+ "step": 426
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.050024770515869e-06,
+ "loss": 0.1817,
+ "step": 427
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.005189417653743e-06,
+ "loss": 0.1073,
+ "step": 428
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 3.960541392253387e-06,
+ "loss": 0.3221,
+ "step": 429
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.916082089488372e-06,
+ "loss": 0.2237,
+ "step": 430
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.8718128986350154e-06,
+ "loss": 0.3927,
+ "step": 431
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.827735203028953e-06,
+ "loss": 0.1443,
+ "step": 432
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.7838503800219393e-06,
+ "loss": 0.1289,
+ "step": 433
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.740159800938784e-06,
+ "loss": 0.1407,
+ "step": 434
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 3.696664831034519e-06,
+ "loss": 0.4103,
+ "step": 435
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6533668294517154e-06,
+ "loss": 0.1538,
+ "step": 436
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.6102671491780393e-06,
+ "loss": 0.4277,
+ "step": 437
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.5673671370039464e-06,
+ "loss": 0.1458,
+ "step": 438
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.5246681334806177e-06,
+ "loss": 0.1699,
+ "step": 439
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.482171472878062e-06,
+ "loss": 0.2724,
+ "step": 440
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.4398784831434127e-06,
+ "loss": 0.2037,
+ "step": 441
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.39779048585945e-06,
+ "loss": 0.3123,
+ "step": 442
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.3559087962032956e-06,
+ "loss": 0.2008,
+ "step": 443
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.314234722905302e-06,
+ "loss": 0.253,
+ "step": 444
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.272769568208183e-06,
+ "loss": 0.1709,
+ "step": 445
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.2315146278263053e-06,
+ "loss": 0.2399,
+ "step": 446
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.1904711909051933e-06,
+ "loss": 0.1039,
+ "step": 447
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.149640539981267e-06,
+ "loss": 0.2212,
+ "step": 448
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.1090239509417364e-06,
+ "loss": 0.2057,
+ "step": 449
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0686226929847617e-06,
+ "loss": 0.1692,
+ "step": 450
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.0284380285797767e-06,
+ "loss": 0.1658,
+ "step": 451
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 2.9884712134280324e-06,
+ "loss": 0.1372,
+ "step": 452
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.948723496423379e-06,
+ "loss": 0.1991,
+ "step": 453
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.909196119613218e-06,
+ "loss": 0.1983,
+ "step": 454
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.869890318159713e-06,
+ "loss": 0.4125,
+ "step": 455
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.8308073203011667e-06,
+ "loss": 0.4987,
+ "step": 456
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.7919483473136678e-06,
+ "loss": 0.1212,
+ "step": 457
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 2.753314613472906e-06,
+ "loss": 0.2399,
+ "step": 458
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.7149073260162416e-06,
+ "loss": 0.1534,
+ "step": 459
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.6767276851049818e-06,
+ "loss": 0.2789,
+ "step": 460
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.63877688378686e-06,
+ "loss": 0.1404,
+ "step": 461
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.6010561079587817e-06,
+ "loss": 0.1757,
+ "step": 462
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.5635665363297424e-06,
+ "loss": 0.1194,
+ "step": 463
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.5263093403840145e-06,
+ "loss": 0.1268,
+ "step": 464
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.489285684344532e-06,
+ "loss": 0.1493,
+ "step": 465
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.452496725136503e-06,
+ "loss": 0.2272,
+ "step": 466
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.4159436123512737e-06,
+ "loss": 0.1863,
+ "step": 467
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.3796274882103964e-06,
+ "loss": 0.1783,
+ "step": 468
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.3435494875299315e-06,
+ "loss": 0.1379,
+ "step": 469
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.3077107376850005e-06,
+ "loss": 0.4607,
+ "step": 470
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.272112358574551e-06,
+ "loss": 0.078,
+ "step": 471
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.2367554625863496e-06,
+ "loss": 0.2215,
+ "step": 472
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.2016411545622497e-06,
+ "loss": 0.1333,
+ "step": 473
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.1667705317636333e-06,
+ "loss": 0.1098,
+ "step": 474
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.132144683837155e-06,
+ "loss": 0.1947,
+ "step": 475
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.0977646927806682e-06,
+ "loss": 0.1148,
+ "step": 476
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.0636316329094317e-06,
+ "loss": 0.2097,
+ "step": 477
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.029746570822524e-06,
+ "loss": 0.127,
+ "step": 478
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.996110565369527e-06,
+ "loss": 0.2963,
+ "step": 479
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.9627246676174363e-06,
+ "loss": 0.172,
+ "step": 480
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.929589920817806e-06,
+ "loss": 0.2083,
+ "step": 481
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 1.896707360374167e-06,
+ "loss": 0.1132,
+ "step": 482
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8640780138096515e-06,
+ "loss": 0.2204,
+ "step": 483
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.8317029007349086e-06,
+ "loss": 0.2599,
+ "step": 484
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.799583032816219e-06,
+ "loss": 0.2668,
+ "step": 485
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7677194137439036e-06,
+ "loss": 0.1575,
+ "step": 486
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7361130392009407e-06,
+ "loss": 0.2251,
+ "step": 487
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 1.7047648968318697e-06,
+ "loss": 0.2407,
+ "step": 488
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6736759662119183e-06,
+ "loss": 0.1042,
+ "step": 489
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.642847218816398e-06,
+ "loss": 0.2138,
+ "step": 490
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.6122796179903355e-06,
+ "loss": 0.185,
+ "step": 491
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.5819741189183902e-06,
+ "loss": 0.1251,
+ "step": 492
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.5519316685949903e-06,
+ "loss": 0.1887,
+ "step": 493
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.522153205794742e-06,
+ "loss": 0.1527,
+ "step": 494
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.492639661043106e-06,
+ "loss": 0.3192,
+ "step": 495
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4633919565873033e-06,
+ "loss": 0.2173,
+ "step": 496
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4344110063675143e-06,
+ "loss": 0.1351,
+ "step": 497
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.4056977159883011e-06,
+ "loss": 0.1096,
+ "step": 498
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.377252982690327e-06,
+ "loss": 0.2903,
+ "step": 499
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.3490776953223107e-06,
+ "loss": 0.1286,
+ "step": 500
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.3211727343132441e-06,
+ "loss": 0.2258,
+ "step": 501
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.2935389716448976e-06,
+ "loss": 0.1646,
+ "step": 502
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.2661772708245535e-06,
+ "loss": 0.2163,
+ "step": 503
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.23908848685804e-06,
+ "loss": 0.2114,
+ "step": 504
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.2122734662229985e-06,
+ "loss": 0.0903,
+ "step": 505
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.1857330468424466e-06,
+ "loss": 0.0616,
+ "step": 506
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1594680580585815e-06,
+ "loss": 0.1434,
+ "step": 507
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1334793206068739e-06,
+ "loss": 0.1386,
+ "step": 508
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.1077676465904209e-06,
+ "loss": 0.1691,
+ "step": 509
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.082333839454559e-06,
+ "loss": 0.1445,
+ "step": 510
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.057178693961771e-06,
+ "loss": 0.2405,
+ "step": 511
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.0323029961668463e-06,
+ "loss": 0.5961,
+ "step": 512
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.0077075233923118e-06,
+ "loss": 0.1181,
+ "step": 513
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.833930442041506e-07,
+ "loss": 0.2651,
+ "step": 514
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.593603183877843e-07,
+ "loss": 0.1735,
+ "step": 515
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.356100969243231e-07,
+ "loss": 0.1939,
+ "step": 516
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 9.121431219671096e-07,
+ "loss": 0.2992,
+ "step": 517
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.889601268185233e-07,
+ "loss": 0.1875,
+ "step": 518
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.660618359070605e-07,
+ "loss": 0.119,
+ "step": 519
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.434489647647093e-07,
+ "loss": 0.1525,
+ "step": 520
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 8.211222200045787e-07,
+ "loss": 0.228,
+ "step": 521
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 7.990822992988267e-07,
+ "loss": 0.2455,
+ "step": 522
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 7.773298913568506e-07,
+ "loss": 0.1871,
+ "step": 523
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.558656759037796e-07,
+ "loss": 0.3385,
+ "step": 524
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.346903236592162e-07,
+ "loss": 0.1667,
+ "step": 525
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 7.13804496316296e-07,
+ "loss": 0.1244,
+ "step": 526
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 6.932088465209941e-07,
+ "loss": 0.2674,
+ "step": 527
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 6.729040178517454e-07,
+ "loss": 0.2326,
+ "step": 528
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 6.528906447993289e-07,
+ "loss": 0.2116,
+ "step": 529
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.331693527470306e-07,
+ "loss": 0.3147,
+ "step": 530
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 6.137407579511212e-07,
+ "loss": 0.2206,
+ "step": 531
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 5.946054675215785e-07,
+ "loss": 0.113,
+ "step": 532
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 5.757640794031361e-07,
+ "loss": 0.2499,
+ "step": 533
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 5.572171823565797e-07,
+ "loss": 0.392,
+ "step": 534
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 5.389653559403629e-07,
+ "loss": 0.1796,
+ "step": 535
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.210091704924947e-07,
+ "loss": 0.1849,
+ "step": 536
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 5.033491871127105e-07,
+ "loss": 0.1075,
+ "step": 537
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 4.859859576449444e-07,
+ "loss": 0.0891,
+ "step": 538
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 4.6892002466008666e-07,
+ "loss": 0.3812,
+ "step": 539
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 4.5215192143902577e-07,
+ "loss": 0.2431,
+ "step": 540
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 4.3568217195598117e-07,
+ "loss": 0.1961,
+ "step": 541
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.1951129086214015e-07,
+ "loss": 0.2062,
+ "step": 542
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 4.03639783469566e-07,
+ "loss": 0.0653,
+ "step": 543
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.8806814573541185e-07,
+ "loss": 0.221,
+ "step": 544
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.7279686424642413e-07,
+ "loss": 0.1787,
+ "step": 545
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.578264162037348e-07,
+ "loss": 0.1328,
+ "step": 546
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 3.4315726940795436e-07,
+ "loss": 0.0782,
+ "step": 547
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.2878988224454346e-07,
+ "loss": 0.22,
+ "step": 548
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.147247036695034e-07,
+ "loss": 0.2186,
+ "step": 549
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 3.0096217319533386e-07,
+ "loss": 0.2259,
+ "step": 550
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 2.875027208773118e-07,
+ "loss": 0.291,
+ "step": 551
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 2.7434676730003886e-07,
+ "loss": 0.1645,
+ "step": 552
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 2.614947235643106e-07,
+ "loss": 0.2621,
+ "step": 553
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.489469912742637e-07,
+ "loss": 0.3108,
+ "step": 554
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.3670396252483018e-07,
+ "loss": 0.3828,
+ "step": 555
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.2476601988947965e-07,
+ "loss": 0.1785,
+ "step": 556
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.1313353640827207e-07,
+ "loss": 0.1605,
+ "step": 557
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.0180687557619816e-07,
+ "loss": 0.2059,
+ "step": 558
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 1.9078639133181532e-07,
+ "loss": 0.1236,
+ "step": 559
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.800724280461963e-07,
+ "loss": 0.4072,
+ "step": 560
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.6966532051216122e-07,
+ "loss": 0.0857,
+ "step": 561
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.5956539393382043e-07,
+ "loss": 0.0805,
+ "step": 562
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.4977296391641026e-07,
+ "loss": 0.1157,
+ "step": 563
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.4028833645643113e-07,
+ "loss": 0.3035,
+ "step": 564
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 1.31111807932085e-07,
+ "loss": 0.0902,
+ "step": 565
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.2224366509401732e-07,
+ "loss": 0.097,
+ "step": 566
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.1368418505635303e-07,
+ "loss": 0.2516,
+ "step": 567
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.0543363528803696e-07,
+ "loss": 0.13,
+ "step": 568
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 9.749227360448143e-08,
+ "loss": 0.1561,
+ "step": 569
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 8.986034815950173e-08,
+ "loss": 0.2111,
+ "step": 570
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 8.253809743756669e-08,
+ "loss": 0.1692,
+ "step": 571
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 7.55257502463469e-08,
+ "loss": 0.1259,
+ "step": 572
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 6.882352570956485e-08,
+ "loss": 0.3749,
+ "step": 573
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 6.243163326014268e-08,
+ "loss": 0.1758,
+ "step": 574
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 5.6350272633664e-08,
+ "loss": 0.2633,
+ "step": 575
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 5.057963386213116e-08,
+ "loss": 0.2307,
+ "step": 576
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 4.5119897268023347e-08,
+ "loss": 0.2843,
+ "step": 577
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 3.9971233458665495e-08,
+ "loss": 0.1782,
+ "step": 578
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 3.5133803320897e-08,
+ "loss": 0.2064,
+ "step": 579
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 3.0607758016043546e-08,
+ "loss": 0.0771,
+ "step": 580
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 2.639323897518975e-08,
+ "loss": 0.4028,
+ "step": 581
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 2.2490377894768266e-08,
+ "loss": 0.2803,
+ "step": 582
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 1.889929673243529e-08,
+ "loss": 0.3806,
+ "step": 583
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.562010770326916e-08,
+ "loss": 0.1015,
+ "step": 584
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.2652913276250956e-08,
+ "loss": 0.1817,
+ "step": 585
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 1.2652913276250956e-08,
+ "loss": 0.1556,
+ "step": 586
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 9.99780617107815e-09,
+ "loss": 0.1861,
+ "step": 587
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 7.654869355252503e-09,
+ "loss": 0.1602,
+ "step": 588
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 5.6241760414987856e-09,
+ "loss": 0.0909,
+ "step": 589
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 3.905789685471062e-09,
+ "loss": 0.3967,
+ "step": 590
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 2.4997639837687217e-09,
+ "loss": 0.1172,
+ "step": 591
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 1.406142872263372e-09,
+ "loss": 0.0301,
+ "step": 592
+ },
+ {
+ "epoch": 1.0,
+ "step": 592,
+ "total_flos": 1850959245312.0,
+ "train_loss": 0.23161595452551706,
+ "train_runtime": 4403.6215,
+ "train_samples_per_second": 1.342,
+ "train_steps_per_second": 0.134
+ }
+ ],
+ "logging_steps": 1.0,
+ "max_steps": 592,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 1,
+ "save_steps": 500,
+ "total_flos": 1850959245312.0,
+ "train_batch_size": 10,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/CheckGuard Models/wholeimage/payee/finetune_lora_llava_mistral.sh b/CheckGuard Models/wholeimage/payee/finetune_lora_llava_mistral.sh
new file mode 100644
index 0000000000000000000000000000000000000000..821a1002cc463e565cdbd5fe12f2156d6e2b5004
--- /dev/null
+++ b/CheckGuard Models/wholeimage/payee/finetune_lora_llava_mistral.sh
@@ -0,0 +1,39 @@
+#!/bin/bash
+
+deepspeed llava/train/train_mem.py \
+ --lora_enable True --lora_r 128 --lora_alpha 256 --mm_projector_lr 2e-5 \
+ --deepspeed ./scripts/zero3.json \
+ --model_name_or_path liuhaotian/llava-v1.6-mistral-7b \
+ --version mistral_instruct \
+ --data_path /home/larry5/project/LLaVA-1.6-ft/data/peft/payee/csv_gt/payee_modified_path_to_train_val_human-gpt-whole-check.json \
+ --image_folder /home/larry5/project/LLaVA-1.6-ft/data/data/ \
+ --vision_tower openai/clip-vit-large-patch14-336 \
+ --mm_projector_type mlp2x_gelu \
+ --mm_vision_select_layer -2 \
+ --mm_use_im_start_end False \
+ --mm_use_im_patch_token False \
+ --mm_patch_merge_type spatial_unpad \
+ --image_aspect_ratio anyres \
+ --group_by_modality_length False \
+ --bf16 False \
+ --fp16 True \
+ --output_dir /home/larry5/project/LLaVA-1.6-ft/scripts_peft/mistral/lora/llava-lora-mistral-r128a256/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model \
+ --num_train_epochs 1 \
+ --per_device_train_batch_size 10 \
+ --per_device_eval_batch_size 1 \
+ --gradient_accumulation_steps 1 \
+ --evaluation_strategy "no" \
+ --save_strategy "steps" \
+ --save_steps 500 \
+ --save_total_limit 5 \
+ --learning_rate 2e-5 \
+ --weight_decay 0. \
+ --warmup_ratio 0.05 \
+ --lr_scheduler_type "cosine" \
+ --logging_steps 1 \
+ --tf32 True \
+ --model_max_length 4096 \
+ --gradient_checkpointing True \
+ --dataloader_num_workers 4 \
+ --lazy_preprocess True \
+ --report_to wandb \
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/README.md b/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..bdb138eee6972419f6d60676388b52fd99ec478e
--- /dev/null
+++ b/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/README.md
@@ -0,0 +1,202 @@
+---
+library_name: peft
+base_model: liuhaotian/llava-v1.6-mistral-7b
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.10.0
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/adapter_config.json b/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..2546d7f5c4c0e793d197d4a8310446465b4d2ede
--- /dev/null
+++ b/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/adapter_config.json
@@ -0,0 +1,34 @@
+{
+ "alpha_pattern": {},
+ "auto_mapping": null,
+ "base_model_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "bias": "none",
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 256,
+ "lora_dropout": 0.05,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "r": 128,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "down_proj",
+ "k_proj",
+ "gate_proj",
+ "q_proj",
+ "o_proj",
+ "up_proj",
+ "v_proj"
+ ],
+ "task_type": "CAUSAL_LM",
+ "use_dora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors b/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..c281d5fb191e9098c81b6c3cc868cff0e9bef2fb
--- /dev/null
+++ b/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:096bc4369107bb2bdfe3944bf8bb696866e306494b750e810cc8938769bb6d30
+size 708924928
diff --git a/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/config.json b/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..93e133af45036a778791b5679a8953a4f6a35a33
--- /dev/null
+++ b/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/config.json
@@ -0,0 +1,70 @@
+{
+ "_name_or_path": "liuhaotian/llava-v1.6-mistral-7b",
+ "architectures": [
+ "LlavaMistralForCausalLM"
+ ],
+ "attention_dropout": 0.0,
+ "bos_token_id": 1,
+ "eos_token_id": 2,
+ "freeze_mm_mlp_adapter": false,
+ "freeze_mm_vision_resampler": false,
+ "hidden_act": "silu",
+ "hidden_size": 4096,
+ "image_aspect_ratio": "anyres",
+ "image_crop_resolution": 224,
+ "image_grid_pinpoints": [
+ [
+ 336,
+ 672
+ ],
+ [
+ 672,
+ 336
+ ],
+ [
+ 672,
+ 672
+ ],
+ [
+ 1008,
+ 336
+ ],
+ [
+ 336,
+ 1008
+ ]
+ ],
+ "image_split_resolution": 224,
+ "initializer_range": 0.02,
+ "intermediate_size": 14336,
+ "max_position_embeddings": 32768,
+ "mm_hidden_size": 1024,
+ "mm_patch_merge_type": "spatial_unpad",
+ "mm_projector_lr": 2e-05,
+ "mm_projector_type": "mlp2x_gelu",
+ "mm_resampler_type": null,
+ "mm_use_im_patch_token": false,
+ "mm_use_im_start_end": false,
+ "mm_vision_select_feature": "patch",
+ "mm_vision_select_layer": -2,
+ "mm_vision_tower": "openai/clip-vit-large-patch14-336",
+ "mm_vision_tower_lr": 2e-06,
+ "model_type": "llava_mistral",
+ "num_attention_heads": 32,
+ "num_hidden_layers": 32,
+ "num_key_value_heads": 8,
+ "rms_norm_eps": 1e-05,
+ "rope_theta": 1000000.0,
+ "sliding_window": null,
+ "tie_word_embeddings": false,
+ "tokenizer_model_max_length": 4096,
+ "tokenizer_padding_side": "right",
+ "torch_dtype": "bfloat16",
+ "transformers_version": "4.37.2",
+ "tune_mm_mlp_adapter": false,
+ "tune_mm_vision_resampler": false,
+ "unfreeze_mm_vision_tower": true,
+ "use_cache": true,
+ "use_mm_proj": true,
+ "vocab_size": 32000
+}
diff --git a/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin b/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin
new file mode 100644
index 0000000000000000000000000000000000000000..c01cafb9f483f66c0580ed9d7da872b5a496d30f
--- /dev/null
+++ b/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/non_lora_trainables.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d9b47307b2d9b9d9badc01bd4896752d30223bbf3544b4b45ee1e1cc0af81440
+size 41961648
diff --git a/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/trainer_state.json b/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..740af236ce258382066f2145ae6f37bf04139722
--- /dev/null
+++ b/CheckGuard Models/wholeimage/payee/llava-lora-mistral-r128a256-10BS-model/trainer_state.json
@@ -0,0 +1,684 @@
+{
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.0,
+ "eval_steps": 500,
+ "global_step": 109,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "epoch": 0.01,
+ "learning_rate": 3.3333333333333333e-06,
+ "loss": 0.7809,
+ "step": 1
+ },
+ {
+ "epoch": 0.02,
+ "learning_rate": 6.666666666666667e-06,
+ "loss": 0.7122,
+ "step": 2
+ },
+ {
+ "epoch": 0.03,
+ "learning_rate": 1e-05,
+ "loss": 0.65,
+ "step": 3
+ },
+ {
+ "epoch": 0.04,
+ "learning_rate": 1.3333333333333333e-05,
+ "loss": 0.6036,
+ "step": 4
+ },
+ {
+ "epoch": 0.05,
+ "learning_rate": 1.6666666666666667e-05,
+ "loss": 0.5032,
+ "step": 5
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 2e-05,
+ "loss": 0.5093,
+ "step": 6
+ },
+ {
+ "epoch": 0.06,
+ "learning_rate": 1.9995348836233517e-05,
+ "loss": 0.3085,
+ "step": 7
+ },
+ {
+ "epoch": 0.07,
+ "learning_rate": 1.998139967159894e-05,
+ "loss": 0.4182,
+ "step": 8
+ },
+ {
+ "epoch": 0.08,
+ "learning_rate": 1.9958165482066094e-05,
+ "loss": 0.2635,
+ "step": 9
+ },
+ {
+ "epoch": 0.09,
+ "learning_rate": 1.992566788083908e-05,
+ "loss": 0.2694,
+ "step": 10
+ },
+ {
+ "epoch": 0.1,
+ "learning_rate": 1.9883937098250962e-05,
+ "loss": 0.4074,
+ "step": 11
+ },
+ {
+ "epoch": 0.11,
+ "learning_rate": 1.9833011953642525e-05,
+ "loss": 0.5041,
+ "step": 12
+ },
+ {
+ "epoch": 0.12,
+ "learning_rate": 1.9772939819251247e-05,
+ "loss": 0.3813,
+ "step": 13
+ },
+ {
+ "epoch": 0.13,
+ "learning_rate": 1.9703776576144106e-05,
+ "loss": 0.3196,
+ "step": 14
+ },
+ {
+ "epoch": 0.14,
+ "learning_rate": 1.962558656223516e-05,
+ "loss": 0.3915,
+ "step": 15
+ },
+ {
+ "epoch": 0.15,
+ "learning_rate": 1.953844251243633e-05,
+ "loss": 0.4946,
+ "step": 16
+ },
+ {
+ "epoch": 0.16,
+ "learning_rate": 1.9442425490996987e-05,
+ "loss": 0.2764,
+ "step": 17
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.933762481609536e-05,
+ "loss": 0.197,
+ "step": 18
+ },
+ {
+ "epoch": 0.17,
+ "learning_rate": 1.9224137976751797e-05,
+ "loss": 0.7322,
+ "step": 19
+ },
+ {
+ "epoch": 0.18,
+ "learning_rate": 1.910207054214133e-05,
+ "loss": 0.3586,
+ "step": 20
+ },
+ {
+ "epoch": 0.19,
+ "learning_rate": 1.8971536063389745e-05,
+ "loss": 0.5291,
+ "step": 21
+ },
+ {
+ "epoch": 0.2,
+ "learning_rate": 1.8832655967944607e-05,
+ "loss": 0.3152,
+ "step": 22
+ },
+ {
+ "epoch": 0.21,
+ "learning_rate": 1.868555944661949e-05,
+ "loss": 0.3683,
+ "step": 23
+ },
+ {
+ "epoch": 0.22,
+ "learning_rate": 1.853038333341642e-05,
+ "loss": 0.5445,
+ "step": 24
+ },
+ {
+ "epoch": 0.23,
+ "learning_rate": 1.8367271978238422e-05,
+ "loss": 0.5318,
+ "step": 25
+ },
+ {
+ "epoch": 0.24,
+ "learning_rate": 1.8196377112610524e-05,
+ "loss": 0.4471,
+ "step": 26
+ },
+ {
+ "epoch": 0.25,
+ "learning_rate": 1.8017857708534107e-05,
+ "loss": 0.4027,
+ "step": 27
+ },
+ {
+ "epoch": 0.26,
+ "learning_rate": 1.783187983060594e-05,
+ "loss": 0.3362,
+ "step": 28
+ },
+ {
+ "epoch": 0.27,
+ "learning_rate": 1.763861648153945e-05,
+ "loss": 0.6705,
+ "step": 29
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.743824744123196e-05,
+ "loss": 0.3705,
+ "step": 30
+ },
+ {
+ "epoch": 0.28,
+ "learning_rate": 1.7230959099527512e-05,
+ "loss": 0.3289,
+ "step": 31
+ },
+ {
+ "epoch": 0.29,
+ "learning_rate": 1.7016944282830935e-05,
+ "loss": 0.4026,
+ "step": 32
+ },
+ {
+ "epoch": 0.3,
+ "learning_rate": 1.6796402074734404e-05,
+ "loss": 0.3437,
+ "step": 33
+ },
+ {
+ "epoch": 0.31,
+ "learning_rate": 1.6569537630823385e-05,
+ "loss": 0.367,
+ "step": 34
+ },
+ {
+ "epoch": 0.32,
+ "learning_rate": 1.6336561987834155e-05,
+ "loss": 0.2763,
+ "step": 35
+ },
+ {
+ "epoch": 0.33,
+ "learning_rate": 1.6097691867340547e-05,
+ "loss": 0.309,
+ "step": 36
+ },
+ {
+ "epoch": 0.34,
+ "learning_rate": 1.585314947415242e-05,
+ "loss": 0.3304,
+ "step": 37
+ },
+ {
+ "epoch": 0.35,
+ "learning_rate": 1.5603162289613503e-05,
+ "loss": 0.3243,
+ "step": 38
+ },
+ {
+ "epoch": 0.36,
+ "learning_rate": 1.5347962859990744e-05,
+ "loss": 0.3459,
+ "step": 39
+ },
+ {
+ "epoch": 0.37,
+ "learning_rate": 1.5087788580152207e-05,
+ "loss": 0.3876,
+ "step": 40
+ },
+ {
+ "epoch": 0.38,
+ "learning_rate": 1.4822881472734563e-05,
+ "loss": 0.2243,
+ "step": 41
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.4553487963005712e-05,
+ "loss": 0.4515,
+ "step": 42
+ },
+ {
+ "epoch": 0.39,
+ "learning_rate": 1.427985864963193e-05,
+ "loss": 0.3135,
+ "step": 43
+ },
+ {
+ "epoch": 0.4,
+ "learning_rate": 1.400224807156278e-05,
+ "loss": 0.3343,
+ "step": 44
+ },
+ {
+ "epoch": 0.41,
+ "learning_rate": 1.3720914471250644e-05,
+ "loss": 0.4998,
+ "step": 45
+ },
+ {
+ "epoch": 0.42,
+ "learning_rate": 1.3720914471250644e-05,
+ "loss": 0.3234,
+ "step": 46
+ },
+ {
+ "epoch": 0.43,
+ "learning_rate": 1.3436119554425133e-05,
+ "loss": 0.2812,
+ "step": 47
+ },
+ {
+ "epoch": 0.44,
+ "learning_rate": 1.314812824664585e-05,
+ "loss": 0.1462,
+ "step": 48
+ },
+ {
+ "epoch": 0.45,
+ "learning_rate": 1.285720844685996e-05,
+ "loss": 0.3018,
+ "step": 49
+ },
+ {
+ "epoch": 0.46,
+ "learning_rate": 1.2563630778193805e-05,
+ "loss": 0.7224,
+ "step": 50
+ },
+ {
+ "epoch": 0.47,
+ "learning_rate": 1.2563630778193805e-05,
+ "loss": 0.5369,
+ "step": 51
+ },
+ {
+ "epoch": 0.48,
+ "learning_rate": 1.2267668336210411e-05,
+ "loss": 0.5186,
+ "step": 52
+ },
+ {
+ "epoch": 0.49,
+ "learning_rate": 1.1969596434867063e-05,
+ "loss": 0.6036,
+ "step": 53
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1669692350409223e-05,
+ "loss": 0.3408,
+ "step": 54
+ },
+ {
+ "epoch": 0.5,
+ "learning_rate": 1.1368235063439103e-05,
+ "loss": 0.1877,
+ "step": 55
+ },
+ {
+ "epoch": 0.51,
+ "learning_rate": 1.1065504999398762e-05,
+ "loss": 0.3792,
+ "step": 56
+ },
+ {
+ "epoch": 0.52,
+ "learning_rate": 1.0761783767709182e-05,
+ "loss": 0.3529,
+ "step": 57
+ },
+ {
+ "epoch": 0.53,
+ "learning_rate": 1.0457353899807947e-05,
+ "loss": 0.1392,
+ "step": 58
+ },
+ {
+ "epoch": 0.54,
+ "learning_rate": 1.015249858632926e-05,
+ "loss": 0.4194,
+ "step": 59
+ },
+ {
+ "epoch": 0.55,
+ "learning_rate": 9.847501413670742e-06,
+ "loss": 0.4476,
+ "step": 60
+ },
+ {
+ "epoch": 0.56,
+ "learning_rate": 9.542646100192056e-06,
+ "loss": 0.501,
+ "step": 61
+ },
+ {
+ "epoch": 0.57,
+ "learning_rate": 9.238216232290821e-06,
+ "loss": 0.5975,
+ "step": 62
+ },
+ {
+ "epoch": 0.58,
+ "learning_rate": 8.934495000601241e-06,
+ "loss": 0.3509,
+ "step": 63
+ },
+ {
+ "epoch": 0.59,
+ "learning_rate": 8.6317649365609e-06,
+ "loss": 0.1972,
+ "step": 64
+ },
+ {
+ "epoch": 0.6,
+ "learning_rate": 8.330307649590782e-06,
+ "loss": 0.3311,
+ "step": 65
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 8.030403565132942e-06,
+ "loss": 0.4667,
+ "step": 66
+ },
+ {
+ "epoch": 0.61,
+ "learning_rate": 7.732331663789592e-06,
+ "loss": 0.2561,
+ "step": 67
+ },
+ {
+ "epoch": 0.62,
+ "learning_rate": 7.436369221806201e-06,
+ "loss": 0.2634,
+ "step": 68
+ },
+ {
+ "epoch": 0.63,
+ "learning_rate": 7.142791553140045e-06,
+ "loss": 0.2677,
+ "step": 69
+ },
+ {
+ "epoch": 0.64,
+ "learning_rate": 6.851871753354154e-06,
+ "loss": 0.3111,
+ "step": 70
+ },
+ {
+ "epoch": 0.65,
+ "learning_rate": 6.563880445574873e-06,
+ "loss": 0.6517,
+ "step": 71
+ },
+ {
+ "epoch": 0.66,
+ "learning_rate": 6.2790855287493605e-06,
+ "loss": 0.2516,
+ "step": 72
+ },
+ {
+ "epoch": 0.67,
+ "learning_rate": 5.99775192843722e-06,
+ "loss": 0.3525,
+ "step": 73
+ },
+ {
+ "epoch": 0.68,
+ "learning_rate": 5.720141350368072e-06,
+ "loss": 0.5402,
+ "step": 74
+ },
+ {
+ "epoch": 0.69,
+ "learning_rate": 5.446512036994287e-06,
+ "loss": 0.3389,
+ "step": 75
+ },
+ {
+ "epoch": 0.7,
+ "learning_rate": 5.177118527265438e-06,
+ "loss": 0.2144,
+ "step": 76
+ },
+ {
+ "epoch": 0.71,
+ "learning_rate": 4.912211419847795e-06,
+ "loss": 0.6116,
+ "step": 77
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.652037140009259e-06,
+ "loss": 0.3376,
+ "step": 78
+ },
+ {
+ "epoch": 0.72,
+ "learning_rate": 4.396837710386503e-06,
+ "loss": 0.36,
+ "step": 79
+ },
+ {
+ "epoch": 0.73,
+ "learning_rate": 4.1468505258475785e-06,
+ "loss": 0.4222,
+ "step": 80
+ },
+ {
+ "epoch": 0.74,
+ "learning_rate": 3.902308132659457e-06,
+ "loss": 0.2264,
+ "step": 81
+ },
+ {
+ "epoch": 0.75,
+ "learning_rate": 3.6634380121658484e-06,
+ "loss": 0.2858,
+ "step": 82
+ },
+ {
+ "epoch": 0.76,
+ "learning_rate": 3.4304623691766193e-06,
+ "loss": 0.3124,
+ "step": 83
+ },
+ {
+ "epoch": 0.77,
+ "learning_rate": 3.203597925265598e-06,
+ "loss": 0.124,
+ "step": 84
+ },
+ {
+ "epoch": 0.78,
+ "learning_rate": 2.98305571716907e-06,
+ "loss": 0.4321,
+ "step": 85
+ },
+ {
+ "epoch": 0.79,
+ "learning_rate": 2.7690409004724883e-06,
+ "loss": 0.4045,
+ "step": 86
+ },
+ {
+ "epoch": 0.8,
+ "learning_rate": 2.56175255876804e-06,
+ "loss": 0.4559,
+ "step": 87
+ },
+ {
+ "epoch": 0.81,
+ "learning_rate": 2.3613835184605527e-06,
+ "loss": 0.1838,
+ "step": 88
+ },
+ {
+ "epoch": 0.82,
+ "learning_rate": 2.1681201693940667e-06,
+ "loss": 0.4033,
+ "step": 89
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.982142291465896e-06,
+ "loss": 0.3051,
+ "step": 90
+ },
+ {
+ "epoch": 0.83,
+ "learning_rate": 1.8036228873894745e-06,
+ "loss": 0.2386,
+ "step": 91
+ },
+ {
+ "epoch": 0.84,
+ "learning_rate": 1.6327280217615793e-06,
+ "loss": 0.3897,
+ "step": 92
+ },
+ {
+ "epoch": 0.85,
+ "learning_rate": 1.4696166665835853e-06,
+ "loss": 0.3724,
+ "step": 93
+ },
+ {
+ "epoch": 0.86,
+ "learning_rate": 1.3144405533805138e-06,
+ "loss": 0.4102,
+ "step": 94
+ },
+ {
+ "epoch": 0.87,
+ "learning_rate": 1.1673440320553941e-06,
+ "loss": 0.2726,
+ "step": 95
+ },
+ {
+ "epoch": 0.88,
+ "learning_rate": 1.02846393661026e-06,
+ "loss": 0.4034,
+ "step": 96
+ },
+ {
+ "epoch": 0.89,
+ "learning_rate": 8.979294578586739e-07,
+ "loss": 0.6888,
+ "step": 97
+ },
+ {
+ "epoch": 0.9,
+ "learning_rate": 7.758620232482083e-07,
+ "loss": 0.3449,
+ "step": 98
+ },
+ {
+ "epoch": 0.91,
+ "learning_rate": 6.623751839046455e-07,
+ "loss": 0.5181,
+ "step": 99
+ },
+ {
+ "epoch": 0.92,
+ "learning_rate": 5.575745090030138e-07,
+ "loss": 0.7589,
+ "step": 100
+ },
+ {
+ "epoch": 0.93,
+ "learning_rate": 4.61557487563673e-07,
+ "loss": 0.3506,
+ "step": 101
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 3.7441343776484116e-07,
+ "loss": 0.5096,
+ "step": 102
+ },
+ {
+ "epoch": 0.94,
+ "learning_rate": 2.9622342385589256e-07,
+ "loss": 0.2922,
+ "step": 103
+ },
+ {
+ "epoch": 0.95,
+ "learning_rate": 2.2706018074875046e-07,
+ "loss": 0.223,
+ "step": 104
+ },
+ {
+ "epoch": 0.96,
+ "learning_rate": 1.669880463574758e-07,
+ "loss": 0.1446,
+ "step": 105
+ },
+ {
+ "epoch": 0.97,
+ "learning_rate": 1.160629017490389e-07,
+ "loss": 0.3283,
+ "step": 106
+ },
+ {
+ "epoch": 0.98,
+ "learning_rate": 7.433211916092143e-08,
+ "loss": 0.2504,
+ "step": 107
+ },
+ {
+ "epoch": 0.99,
+ "learning_rate": 4.183451793390747e-08,
+ "loss": 0.4626,
+ "step": 108
+ },
+ {
+ "epoch": 1.0,
+ "learning_rate": 1.860032840106163e-08,
+ "loss": 0.161,
+ "step": 109
+ },
+ {
+ "epoch": 1.0,
+ "step": 109,
+ "total_flos": 394052874240.0,
+ "train_loss": 0.3891204473895764,
+ "train_runtime": 831.1939,
+ "train_samples_per_second": 2.62,
+ "train_steps_per_second": 0.131
+ }
+ ],
+ "logging_steps": 1.0,
+ "max_steps": 109,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 1,
+ "save_steps": 500,
+ "total_flos": 394052874240.0,
+ "train_batch_size": 10,
+ "trial_name": null,
+ "trial_params": null
+}