Commit
•
3bdd32d
1
Parent(s):
ce7e7b5
Model save
Browse files- README.md +69 -0
- all_results.json +9 -0
- train_results.json +9 -0
- trainer_state.json +3258 -0
README.md
ADDED
|
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
|
| 3 |
+
library_name: peft
|
| 4 |
+
license: llama3.1
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- sft
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: barc-llama3.1-8b-instruct-lora64-induction-gpt4mini20k_lr2e-4_epoch3
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# barc-llama3.1-8b-instruct-lora64-induction-gpt4mini20k_lr2e-4_epoch3
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct) on an unknown dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.3327
|
| 22 |
+
|
| 23 |
+
## Model description
|
| 24 |
+
|
| 25 |
+
More information needed
|
| 26 |
+
|
| 27 |
+
## Intended uses & limitations
|
| 28 |
+
|
| 29 |
+
More information needed
|
| 30 |
+
|
| 31 |
+
## Training and evaluation data
|
| 32 |
+
|
| 33 |
+
More information needed
|
| 34 |
+
|
| 35 |
+
## Training procedure
|
| 36 |
+
|
| 37 |
+
### Training hyperparameters
|
| 38 |
+
|
| 39 |
+
The following hyperparameters were used during training:
|
| 40 |
+
- learning_rate: 0.0002
|
| 41 |
+
- train_batch_size: 8
|
| 42 |
+
- eval_batch_size: 4
|
| 43 |
+
- seed: 42
|
| 44 |
+
- distributed_type: multi-GPU
|
| 45 |
+
- num_devices: 8
|
| 46 |
+
- gradient_accumulation_steps: 2
|
| 47 |
+
- total_train_batch_size: 128
|
| 48 |
+
- total_eval_batch_size: 32
|
| 49 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 50 |
+
- lr_scheduler_type: cosine
|
| 51 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 52 |
+
- num_epochs: 3
|
| 53 |
+
|
| 54 |
+
### Training results
|
| 55 |
+
|
| 56 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 57 |
+
|:-------------:|:-----:|:----:|:---------------:|
|
| 58 |
+
| 0.3519 | 1.0 | 152 | 0.3529 |
|
| 59 |
+
| 0.3233 | 2.0 | 304 | 0.3357 |
|
| 60 |
+
| 0.2989 | 3.0 | 456 | 0.3327 |
|
| 61 |
+
|
| 62 |
+
|
| 63 |
+
### Framework versions
|
| 64 |
+
|
| 65 |
+
- PEFT 0.13.0
|
| 66 |
+
- Transformers 4.45.0.dev0
|
| 67 |
+
- Pytorch 2.4.1+cu121
|
| 68 |
+
- Datasets 3.0.0
|
| 69 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 1.6172409563591475e+17,
|
| 4 |
+
"train_loss": 0.366262159849468,
|
| 5 |
+
"train_runtime": 5417.5508,
|
| 6 |
+
"train_samples": 19396,
|
| 7 |
+
"train_samples_per_second": 10.741,
|
| 8 |
+
"train_steps_per_second": 0.084
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 1.6172409563591475e+17,
|
| 4 |
+
"train_loss": 0.366262159849468,
|
| 5 |
+
"train_runtime": 5417.5508,
|
| 6 |
+
"train_samples": 19396,
|
| 7 |
+
"train_samples_per_second": 10.741,
|
| 8 |
+
"train_steps_per_second": 0.084
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3258 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 456,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.006578947368421052,
|
| 13 |
+
"grad_norm": 0.9301734884942886,
|
| 14 |
+
"learning_rate": 4.347826086956522e-06,
|
| 15 |
+
"loss": 1.3345,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.013157894736842105,
|
| 20 |
+
"grad_norm": 0.8660026064620485,
|
| 21 |
+
"learning_rate": 8.695652173913044e-06,
|
| 22 |
+
"loss": 1.2699,
|
| 23 |
+
"step": 2
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.019736842105263157,
|
| 27 |
+
"grad_norm": 0.9006761056531956,
|
| 28 |
+
"learning_rate": 1.3043478260869566e-05,
|
| 29 |
+
"loss": 1.3134,
|
| 30 |
+
"step": 3
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.02631578947368421,
|
| 34 |
+
"grad_norm": 0.8334670285502003,
|
| 35 |
+
"learning_rate": 1.739130434782609e-05,
|
| 36 |
+
"loss": 1.2548,
|
| 37 |
+
"step": 4
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.03289473684210526,
|
| 41 |
+
"grad_norm": 0.8574803030028485,
|
| 42 |
+
"learning_rate": 2.173913043478261e-05,
|
| 43 |
+
"loss": 1.2843,
|
| 44 |
+
"step": 5
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.039473684210526314,
|
| 48 |
+
"grad_norm": 0.8390817055257793,
|
| 49 |
+
"learning_rate": 2.608695652173913e-05,
|
| 50 |
+
"loss": 1.2982,
|
| 51 |
+
"step": 6
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.046052631578947366,
|
| 55 |
+
"grad_norm": 0.7141521506417161,
|
| 56 |
+
"learning_rate": 3.0434782608695656e-05,
|
| 57 |
+
"loss": 1.1906,
|
| 58 |
+
"step": 7
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.05263157894736842,
|
| 62 |
+
"grad_norm": 0.6047790773556203,
|
| 63 |
+
"learning_rate": 3.478260869565218e-05,
|
| 64 |
+
"loss": 1.1499,
|
| 65 |
+
"step": 8
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.05921052631578947,
|
| 69 |
+
"grad_norm": 0.5265109630433017,
|
| 70 |
+
"learning_rate": 3.91304347826087e-05,
|
| 71 |
+
"loss": 1.0976,
|
| 72 |
+
"step": 9
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.06578947368421052,
|
| 76 |
+
"grad_norm": 0.5245802695709278,
|
| 77 |
+
"learning_rate": 4.347826086956522e-05,
|
| 78 |
+
"loss": 1.0108,
|
| 79 |
+
"step": 10
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.07236842105263158,
|
| 83 |
+
"grad_norm": 0.5599855466933923,
|
| 84 |
+
"learning_rate": 4.782608695652174e-05,
|
| 85 |
+
"loss": 0.9675,
|
| 86 |
+
"step": 11
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.07894736842105263,
|
| 90 |
+
"grad_norm": 0.5570743219368649,
|
| 91 |
+
"learning_rate": 5.217391304347826e-05,
|
| 92 |
+
"loss": 0.8945,
|
| 93 |
+
"step": 12
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.08552631578947369,
|
| 97 |
+
"grad_norm": 0.5874369878213753,
|
| 98 |
+
"learning_rate": 5.652173913043478e-05,
|
| 99 |
+
"loss": 0.8503,
|
| 100 |
+
"step": 13
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.09210526315789473,
|
| 104 |
+
"grad_norm": 0.522753964863263,
|
| 105 |
+
"learning_rate": 6.086956521739131e-05,
|
| 106 |
+
"loss": 0.7773,
|
| 107 |
+
"step": 14
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.09868421052631579,
|
| 111 |
+
"grad_norm": 0.5330305040727032,
|
| 112 |
+
"learning_rate": 6.521739130434783e-05,
|
| 113 |
+
"loss": 0.7115,
|
| 114 |
+
"step": 15
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.10526315789473684,
|
| 118 |
+
"grad_norm": 0.46085140769365757,
|
| 119 |
+
"learning_rate": 6.956521739130436e-05,
|
| 120 |
+
"loss": 0.6463,
|
| 121 |
+
"step": 16
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.1118421052631579,
|
| 125 |
+
"grad_norm": 0.40848688620706747,
|
| 126 |
+
"learning_rate": 7.391304347826086e-05,
|
| 127 |
+
"loss": 0.6151,
|
| 128 |
+
"step": 17
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.11842105263157894,
|
| 132 |
+
"grad_norm": 0.2704450670383811,
|
| 133 |
+
"learning_rate": 7.82608695652174e-05,
|
| 134 |
+
"loss": 0.5724,
|
| 135 |
+
"step": 18
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.125,
|
| 139 |
+
"grad_norm": 0.26085072447544416,
|
| 140 |
+
"learning_rate": 8.260869565217392e-05,
|
| 141 |
+
"loss": 0.5528,
|
| 142 |
+
"step": 19
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.13157894736842105,
|
| 146 |
+
"grad_norm": 0.2339583161738833,
|
| 147 |
+
"learning_rate": 8.695652173913044e-05,
|
| 148 |
+
"loss": 0.5369,
|
| 149 |
+
"step": 20
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.13815789473684212,
|
| 153 |
+
"grad_norm": 0.1949857904626772,
|
| 154 |
+
"learning_rate": 9.130434782608696e-05,
|
| 155 |
+
"loss": 0.512,
|
| 156 |
+
"step": 21
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.14473684210526316,
|
| 160 |
+
"grad_norm": 0.2435718076959428,
|
| 161 |
+
"learning_rate": 9.565217391304348e-05,
|
| 162 |
+
"loss": 0.5202,
|
| 163 |
+
"step": 22
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.1513157894736842,
|
| 167 |
+
"grad_norm": 0.22205719081769748,
|
| 168 |
+
"learning_rate": 0.0001,
|
| 169 |
+
"loss": 0.5221,
|
| 170 |
+
"step": 23
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.15789473684210525,
|
| 174 |
+
"grad_norm": 0.21129557114647404,
|
| 175 |
+
"learning_rate": 0.00010434782608695653,
|
| 176 |
+
"loss": 0.5128,
|
| 177 |
+
"step": 24
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.16447368421052633,
|
| 181 |
+
"grad_norm": 0.18916988013906638,
|
| 182 |
+
"learning_rate": 0.00010869565217391305,
|
| 183 |
+
"loss": 0.4869,
|
| 184 |
+
"step": 25
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.17105263157894737,
|
| 188 |
+
"grad_norm": 0.14949200071673818,
|
| 189 |
+
"learning_rate": 0.00011304347826086956,
|
| 190 |
+
"loss": 0.4821,
|
| 191 |
+
"step": 26
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.17763157894736842,
|
| 195 |
+
"grad_norm": 0.1581846054226008,
|
| 196 |
+
"learning_rate": 0.0001173913043478261,
|
| 197 |
+
"loss": 0.4661,
|
| 198 |
+
"step": 27
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.18421052631578946,
|
| 202 |
+
"grad_norm": 0.134913498610646,
|
| 203 |
+
"learning_rate": 0.00012173913043478263,
|
| 204 |
+
"loss": 0.4758,
|
| 205 |
+
"step": 28
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.19078947368421054,
|
| 209 |
+
"grad_norm": 0.12244728979819958,
|
| 210 |
+
"learning_rate": 0.00012608695652173915,
|
| 211 |
+
"loss": 0.4504,
|
| 212 |
+
"step": 29
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.19736842105263158,
|
| 216 |
+
"grad_norm": 0.12177957867071984,
|
| 217 |
+
"learning_rate": 0.00013043478260869567,
|
| 218 |
+
"loss": 0.4788,
|
| 219 |
+
"step": 30
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.20394736842105263,
|
| 223 |
+
"grad_norm": 0.109229502136198,
|
| 224 |
+
"learning_rate": 0.0001347826086956522,
|
| 225 |
+
"loss": 0.4434,
|
| 226 |
+
"step": 31
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.21052631578947367,
|
| 230 |
+
"grad_norm": 0.11628917220798876,
|
| 231 |
+
"learning_rate": 0.0001391304347826087,
|
| 232 |
+
"loss": 0.4536,
|
| 233 |
+
"step": 32
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.21710526315789475,
|
| 237 |
+
"grad_norm": 0.11733881064063928,
|
| 238 |
+
"learning_rate": 0.0001434782608695652,
|
| 239 |
+
"loss": 0.4742,
|
| 240 |
+
"step": 33
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.2236842105263158,
|
| 244 |
+
"grad_norm": 0.11627552292165463,
|
| 245 |
+
"learning_rate": 0.00014782608695652173,
|
| 246 |
+
"loss": 0.4494,
|
| 247 |
+
"step": 34
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.23026315789473684,
|
| 251 |
+
"grad_norm": 0.11476166833257723,
|
| 252 |
+
"learning_rate": 0.00015217391304347827,
|
| 253 |
+
"loss": 0.4319,
|
| 254 |
+
"step": 35
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.23684210526315788,
|
| 258 |
+
"grad_norm": 0.10237139066203954,
|
| 259 |
+
"learning_rate": 0.0001565217391304348,
|
| 260 |
+
"loss": 0.4263,
|
| 261 |
+
"step": 36
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.24342105263157895,
|
| 265 |
+
"grad_norm": 0.10918908589611533,
|
| 266 |
+
"learning_rate": 0.00016086956521739132,
|
| 267 |
+
"loss": 0.4254,
|
| 268 |
+
"step": 37
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.25,
|
| 272 |
+
"grad_norm": 0.10679450577166048,
|
| 273 |
+
"learning_rate": 0.00016521739130434784,
|
| 274 |
+
"loss": 0.4363,
|
| 275 |
+
"step": 38
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.2565789473684211,
|
| 279 |
+
"grad_norm": 0.10517151889540283,
|
| 280 |
+
"learning_rate": 0.00016956521739130436,
|
| 281 |
+
"loss": 0.4353,
|
| 282 |
+
"step": 39
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.2631578947368421,
|
| 286 |
+
"grad_norm": 0.10453003257689569,
|
| 287 |
+
"learning_rate": 0.00017391304347826088,
|
| 288 |
+
"loss": 0.4313,
|
| 289 |
+
"step": 40
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.26973684210526316,
|
| 293 |
+
"grad_norm": 0.09705593623180399,
|
| 294 |
+
"learning_rate": 0.0001782608695652174,
|
| 295 |
+
"loss": 0.4146,
|
| 296 |
+
"step": 41
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.27631578947368424,
|
| 300 |
+
"grad_norm": 0.09840834125049662,
|
| 301 |
+
"learning_rate": 0.00018260869565217392,
|
| 302 |
+
"loss": 0.414,
|
| 303 |
+
"step": 42
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.28289473684210525,
|
| 307 |
+
"grad_norm": 0.09733214025720571,
|
| 308 |
+
"learning_rate": 0.00018695652173913045,
|
| 309 |
+
"loss": 0.4163,
|
| 310 |
+
"step": 43
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.2894736842105263,
|
| 314 |
+
"grad_norm": 0.09403117831890086,
|
| 315 |
+
"learning_rate": 0.00019130434782608697,
|
| 316 |
+
"loss": 0.4095,
|
| 317 |
+
"step": 44
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.29605263157894735,
|
| 321 |
+
"grad_norm": 0.0998536829474165,
|
| 322 |
+
"learning_rate": 0.0001956521739130435,
|
| 323 |
+
"loss": 0.4075,
|
| 324 |
+
"step": 45
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.3026315789473684,
|
| 328 |
+
"grad_norm": 0.09752527766251139,
|
| 329 |
+
"learning_rate": 0.0002,
|
| 330 |
+
"loss": 0.4194,
|
| 331 |
+
"step": 46
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.3092105263157895,
|
| 335 |
+
"grad_norm": 0.09784770986893948,
|
| 336 |
+
"learning_rate": 0.0001999970643795027,
|
| 337 |
+
"loss": 0.4109,
|
| 338 |
+
"step": 47
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.3157894736842105,
|
| 342 |
+
"grad_norm": 0.09501975538592294,
|
| 343 |
+
"learning_rate": 0.00019998825769036819,
|
| 344 |
+
"loss": 0.4116,
|
| 345 |
+
"step": 48
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.3223684210526316,
|
| 349 |
+
"grad_norm": 0.09520308666764123,
|
| 350 |
+
"learning_rate": 0.0001999735804496583,
|
| 351 |
+
"loss": 0.4016,
|
| 352 |
+
"step": 49
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.32894736842105265,
|
| 356 |
+
"grad_norm": 0.09458518327177581,
|
| 357 |
+
"learning_rate": 0.0001999530335191093,
|
| 358 |
+
"loss": 0.4038,
|
| 359 |
+
"step": 50
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.3355263157894737,
|
| 363 |
+
"grad_norm": 0.08881816935147156,
|
| 364 |
+
"learning_rate": 0.000199926618105081,
|
| 365 |
+
"loss": 0.3881,
|
| 366 |
+
"step": 51
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.34210526315789475,
|
| 370 |
+
"grad_norm": 0.09507811168988711,
|
| 371 |
+
"learning_rate": 0.000199894335758486,
|
| 372 |
+
"loss": 0.4101,
|
| 373 |
+
"step": 52
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.34868421052631576,
|
| 377 |
+
"grad_norm": 0.0892408897509568,
|
| 378 |
+
"learning_rate": 0.00019985618837469864,
|
| 379 |
+
"loss": 0.3976,
|
| 380 |
+
"step": 53
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.35526315789473684,
|
| 384 |
+
"grad_norm": 0.08932074970893863,
|
| 385 |
+
"learning_rate": 0.0001998121781934438,
|
| 386 |
+
"loss": 0.3869,
|
| 387 |
+
"step": 54
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.3618421052631579,
|
| 391 |
+
"grad_norm": 0.09109394290790748,
|
| 392 |
+
"learning_rate": 0.00019976230779866525,
|
| 393 |
+
"loss": 0.3898,
|
| 394 |
+
"step": 55
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.3684210526315789,
|
| 398 |
+
"grad_norm": 0.09116027892556224,
|
| 399 |
+
"learning_rate": 0.00019970658011837404,
|
| 400 |
+
"loss": 0.398,
|
| 401 |
+
"step": 56
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.375,
|
| 405 |
+
"grad_norm": 0.09891633296064052,
|
| 406 |
+
"learning_rate": 0.00019964499842447663,
|
| 407 |
+
"loss": 0.4046,
|
| 408 |
+
"step": 57
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.3815789473684211,
|
| 412 |
+
"grad_norm": 0.0939928712835388,
|
| 413 |
+
"learning_rate": 0.00019957756633258265,
|
| 414 |
+
"loss": 0.3864,
|
| 415 |
+
"step": 58
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.3881578947368421,
|
| 419 |
+
"grad_norm": 0.09374915175716102,
|
| 420 |
+
"learning_rate": 0.00019950428780179274,
|
| 421 |
+
"loss": 0.3951,
|
| 422 |
+
"step": 59
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.39473684210526316,
|
| 426 |
+
"grad_norm": 0.09474512693386154,
|
| 427 |
+
"learning_rate": 0.000199425167134466,
|
| 428 |
+
"loss": 0.3911,
|
| 429 |
+
"step": 60
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.40131578947368424,
|
| 433 |
+
"grad_norm": 0.09534030624628262,
|
| 434 |
+
"learning_rate": 0.0001993402089759675,
|
| 435 |
+
"loss": 0.3917,
|
| 436 |
+
"step": 61
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.40789473684210525,
|
| 440 |
+
"grad_norm": 0.09308439623268501,
|
| 441 |
+
"learning_rate": 0.0001992494183143955,
|
| 442 |
+
"loss": 0.3799,
|
| 443 |
+
"step": 62
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.4144736842105263,
|
| 447 |
+
"grad_norm": 0.0963241819886369,
|
| 448 |
+
"learning_rate": 0.00019915280048028852,
|
| 449 |
+
"loss": 0.3958,
|
| 450 |
+
"step": 63
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.42105263157894735,
|
| 454 |
+
"grad_norm": 0.09034072116748955,
|
| 455 |
+
"learning_rate": 0.00019905036114631247,
|
| 456 |
+
"loss": 0.3735,
|
| 457 |
+
"step": 64
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.4276315789473684,
|
| 461 |
+
"grad_norm": 0.09570337959260364,
|
| 462 |
+
"learning_rate": 0.00019894210632692745,
|
| 463 |
+
"loss": 0.3854,
|
| 464 |
+
"step": 65
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.4342105263157895,
|
| 468 |
+
"grad_norm": 0.091943482076957,
|
| 469 |
+
"learning_rate": 0.00019882804237803488,
|
| 470 |
+
"loss": 0.3839,
|
| 471 |
+
"step": 66
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.4407894736842105,
|
| 475 |
+
"grad_norm": 0.09392178800966167,
|
| 476 |
+
"learning_rate": 0.000198708175996604,
|
| 477 |
+
"loss": 0.3922,
|
| 478 |
+
"step": 67
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.4473684210526316,
|
| 482 |
+
"grad_norm": 0.09032901472780272,
|
| 483 |
+
"learning_rate": 0.00019858251422027903,
|
| 484 |
+
"loss": 0.3827,
|
| 485 |
+
"step": 68
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.45394736842105265,
|
| 489 |
+
"grad_norm": 0.08955279616077026,
|
| 490 |
+
"learning_rate": 0.00019845106442696562,
|
| 491 |
+
"loss": 0.3782,
|
| 492 |
+
"step": 69
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.4605263157894737,
|
| 496 |
+
"grad_norm": 0.08888768341850095,
|
| 497 |
+
"learning_rate": 0.00019831383433439797,
|
| 498 |
+
"loss": 0.3803,
|
| 499 |
+
"step": 70
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.46710526315789475,
|
| 503 |
+
"grad_norm": 0.10135869270529443,
|
| 504 |
+
"learning_rate": 0.00019817083199968552,
|
| 505 |
+
"loss": 0.3876,
|
| 506 |
+
"step": 71
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.47368421052631576,
|
| 510 |
+
"grad_norm": 0.09494931509695588,
|
| 511 |
+
"learning_rate": 0.00019802206581883994,
|
| 512 |
+
"loss": 0.3878,
|
| 513 |
+
"step": 72
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.48026315789473684,
|
| 517 |
+
"grad_norm": 0.09731569285786043,
|
| 518 |
+
"learning_rate": 0.00019786754452628225,
|
| 519 |
+
"loss": 0.3731,
|
| 520 |
+
"step": 73
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.4868421052631579,
|
| 524 |
+
"grad_norm": 0.09286168859706302,
|
| 525 |
+
"learning_rate": 0.00019770727719432994,
|
| 526 |
+
"loss": 0.3852,
|
| 527 |
+
"step": 74
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.4934210526315789,
|
| 531 |
+
"grad_norm": 0.09331628340360382,
|
| 532 |
+
"learning_rate": 0.00019754127323266428,
|
| 533 |
+
"loss": 0.3679,
|
| 534 |
+
"step": 75
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.5,
|
| 538 |
+
"grad_norm": 0.0985838743425182,
|
| 539 |
+
"learning_rate": 0.00019736954238777792,
|
| 540 |
+
"loss": 0.3751,
|
| 541 |
+
"step": 76
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.506578947368421,
|
| 545 |
+
"grad_norm": 0.08784090808893116,
|
| 546 |
+
"learning_rate": 0.00019719209474240262,
|
| 547 |
+
"loss": 0.3857,
|
| 548 |
+
"step": 77
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.5131578947368421,
|
| 552 |
+
"grad_norm": 0.09595397552494361,
|
| 553 |
+
"learning_rate": 0.00019700894071491732,
|
| 554 |
+
"loss": 0.3772,
|
| 555 |
+
"step": 78
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.5197368421052632,
|
| 559 |
+
"grad_norm": 0.09751786827076832,
|
| 560 |
+
"learning_rate": 0.00019682009105873635,
|
| 561 |
+
"loss": 0.3783,
|
| 562 |
+
"step": 79
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.5263157894736842,
|
| 566 |
+
"grad_norm": 0.08800370843142873,
|
| 567 |
+
"learning_rate": 0.00019662555686167808,
|
| 568 |
+
"loss": 0.3662,
|
| 569 |
+
"step": 80
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.5328947368421053,
|
| 573 |
+
"grad_norm": 0.0897695028264937,
|
| 574 |
+
"learning_rate": 0.0001964253495453141,
|
| 575 |
+
"loss": 0.3524,
|
| 576 |
+
"step": 81
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.5394736842105263,
|
| 580 |
+
"grad_norm": 0.09955413713238825,
|
| 581 |
+
"learning_rate": 0.00019621948086429844,
|
| 582 |
+
"loss": 0.3893,
|
| 583 |
+
"step": 82
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.5460526315789473,
|
| 587 |
+
"grad_norm": 0.09349570369317932,
|
| 588 |
+
"learning_rate": 0.00019600796290567747,
|
| 589 |
+
"loss": 0.3677,
|
| 590 |
+
"step": 83
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.5526315789473685,
|
| 594 |
+
"grad_norm": 0.08892170997103152,
|
| 595 |
+
"learning_rate": 0.00019579080808818032,
|
| 596 |
+
"loss": 0.3695,
|
| 597 |
+
"step": 84
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.5592105263157895,
|
| 601 |
+
"grad_norm": 0.09487589068069692,
|
| 602 |
+
"learning_rate": 0.00019556802916148962,
|
| 603 |
+
"loss": 0.3758,
|
| 604 |
+
"step": 85
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.5657894736842105,
|
| 608 |
+
"grad_norm": 0.08802853628490082,
|
| 609 |
+
"learning_rate": 0.00019533963920549306,
|
| 610 |
+
"loss": 0.3611,
|
| 611 |
+
"step": 86
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.5723684210526315,
|
| 615 |
+
"grad_norm": 0.0952039827447668,
|
| 616 |
+
"learning_rate": 0.00019510565162951537,
|
| 617 |
+
"loss": 0.3748,
|
| 618 |
+
"step": 87
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.5789473684210527,
|
| 622 |
+
"grad_norm": 0.08848254195231814,
|
| 623 |
+
"learning_rate": 0.000194866080171531,
|
| 624 |
+
"loss": 0.3599,
|
| 625 |
+
"step": 88
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.5855263157894737,
|
| 629 |
+
"grad_norm": 0.09613545225239202,
|
| 630 |
+
"learning_rate": 0.00019462093889735765,
|
| 631 |
+
"loss": 0.3733,
|
| 632 |
+
"step": 89
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.5921052631578947,
|
| 636 |
+
"grad_norm": 0.08924377105304611,
|
| 637 |
+
"learning_rate": 0.00019437024219983028,
|
| 638 |
+
"loss": 0.3788,
|
| 639 |
+
"step": 90
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.5986842105263158,
|
| 643 |
+
"grad_norm": 0.0970473798319214,
|
| 644 |
+
"learning_rate": 0.00019411400479795617,
|
| 645 |
+
"loss": 0.3768,
|
| 646 |
+
"step": 91
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.6052631578947368,
|
| 650 |
+
"grad_norm": 0.09487853197595246,
|
| 651 |
+
"learning_rate": 0.0001938522417360507,
|
| 652 |
+
"loss": 0.3707,
|
| 653 |
+
"step": 92
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.6118421052631579,
|
| 657 |
+
"grad_norm": 0.08878860279782896,
|
| 658 |
+
"learning_rate": 0.00019358496838285406,
|
| 659 |
+
"loss": 0.3709,
|
| 660 |
+
"step": 93
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.618421052631579,
|
| 664 |
+
"grad_norm": 0.0894046820861241,
|
| 665 |
+
"learning_rate": 0.00019331220043062894,
|
| 666 |
+
"loss": 0.3732,
|
| 667 |
+
"step": 94
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.625,
|
| 671 |
+
"grad_norm": 0.0873587537636765,
|
| 672 |
+
"learning_rate": 0.00019303395389423918,
|
| 673 |
+
"loss": 0.378,
|
| 674 |
+
"step": 95
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.631578947368421,
|
| 678 |
+
"grad_norm": 0.09689373631878609,
|
| 679 |
+
"learning_rate": 0.0001927502451102095,
|
| 680 |
+
"loss": 0.3696,
|
| 681 |
+
"step": 96
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.6381578947368421,
|
| 685 |
+
"grad_norm": 0.09705847084532389,
|
| 686 |
+
"learning_rate": 0.00019246109073576628,
|
| 687 |
+
"loss": 0.3604,
|
| 688 |
+
"step": 97
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.6447368421052632,
|
| 692 |
+
"grad_norm": 0.09181701291383483,
|
| 693 |
+
"learning_rate": 0.00019216650774785972,
|
| 694 |
+
"loss": 0.3711,
|
| 695 |
+
"step": 98
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.6513157894736842,
|
| 699 |
+
"grad_norm": 0.09684384392230812,
|
| 700 |
+
"learning_rate": 0.00019186651344216702,
|
| 701 |
+
"loss": 0.3736,
|
| 702 |
+
"step": 99
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.6578947368421053,
|
| 706 |
+
"grad_norm": 0.10298393644504046,
|
| 707 |
+
"learning_rate": 0.00019156112543207673,
|
| 708 |
+
"loss": 0.3802,
|
| 709 |
+
"step": 100
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.6644736842105263,
|
| 713 |
+
"grad_norm": 0.09420976346161887,
|
| 714 |
+
"learning_rate": 0.000191250361647655,
|
| 715 |
+
"loss": 0.3672,
|
| 716 |
+
"step": 101
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.6710526315789473,
|
| 720 |
+
"grad_norm": 0.08630124052676108,
|
| 721 |
+
"learning_rate": 0.00019093424033459248,
|
| 722 |
+
"loss": 0.3709,
|
| 723 |
+
"step": 102
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.6776315789473685,
|
| 727 |
+
"grad_norm": 0.09501717321245914,
|
| 728 |
+
"learning_rate": 0.0001906127800531333,
|
| 729 |
+
"loss": 0.366,
|
| 730 |
+
"step": 103
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.6842105263157895,
|
| 734 |
+
"grad_norm": 0.08821786734234693,
|
| 735 |
+
"learning_rate": 0.00019028599967698531,
|
| 736 |
+
"loss": 0.3689,
|
| 737 |
+
"step": 104
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.6907894736842105,
|
| 741 |
+
"grad_norm": 0.08658177205492697,
|
| 742 |
+
"learning_rate": 0.0001899539183922119,
|
| 743 |
+
"loss": 0.3541,
|
| 744 |
+
"step": 105
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.6973684210526315,
|
| 748 |
+
"grad_norm": 0.09325609783945248,
|
| 749 |
+
"learning_rate": 0.00018961655569610557,
|
| 750 |
+
"loss": 0.3721,
|
| 751 |
+
"step": 106
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.7039473684210527,
|
| 755 |
+
"grad_norm": 0.08841280838509634,
|
| 756 |
+
"learning_rate": 0.00018927393139604325,
|
| 757 |
+
"loss": 0.3519,
|
| 758 |
+
"step": 107
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.7105263157894737,
|
| 762 |
+
"grad_norm": 0.09690514694100973,
|
| 763 |
+
"learning_rate": 0.00018892606560832337,
|
| 764 |
+
"loss": 0.3774,
|
| 765 |
+
"step": 108
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.7171052631578947,
|
| 769 |
+
"grad_norm": 0.09722105218508262,
|
| 770 |
+
"learning_rate": 0.00018857297875698455,
|
| 771 |
+
"loss": 0.3558,
|
| 772 |
+
"step": 109
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.7236842105263158,
|
| 776 |
+
"grad_norm": 0.08839435869483182,
|
| 777 |
+
"learning_rate": 0.00018821469157260685,
|
| 778 |
+
"loss": 0.3565,
|
| 779 |
+
"step": 110
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.7302631578947368,
|
| 783 |
+
"grad_norm": 0.08814358581045939,
|
| 784 |
+
"learning_rate": 0.00018785122509109426,
|
| 785 |
+
"loss": 0.362,
|
| 786 |
+
"step": 111
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.7368421052631579,
|
| 790 |
+
"grad_norm": 0.08994997062082188,
|
| 791 |
+
"learning_rate": 0.00018748260065243986,
|
| 792 |
+
"loss": 0.3455,
|
| 793 |
+
"step": 112
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.743421052631579,
|
| 797 |
+
"grad_norm": 0.08543302223033665,
|
| 798 |
+
"learning_rate": 0.0001871088398994728,
|
| 799 |
+
"loss": 0.362,
|
| 800 |
+
"step": 113
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.75,
|
| 804 |
+
"grad_norm": 0.08894111865879421,
|
| 805 |
+
"learning_rate": 0.00018672996477658767,
|
| 806 |
+
"loss": 0.3643,
|
| 807 |
+
"step": 114
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.756578947368421,
|
| 811 |
+
"grad_norm": 0.08853456191177575,
|
| 812 |
+
"learning_rate": 0.00018634599752845592,
|
| 813 |
+
"loss": 0.3658,
|
| 814 |
+
"step": 115
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.7631578947368421,
|
| 818 |
+
"grad_norm": 0.09412265410116702,
|
| 819 |
+
"learning_rate": 0.00018595696069872013,
|
| 820 |
+
"loss": 0.3732,
|
| 821 |
+
"step": 116
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.7697368421052632,
|
| 825 |
+
"grad_norm": 0.08843510819092883,
|
| 826 |
+
"learning_rate": 0.00018556287712867005,
|
| 827 |
+
"loss": 0.3498,
|
| 828 |
+
"step": 117
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.7763157894736842,
|
| 832 |
+
"grad_norm": 0.09005908717348306,
|
| 833 |
+
"learning_rate": 0.00018516376995590187,
|
| 834 |
+
"loss": 0.343,
|
| 835 |
+
"step": 118
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.7828947368421053,
|
| 839 |
+
"grad_norm": 0.09003456135164496,
|
| 840 |
+
"learning_rate": 0.00018475966261295948,
|
| 841 |
+
"loss": 0.3708,
|
| 842 |
+
"step": 119
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.7894736842105263,
|
| 846 |
+
"grad_norm": 0.09116449509404734,
|
| 847 |
+
"learning_rate": 0.00018435057882595882,
|
| 848 |
+
"loss": 0.3548,
|
| 849 |
+
"step": 120
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.7960526315789473,
|
| 853 |
+
"grad_norm": 0.09200874674641456,
|
| 854 |
+
"learning_rate": 0.000183936542613195,
|
| 855 |
+
"loss": 0.3728,
|
| 856 |
+
"step": 121
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.8026315789473685,
|
| 860 |
+
"grad_norm": 0.08854192816932756,
|
| 861 |
+
"learning_rate": 0.0001835175782837318,
|
| 862 |
+
"loss": 0.3554,
|
| 863 |
+
"step": 122
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.8092105263157895,
|
| 867 |
+
"grad_norm": 0.0902705983711389,
|
| 868 |
+
"learning_rate": 0.00018309371043597473,
|
| 869 |
+
"loss": 0.3708,
|
| 870 |
+
"step": 123
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.8157894736842105,
|
| 874 |
+
"grad_norm": 0.08478855503527433,
|
| 875 |
+
"learning_rate": 0.0001826649639562266,
|
| 876 |
+
"loss": 0.3455,
|
| 877 |
+
"step": 124
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.8223684210526315,
|
| 881 |
+
"grad_norm": 0.08988565845179716,
|
| 882 |
+
"learning_rate": 0.0001822313640172265,
|
| 883 |
+
"loss": 0.3552,
|
| 884 |
+
"step": 125
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.8289473684210527,
|
| 888 |
+
"grad_norm": 0.08874881990127954,
|
| 889 |
+
"learning_rate": 0.00018179293607667178,
|
| 890 |
+
"loss": 0.3587,
|
| 891 |
+
"step": 126
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.8355263157894737,
|
| 895 |
+
"grad_norm": 0.09180650586751762,
|
| 896 |
+
"learning_rate": 0.00018134970587572343,
|
| 897 |
+
"loss": 0.3577,
|
| 898 |
+
"step": 127
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.8421052631578947,
|
| 902 |
+
"grad_norm": 0.09068943599513157,
|
| 903 |
+
"learning_rate": 0.00018090169943749476,
|
| 904 |
+
"loss": 0.3616,
|
| 905 |
+
"step": 128
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.8486842105263158,
|
| 909 |
+
"grad_norm": 0.09063538270613061,
|
| 910 |
+
"learning_rate": 0.00018044894306552337,
|
| 911 |
+
"loss": 0.3732,
|
| 912 |
+
"step": 129
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.8552631578947368,
|
| 916 |
+
"grad_norm": 0.0901373223301407,
|
| 917 |
+
"learning_rate": 0.00017999146334222695,
|
| 918 |
+
"loss": 0.3345,
|
| 919 |
+
"step": 130
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.8618421052631579,
|
| 923 |
+
"grad_norm": 0.0885794928608872,
|
| 924 |
+
"learning_rate": 0.00017952928712734268,
|
| 925 |
+
"loss": 0.3498,
|
| 926 |
+
"step": 131
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.868421052631579,
|
| 930 |
+
"grad_norm": 0.08860093481405781,
|
| 931 |
+
"learning_rate": 0.00017906244155634982,
|
| 932 |
+
"loss": 0.3658,
|
| 933 |
+
"step": 132
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.875,
|
| 937 |
+
"grad_norm": 0.08977436227076488,
|
| 938 |
+
"learning_rate": 0.00017859095403887696,
|
| 939 |
+
"loss": 0.3579,
|
| 940 |
+
"step": 133
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.881578947368421,
|
| 944 |
+
"grad_norm": 0.08940660194253588,
|
| 945 |
+
"learning_rate": 0.00017811485225709256,
|
| 946 |
+
"loss": 0.3537,
|
| 947 |
+
"step": 134
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.8881578947368421,
|
| 951 |
+
"grad_norm": 0.08080512906819583,
|
| 952 |
+
"learning_rate": 0.00017763416416407952,
|
| 953 |
+
"loss": 0.3379,
|
| 954 |
+
"step": 135
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.8947368421052632,
|
| 958 |
+
"grad_norm": 0.08837755359593251,
|
| 959 |
+
"learning_rate": 0.0001771489179821943,
|
| 960 |
+
"loss": 0.3524,
|
| 961 |
+
"step": 136
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.9013157894736842,
|
| 965 |
+
"grad_norm": 0.08523392262655982,
|
| 966 |
+
"learning_rate": 0.00017665914220140963,
|
| 967 |
+
"loss": 0.3386,
|
| 968 |
+
"step": 137
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.9078947368421053,
|
| 972 |
+
"grad_norm": 0.09103630477831169,
|
| 973 |
+
"learning_rate": 0.00017616486557764187,
|
| 974 |
+
"loss": 0.3659,
|
| 975 |
+
"step": 138
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.9144736842105263,
|
| 979 |
+
"grad_norm": 0.09441590695909564,
|
| 980 |
+
"learning_rate": 0.00017566611713106287,
|
| 981 |
+
"loss": 0.3507,
|
| 982 |
+
"step": 139
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.9210526315789473,
|
| 986 |
+
"grad_norm": 0.09451841778053577,
|
| 987 |
+
"learning_rate": 0.00017516292614439585,
|
| 988 |
+
"loss": 0.357,
|
| 989 |
+
"step": 140
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.9276315789473685,
|
| 993 |
+
"grad_norm": 0.08792822153458119,
|
| 994 |
+
"learning_rate": 0.00017465532216119625,
|
| 995 |
+
"loss": 0.3369,
|
| 996 |
+
"step": 141
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.9342105263157895,
|
| 1000 |
+
"grad_norm": 0.09847169501416486,
|
| 1001 |
+
"learning_rate": 0.00017414333498411733,
|
| 1002 |
+
"loss": 0.3537,
|
| 1003 |
+
"step": 142
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.9407894736842105,
|
| 1007 |
+
"grad_norm": 0.09535814851654961,
|
| 1008 |
+
"learning_rate": 0.00017362699467316005,
|
| 1009 |
+
"loss": 0.3688,
|
| 1010 |
+
"step": 143
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.9473684210526315,
|
| 1014 |
+
"grad_norm": 0.0894204821997548,
|
| 1015 |
+
"learning_rate": 0.0001731063315439084,
|
| 1016 |
+
"loss": 0.3562,
|
| 1017 |
+
"step": 144
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.9539473684210527,
|
| 1021 |
+
"grad_norm": 0.08747335966415704,
|
| 1022 |
+
"learning_rate": 0.0001725813761657495,
|
| 1023 |
+
"loss": 0.3431,
|
| 1024 |
+
"step": 145
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.9605263157894737,
|
| 1028 |
+
"grad_norm": 0.09306693938804542,
|
| 1029 |
+
"learning_rate": 0.0001720521593600787,
|
| 1030 |
+
"loss": 0.3463,
|
| 1031 |
+
"step": 146
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.9671052631578947,
|
| 1035 |
+
"grad_norm": 0.09389776996796227,
|
| 1036 |
+
"learning_rate": 0.00017151871219849008,
|
| 1037 |
+
"loss": 0.3585,
|
| 1038 |
+
"step": 147
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.9736842105263158,
|
| 1042 |
+
"grad_norm": 0.08560502715505004,
|
| 1043 |
+
"learning_rate": 0.00017098106600095205,
|
| 1044 |
+
"loss": 0.3465,
|
| 1045 |
+
"step": 148
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.9802631578947368,
|
| 1049 |
+
"grad_norm": 0.08501659921152092,
|
| 1050 |
+
"learning_rate": 0.00017043925233396854,
|
| 1051 |
+
"loss": 0.3481,
|
| 1052 |
+
"step": 149
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.9868421052631579,
|
| 1056 |
+
"grad_norm": 0.09086546243483316,
|
| 1057 |
+
"learning_rate": 0.00016989330300872576,
|
| 1058 |
+
"loss": 0.355,
|
| 1059 |
+
"step": 150
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.993421052631579,
|
| 1063 |
+
"grad_norm": 0.08269896391914834,
|
| 1064 |
+
"learning_rate": 0.00016934325007922417,
|
| 1065 |
+
"loss": 0.347,
|
| 1066 |
+
"step": 151
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 1.0,
|
| 1070 |
+
"grad_norm": 0.08946228022278917,
|
| 1071 |
+
"learning_rate": 0.000168789125840397,
|
| 1072 |
+
"loss": 0.3519,
|
| 1073 |
+
"step": 152
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 1.0,
|
| 1077 |
+
"eval_loss": 0.3528538942337036,
|
| 1078 |
+
"eval_runtime": 165.9892,
|
| 1079 |
+
"eval_samples_per_second": 30.755,
|
| 1080 |
+
"eval_steps_per_second": 0.964,
|
| 1081 |
+
"step": 152
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 1.006578947368421,
|
| 1085 |
+
"grad_norm": 0.08542638599767739,
|
| 1086 |
+
"learning_rate": 0.00016823096282621365,
|
| 1087 |
+
"loss": 0.3457,
|
| 1088 |
+
"step": 153
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 1.013157894736842,
|
| 1092 |
+
"grad_norm": 0.08540863800720608,
|
| 1093 |
+
"learning_rate": 0.0001676687938077698,
|
| 1094 |
+
"loss": 0.3323,
|
| 1095 |
+
"step": 154
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 1.019736842105263,
|
| 1099 |
+
"grad_norm": 0.08863302632538074,
|
| 1100 |
+
"learning_rate": 0.0001671026517913634,
|
| 1101 |
+
"loss": 0.3447,
|
| 1102 |
+
"step": 155
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 1.0263157894736843,
|
| 1106 |
+
"grad_norm": 0.08334205967004667,
|
| 1107 |
+
"learning_rate": 0.00016653257001655652,
|
| 1108 |
+
"loss": 0.3337,
|
| 1109 |
+
"step": 156
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.0328947368421053,
|
| 1113 |
+
"grad_norm": 0.08918626067677893,
|
| 1114 |
+
"learning_rate": 0.0001659585819542241,
|
| 1115 |
+
"loss": 0.343,
|
| 1116 |
+
"step": 157
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 1.0394736842105263,
|
| 1120 |
+
"grad_norm": 0.09996420454196991,
|
| 1121 |
+
"learning_rate": 0.00016538072130458853,
|
| 1122 |
+
"loss": 0.3529,
|
| 1123 |
+
"step": 158
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 1.0460526315789473,
|
| 1127 |
+
"grad_norm": 0.08514205522788368,
|
| 1128 |
+
"learning_rate": 0.00016479902199524116,
|
| 1129 |
+
"loss": 0.3338,
|
| 1130 |
+
"step": 159
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 1.0526315789473684,
|
| 1134 |
+
"grad_norm": 0.08426879308880401,
|
| 1135 |
+
"learning_rate": 0.00016421351817915024,
|
| 1136 |
+
"loss": 0.3347,
|
| 1137 |
+
"step": 160
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 1.0592105263157894,
|
| 1141 |
+
"grad_norm": 0.09335534193359232,
|
| 1142 |
+
"learning_rate": 0.00016362424423265598,
|
| 1143 |
+
"loss": 0.3396,
|
| 1144 |
+
"step": 161
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 1.0657894736842106,
|
| 1148 |
+
"grad_norm": 0.09585726718482877,
|
| 1149 |
+
"learning_rate": 0.00016303123475345182,
|
| 1150 |
+
"loss": 0.3471,
|
| 1151 |
+
"step": 162
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 1.0723684210526316,
|
| 1155 |
+
"grad_norm": 0.08521089701402258,
|
| 1156 |
+
"learning_rate": 0.00016243452455855342,
|
| 1157 |
+
"loss": 0.333,
|
| 1158 |
+
"step": 163
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 1.0789473684210527,
|
| 1162 |
+
"grad_norm": 0.08924725173949923,
|
| 1163 |
+
"learning_rate": 0.00016183414868225433,
|
| 1164 |
+
"loss": 0.3311,
|
| 1165 |
+
"step": 164
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 1.0855263157894737,
|
| 1169 |
+
"grad_norm": 0.08988680540431337,
|
| 1170 |
+
"learning_rate": 0.00016123014237406913,
|
| 1171 |
+
"loss": 0.3298,
|
| 1172 |
+
"step": 165
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 1.0921052631578947,
|
| 1176 |
+
"grad_norm": 0.09878832595880599,
|
| 1177 |
+
"learning_rate": 0.0001606225410966638,
|
| 1178 |
+
"loss": 0.3322,
|
| 1179 |
+
"step": 166
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 1.0986842105263157,
|
| 1183 |
+
"grad_norm": 0.08804802839862416,
|
| 1184 |
+
"learning_rate": 0.0001600113805237737,
|
| 1185 |
+
"loss": 0.3265,
|
| 1186 |
+
"step": 167
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 1.1052631578947367,
|
| 1190 |
+
"grad_norm": 0.08862630237882663,
|
| 1191 |
+
"learning_rate": 0.0001593966965381088,
|
| 1192 |
+
"loss": 0.3273,
|
| 1193 |
+
"step": 168
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 1.111842105263158,
|
| 1197 |
+
"grad_norm": 0.09796877182879883,
|
| 1198 |
+
"learning_rate": 0.00015877852522924732,
|
| 1199 |
+
"loss": 0.35,
|
| 1200 |
+
"step": 169
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 1.118421052631579,
|
| 1204 |
+
"grad_norm": 0.08598713651781048,
|
| 1205 |
+
"learning_rate": 0.00015815690289151658,
|
| 1206 |
+
"loss": 0.3297,
|
| 1207 |
+
"step": 170
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.125,
|
| 1211 |
+
"grad_norm": 0.08765944123128291,
|
| 1212 |
+
"learning_rate": 0.00015753186602186209,
|
| 1213 |
+
"loss": 0.3214,
|
| 1214 |
+
"step": 171
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 1.131578947368421,
|
| 1218 |
+
"grad_norm": 0.09209895613253609,
|
| 1219 |
+
"learning_rate": 0.00015690345131770474,
|
| 1220 |
+
"loss": 0.3315,
|
| 1221 |
+
"step": 172
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 1.138157894736842,
|
| 1225 |
+
"grad_norm": 0.09108557919395985,
|
| 1226 |
+
"learning_rate": 0.0001562716956747863,
|
| 1227 |
+
"loss": 0.3367,
|
| 1228 |
+
"step": 173
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 1.1447368421052633,
|
| 1232 |
+
"grad_norm": 0.08895539817943555,
|
| 1233 |
+
"learning_rate": 0.00015563663618500302,
|
| 1234 |
+
"loss": 0.3341,
|
| 1235 |
+
"step": 174
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 1.1513157894736843,
|
| 1239 |
+
"grad_norm": 0.08973951614406138,
|
| 1240 |
+
"learning_rate": 0.00015499831013422803,
|
| 1241 |
+
"loss": 0.3339,
|
| 1242 |
+
"step": 175
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 1.1578947368421053,
|
| 1246 |
+
"grad_norm": 0.08620142874315022,
|
| 1247 |
+
"learning_rate": 0.00015435675500012212,
|
| 1248 |
+
"loss": 0.335,
|
| 1249 |
+
"step": 176
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.1644736842105263,
|
| 1253 |
+
"grad_norm": 0.09126492238951715,
|
| 1254 |
+
"learning_rate": 0.00015371200844993332,
|
| 1255 |
+
"loss": 0.3312,
|
| 1256 |
+
"step": 177
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 1.1710526315789473,
|
| 1260 |
+
"grad_norm": 0.09587526018453588,
|
| 1261 |
+
"learning_rate": 0.00015306410833828535,
|
| 1262 |
+
"loss": 0.3294,
|
| 1263 |
+
"step": 178
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.1776315789473684,
|
| 1267 |
+
"grad_norm": 0.09087674252888882,
|
| 1268 |
+
"learning_rate": 0.00015241309270495524,
|
| 1269 |
+
"loss": 0.3372,
|
| 1270 |
+
"step": 179
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 1.1842105263157894,
|
| 1274 |
+
"grad_norm": 0.0883641678692255,
|
| 1275 |
+
"learning_rate": 0.00015175899977263964,
|
| 1276 |
+
"loss": 0.3386,
|
| 1277 |
+
"step": 180
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 1.1907894736842106,
|
| 1281 |
+
"grad_norm": 0.09379689103796256,
|
| 1282 |
+
"learning_rate": 0.00015110186794471103,
|
| 1283 |
+
"loss": 0.3377,
|
| 1284 |
+
"step": 181
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 1.1973684210526316,
|
| 1288 |
+
"grad_norm": 0.09279921976435307,
|
| 1289 |
+
"learning_rate": 0.00015044173580296266,
|
| 1290 |
+
"loss": 0.3331,
|
| 1291 |
+
"step": 182
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 1.2039473684210527,
|
| 1295 |
+
"grad_norm": 0.09523467660834971,
|
| 1296 |
+
"learning_rate": 0.0001497786421053434,
|
| 1297 |
+
"loss": 0.3468,
|
| 1298 |
+
"step": 183
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 1.2105263157894737,
|
| 1302 |
+
"grad_norm": 0.09692763912842987,
|
| 1303 |
+
"learning_rate": 0.00014911262578368232,
|
| 1304 |
+
"loss": 0.3254,
|
| 1305 |
+
"step": 184
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 1.2171052631578947,
|
| 1309 |
+
"grad_norm": 0.0905572094984609,
|
| 1310 |
+
"learning_rate": 0.0001484437259414027,
|
| 1311 |
+
"loss": 0.3428,
|
| 1312 |
+
"step": 185
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 1.2236842105263157,
|
| 1316 |
+
"grad_norm": 0.0895303347607193,
|
| 1317 |
+
"learning_rate": 0.0001477719818512263,
|
| 1318 |
+
"loss": 0.3343,
|
| 1319 |
+
"step": 186
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 1.2302631578947367,
|
| 1323 |
+
"grad_norm": 0.09084314989641666,
|
| 1324 |
+
"learning_rate": 0.00014709743295286752,
|
| 1325 |
+
"loss": 0.3428,
|
| 1326 |
+
"step": 187
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 1.236842105263158,
|
| 1330 |
+
"grad_norm": 0.09188016999870963,
|
| 1331 |
+
"learning_rate": 0.0001464201188507178,
|
| 1332 |
+
"loss": 0.3392,
|
| 1333 |
+
"step": 188
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 1.243421052631579,
|
| 1337 |
+
"grad_norm": 0.09034095867956521,
|
| 1338 |
+
"learning_rate": 0.00014574007931152037,
|
| 1339 |
+
"loss": 0.3305,
|
| 1340 |
+
"step": 189
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 1.25,
|
| 1344 |
+
"grad_norm": 0.09422416865836616,
|
| 1345 |
+
"learning_rate": 0.00014505735426203543,
|
| 1346 |
+
"loss": 0.3519,
|
| 1347 |
+
"step": 190
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 1.256578947368421,
|
| 1351 |
+
"grad_norm": 0.09146229714998408,
|
| 1352 |
+
"learning_rate": 0.00014437198378669598,
|
| 1353 |
+
"loss": 0.3246,
|
| 1354 |
+
"step": 191
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 1.263157894736842,
|
| 1358 |
+
"grad_norm": 0.09150826987109315,
|
| 1359 |
+
"learning_rate": 0.00014368400812525434,
|
| 1360 |
+
"loss": 0.3385,
|
| 1361 |
+
"step": 192
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 1.2697368421052633,
|
| 1365 |
+
"grad_norm": 0.09170822182104935,
|
| 1366 |
+
"learning_rate": 0.00014299346767041954,
|
| 1367 |
+
"loss": 0.3202,
|
| 1368 |
+
"step": 193
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 1.2763157894736843,
|
| 1372 |
+
"grad_norm": 0.08611790992292927,
|
| 1373 |
+
"learning_rate": 0.0001423004029654859,
|
| 1374 |
+
"loss": 0.3309,
|
| 1375 |
+
"step": 194
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 1.2828947368421053,
|
| 1379 |
+
"grad_norm": 0.09296272729378167,
|
| 1380 |
+
"learning_rate": 0.00014160485470195245,
|
| 1381 |
+
"loss": 0.3334,
|
| 1382 |
+
"step": 195
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 1.2894736842105263,
|
| 1386 |
+
"grad_norm": 0.09167809377202646,
|
| 1387 |
+
"learning_rate": 0.00014090686371713402,
|
| 1388 |
+
"loss": 0.3311,
|
| 1389 |
+
"step": 196
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 1.2960526315789473,
|
| 1393 |
+
"grad_norm": 0.08975846040243303,
|
| 1394 |
+
"learning_rate": 0.0001402064709917634,
|
| 1395 |
+
"loss": 0.3435,
|
| 1396 |
+
"step": 197
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 1.3026315789473684,
|
| 1400 |
+
"grad_norm": 0.09083148769815752,
|
| 1401 |
+
"learning_rate": 0.00013950371764758542,
|
| 1402 |
+
"loss": 0.3252,
|
| 1403 |
+
"step": 198
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 1.3092105263157894,
|
| 1407 |
+
"grad_norm": 0.09070017702344292,
|
| 1408 |
+
"learning_rate": 0.0001387986449449425,
|
| 1409 |
+
"loss": 0.3417,
|
| 1410 |
+
"step": 199
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 1.3157894736842106,
|
| 1414 |
+
"grad_norm": 0.09406507310996726,
|
| 1415 |
+
"learning_rate": 0.00013809129428035227,
|
| 1416 |
+
"loss": 0.3474,
|
| 1417 |
+
"step": 200
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.3223684210526316,
|
| 1421 |
+
"grad_norm": 0.08965350812782857,
|
| 1422 |
+
"learning_rate": 0.00013738170718407687,
|
| 1423 |
+
"loss": 0.3309,
|
| 1424 |
+
"step": 201
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 1.3289473684210527,
|
| 1428 |
+
"grad_norm": 0.08959778822119953,
|
| 1429 |
+
"learning_rate": 0.00013666992531768482,
|
| 1430 |
+
"loss": 0.3402,
|
| 1431 |
+
"step": 202
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 1.3355263157894737,
|
| 1435 |
+
"grad_norm": 0.092084592265277,
|
| 1436 |
+
"learning_rate": 0.0001359559904716048,
|
| 1437 |
+
"loss": 0.3375,
|
| 1438 |
+
"step": 203
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 1.3421052631578947,
|
| 1442 |
+
"grad_norm": 0.09457199959869429,
|
| 1443 |
+
"learning_rate": 0.0001352399445626722,
|
| 1444 |
+
"loss": 0.3389,
|
| 1445 |
+
"step": 204
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 1.3486842105263157,
|
| 1449 |
+
"grad_norm": 0.09197177008454786,
|
| 1450 |
+
"learning_rate": 0.00013452182963166794,
|
| 1451 |
+
"loss": 0.3374,
|
| 1452 |
+
"step": 205
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 1.3552631578947367,
|
| 1456 |
+
"grad_norm": 0.09399945833367072,
|
| 1457 |
+
"learning_rate": 0.00013380168784085027,
|
| 1458 |
+
"loss": 0.3393,
|
| 1459 |
+
"step": 206
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 1.361842105263158,
|
| 1463 |
+
"grad_norm": 0.08785100494819592,
|
| 1464 |
+
"learning_rate": 0.00013307956147147923,
|
| 1465 |
+
"loss": 0.339,
|
| 1466 |
+
"step": 207
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 1.368421052631579,
|
| 1470 |
+
"grad_norm": 0.09128615601816815,
|
| 1471 |
+
"learning_rate": 0.00013235549292133425,
|
| 1472 |
+
"loss": 0.3319,
|
| 1473 |
+
"step": 208
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 1.375,
|
| 1477 |
+
"grad_norm": 0.08674366087344106,
|
| 1478 |
+
"learning_rate": 0.00013162952470222486,
|
| 1479 |
+
"loss": 0.3262,
|
| 1480 |
+
"step": 209
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 1.381578947368421,
|
| 1484 |
+
"grad_norm": 0.09650968662374713,
|
| 1485 |
+
"learning_rate": 0.00013090169943749476,
|
| 1486 |
+
"loss": 0.3218,
|
| 1487 |
+
"step": 210
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 1.388157894736842,
|
| 1491 |
+
"grad_norm": 0.09647571976953839,
|
| 1492 |
+
"learning_rate": 0.00013017205985951926,
|
| 1493 |
+
"loss": 0.3297,
|
| 1494 |
+
"step": 211
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 1.3947368421052633,
|
| 1498 |
+
"grad_norm": 0.08840745611440821,
|
| 1499 |
+
"learning_rate": 0.00012944064880719632,
|
| 1500 |
+
"loss": 0.3183,
|
| 1501 |
+
"step": 212
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.4013157894736843,
|
| 1505 |
+
"grad_norm": 0.09429827798710846,
|
| 1506 |
+
"learning_rate": 0.0001287075092234316,
|
| 1507 |
+
"loss": 0.3441,
|
| 1508 |
+
"step": 213
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 1.4078947368421053,
|
| 1512 |
+
"grad_norm": 0.09400151382755063,
|
| 1513 |
+
"learning_rate": 0.00012797268415261682,
|
| 1514 |
+
"loss": 0.3374,
|
| 1515 |
+
"step": 214
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 1.4144736842105263,
|
| 1519 |
+
"grad_norm": 0.094388991307675,
|
| 1520 |
+
"learning_rate": 0.00012723621673810275,
|
| 1521 |
+
"loss": 0.3425,
|
| 1522 |
+
"step": 215
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 1.4210526315789473,
|
| 1526 |
+
"grad_norm": 0.09040719138127835,
|
| 1527 |
+
"learning_rate": 0.0001264981502196662,
|
| 1528 |
+
"loss": 0.3466,
|
| 1529 |
+
"step": 216
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 1.4276315789473684,
|
| 1533 |
+
"grad_norm": 0.08410156088705376,
|
| 1534 |
+
"learning_rate": 0.0001257585279309711,
|
| 1535 |
+
"loss": 0.3293,
|
| 1536 |
+
"step": 217
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 1.4342105263157894,
|
| 1540 |
+
"grad_norm": 0.09012762447734174,
|
| 1541 |
+
"learning_rate": 0.00012501739329702453,
|
| 1542 |
+
"loss": 0.3241,
|
| 1543 |
+
"step": 218
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 1.4407894736842106,
|
| 1547 |
+
"grad_norm": 0.08981619956473173,
|
| 1548 |
+
"learning_rate": 0.00012427478983162695,
|
| 1549 |
+
"loss": 0.3309,
|
| 1550 |
+
"step": 219
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 1.4473684210526316,
|
| 1554 |
+
"grad_norm": 0.09690483937945461,
|
| 1555 |
+
"learning_rate": 0.0001235307611348174,
|
| 1556 |
+
"loss": 0.3353,
|
| 1557 |
+
"step": 220
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 1.4539473684210527,
|
| 1561 |
+
"grad_norm": 0.09199281173602707,
|
| 1562 |
+
"learning_rate": 0.00012278535089031378,
|
| 1563 |
+
"loss": 0.3478,
|
| 1564 |
+
"step": 221
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 1.4605263157894737,
|
| 1568 |
+
"grad_norm": 0.09383616607484997,
|
| 1569 |
+
"learning_rate": 0.0001220386028629479,
|
| 1570 |
+
"loss": 0.3226,
|
| 1571 |
+
"step": 222
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.4671052631578947,
|
| 1575 |
+
"grad_norm": 0.0888565150769418,
|
| 1576 |
+
"learning_rate": 0.00012129056089609609,
|
| 1577 |
+
"loss": 0.3368,
|
| 1578 |
+
"step": 223
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 1.4736842105263157,
|
| 1582 |
+
"grad_norm": 0.08846739233573914,
|
| 1583 |
+
"learning_rate": 0.00012054126890910498,
|
| 1584 |
+
"loss": 0.3363,
|
| 1585 |
+
"step": 224
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.4802631578947367,
|
| 1589 |
+
"grad_norm": 0.08828966342515125,
|
| 1590 |
+
"learning_rate": 0.00011979077089471287,
|
| 1591 |
+
"loss": 0.3426,
|
| 1592 |
+
"step": 225
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 1.486842105263158,
|
| 1596 |
+
"grad_norm": 0.09066446045227519,
|
| 1597 |
+
"learning_rate": 0.00011903911091646684,
|
| 1598 |
+
"loss": 0.3305,
|
| 1599 |
+
"step": 226
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 1.493421052631579,
|
| 1603 |
+
"grad_norm": 0.0885766678361643,
|
| 1604 |
+
"learning_rate": 0.00011828633310613568,
|
| 1605 |
+
"loss": 0.3359,
|
| 1606 |
+
"step": 227
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 1.5,
|
| 1610 |
+
"grad_norm": 0.08447486357210816,
|
| 1611 |
+
"learning_rate": 0.00011753248166111878,
|
| 1612 |
+
"loss": 0.342,
|
| 1613 |
+
"step": 228
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 1.506578947368421,
|
| 1617 |
+
"grad_norm": 0.08995420125881844,
|
| 1618 |
+
"learning_rate": 0.00011677760084185123,
|
| 1619 |
+
"loss": 0.3239,
|
| 1620 |
+
"step": 229
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 1.513157894736842,
|
| 1624 |
+
"grad_norm": 0.08976025618793185,
|
| 1625 |
+
"learning_rate": 0.0001160217349692051,
|
| 1626 |
+
"loss": 0.3234,
|
| 1627 |
+
"step": 230
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 1.5197368421052633,
|
| 1631 |
+
"grad_norm": 0.08403343805686422,
|
| 1632 |
+
"learning_rate": 0.00011526492842188745,
|
| 1633 |
+
"loss": 0.3241,
|
| 1634 |
+
"step": 231
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 1.526315789473684,
|
| 1638 |
+
"grad_norm": 0.08731505199027503,
|
| 1639 |
+
"learning_rate": 0.00011450722563383449,
|
| 1640 |
+
"loss": 0.323,
|
| 1641 |
+
"step": 232
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 1.5328947368421053,
|
| 1645 |
+
"grad_norm": 0.08919530304419715,
|
| 1646 |
+
"learning_rate": 0.00011374867109160294,
|
| 1647 |
+
"loss": 0.335,
|
| 1648 |
+
"step": 233
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 1.5394736842105263,
|
| 1652 |
+
"grad_norm": 0.08543608378136278,
|
| 1653 |
+
"learning_rate": 0.00011298930933175804,
|
| 1654 |
+
"loss": 0.3257,
|
| 1655 |
+
"step": 234
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 1.5460526315789473,
|
| 1659 |
+
"grad_norm": 0.0843872841143867,
|
| 1660 |
+
"learning_rate": 0.00011222918493825876,
|
| 1661 |
+
"loss": 0.322,
|
| 1662 |
+
"step": 235
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 1.5526315789473686,
|
| 1666 |
+
"grad_norm": 0.09015699624357591,
|
| 1667 |
+
"learning_rate": 0.00011146834253984006,
|
| 1668 |
+
"loss": 0.3393,
|
| 1669 |
+
"step": 236
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 1.5592105263157894,
|
| 1673 |
+
"grad_norm": 0.08527720509613376,
|
| 1674 |
+
"learning_rate": 0.00011070682680739275,
|
| 1675 |
+
"loss": 0.3228,
|
| 1676 |
+
"step": 237
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 1.5657894736842106,
|
| 1680 |
+
"grad_norm": 0.08882512008213382,
|
| 1681 |
+
"learning_rate": 0.00010994468245134071,
|
| 1682 |
+
"loss": 0.3307,
|
| 1683 |
+
"step": 238
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 1.5723684210526314,
|
| 1687 |
+
"grad_norm": 0.0873699179145412,
|
| 1688 |
+
"learning_rate": 0.00010918195421901583,
|
| 1689 |
+
"loss": 0.3315,
|
| 1690 |
+
"step": 239
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 1.5789473684210527,
|
| 1694 |
+
"grad_norm": 0.0856032512902339,
|
| 1695 |
+
"learning_rate": 0.00010841868689203071,
|
| 1696 |
+
"loss": 0.3249,
|
| 1697 |
+
"step": 240
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 1.5855263157894737,
|
| 1701 |
+
"grad_norm": 0.08508957717963754,
|
| 1702 |
+
"learning_rate": 0.0001076549252836496,
|
| 1703 |
+
"loss": 0.3198,
|
| 1704 |
+
"step": 241
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 1.5921052631578947,
|
| 1708 |
+
"grad_norm": 0.08646553066877591,
|
| 1709 |
+
"learning_rate": 0.00010689071423615709,
|
| 1710 |
+
"loss": 0.3201,
|
| 1711 |
+
"step": 242
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 1.598684210526316,
|
| 1715 |
+
"grad_norm": 0.08651961917444705,
|
| 1716 |
+
"learning_rate": 0.00010612609861822556,
|
| 1717 |
+
"loss": 0.3273,
|
| 1718 |
+
"step": 243
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 1.6052631578947367,
|
| 1722 |
+
"grad_norm": 0.08675562470697942,
|
| 1723 |
+
"learning_rate": 0.00010536112332228057,
|
| 1724 |
+
"loss": 0.3344,
|
| 1725 |
+
"step": 244
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 1.611842105263158,
|
| 1729 |
+
"grad_norm": 0.09458426431125393,
|
| 1730 |
+
"learning_rate": 0.00010459583326186533,
|
| 1731 |
+
"loss": 0.3411,
|
| 1732 |
+
"step": 245
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 1.618421052631579,
|
| 1736 |
+
"grad_norm": 0.09120208666424241,
|
| 1737 |
+
"learning_rate": 0.00010383027336900355,
|
| 1738 |
+
"loss": 0.3359,
|
| 1739 |
+
"step": 246
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 1.625,
|
| 1743 |
+
"grad_norm": 0.09024986688209398,
|
| 1744 |
+
"learning_rate": 0.00010306448859156155,
|
| 1745 |
+
"loss": 0.3244,
|
| 1746 |
+
"step": 247
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 1.631578947368421,
|
| 1750 |
+
"grad_norm": 0.08766784765589664,
|
| 1751 |
+
"learning_rate": 0.00010229852389060905,
|
| 1752 |
+
"loss": 0.3165,
|
| 1753 |
+
"step": 248
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 1.638157894736842,
|
| 1757 |
+
"grad_norm": 0.08799575986715236,
|
| 1758 |
+
"learning_rate": 0.00010153242423777965,
|
| 1759 |
+
"loss": 0.3274,
|
| 1760 |
+
"step": 249
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 1.6447368421052633,
|
| 1764 |
+
"grad_norm": 0.0920628008990826,
|
| 1765 |
+
"learning_rate": 0.00010076623461263018,
|
| 1766 |
+
"loss": 0.3228,
|
| 1767 |
+
"step": 250
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 1.651315789473684,
|
| 1771 |
+
"grad_norm": 0.09132220875461658,
|
| 1772 |
+
"learning_rate": 0.0001,
|
| 1773 |
+
"loss": 0.3291,
|
| 1774 |
+
"step": 251
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 1.6578947368421053,
|
| 1778 |
+
"grad_norm": 0.0861257482862008,
|
| 1779 |
+
"learning_rate": 9.923376538736985e-05,
|
| 1780 |
+
"loss": 0.3257,
|
| 1781 |
+
"step": 252
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 1.6644736842105263,
|
| 1785 |
+
"grad_norm": 0.08970245386690526,
|
| 1786 |
+
"learning_rate": 9.846757576222038e-05,
|
| 1787 |
+
"loss": 0.333,
|
| 1788 |
+
"step": 253
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 1.6710526315789473,
|
| 1792 |
+
"grad_norm": 0.08745399881851679,
|
| 1793 |
+
"learning_rate": 9.770147610939097e-05,
|
| 1794 |
+
"loss": 0.3243,
|
| 1795 |
+
"step": 254
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 1.6776315789473686,
|
| 1799 |
+
"grad_norm": 0.0942879278062166,
|
| 1800 |
+
"learning_rate": 9.693551140843847e-05,
|
| 1801 |
+
"loss": 0.3305,
|
| 1802 |
+
"step": 255
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 1.6842105263157894,
|
| 1806 |
+
"grad_norm": 0.08491033600054997,
|
| 1807 |
+
"learning_rate": 9.616972663099647e-05,
|
| 1808 |
+
"loss": 0.3219,
|
| 1809 |
+
"step": 256
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 1.6907894736842106,
|
| 1813 |
+
"grad_norm": 0.0824170398183598,
|
| 1814 |
+
"learning_rate": 9.54041667381347e-05,
|
| 1815 |
+
"loss": 0.3067,
|
| 1816 |
+
"step": 257
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 1.6973684210526314,
|
| 1820 |
+
"grad_norm": 0.08689158883776966,
|
| 1821 |
+
"learning_rate": 9.463887667771946e-05,
|
| 1822 |
+
"loss": 0.3243,
|
| 1823 |
+
"step": 258
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 1.7039473684210527,
|
| 1827 |
+
"grad_norm": 0.09175347122606316,
|
| 1828 |
+
"learning_rate": 9.387390138177448e-05,
|
| 1829 |
+
"loss": 0.3315,
|
| 1830 |
+
"step": 259
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 1.7105263157894737,
|
| 1834 |
+
"grad_norm": 0.08682953977254314,
|
| 1835 |
+
"learning_rate": 9.310928576384293e-05,
|
| 1836 |
+
"loss": 0.3183,
|
| 1837 |
+
"step": 260
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 1.7171052631578947,
|
| 1841 |
+
"grad_norm": 0.0844793823883941,
|
| 1842 |
+
"learning_rate": 9.234507471635043e-05,
|
| 1843 |
+
"loss": 0.3262,
|
| 1844 |
+
"step": 261
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 1.723684210526316,
|
| 1848 |
+
"grad_norm": 0.08904925607635152,
|
| 1849 |
+
"learning_rate": 9.15813131079693e-05,
|
| 1850 |
+
"loss": 0.3318,
|
| 1851 |
+
"step": 262
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 1.7302631578947367,
|
| 1855 |
+
"grad_norm": 0.08897767966029022,
|
| 1856 |
+
"learning_rate": 9.08180457809842e-05,
|
| 1857 |
+
"loss": 0.3338,
|
| 1858 |
+
"step": 263
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 1.736842105263158,
|
| 1862 |
+
"grad_norm": 0.08418525864024147,
|
| 1863 |
+
"learning_rate": 9.005531754865929e-05,
|
| 1864 |
+
"loss": 0.3224,
|
| 1865 |
+
"step": 264
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 1.743421052631579,
|
| 1869 |
+
"grad_norm": 0.08369708951027545,
|
| 1870 |
+
"learning_rate": 8.929317319260726e-05,
|
| 1871 |
+
"loss": 0.3344,
|
| 1872 |
+
"step": 265
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 1.75,
|
| 1876 |
+
"grad_norm": 0.08344688234687514,
|
| 1877 |
+
"learning_rate": 8.853165746015997e-05,
|
| 1878 |
+
"loss": 0.3186,
|
| 1879 |
+
"step": 266
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 1.756578947368421,
|
| 1883 |
+
"grad_norm": 0.08882311328649488,
|
| 1884 |
+
"learning_rate": 8.777081506174127e-05,
|
| 1885 |
+
"loss": 0.3189,
|
| 1886 |
+
"step": 267
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 1.763157894736842,
|
| 1890 |
+
"grad_norm": 0.08911077215219455,
|
| 1891 |
+
"learning_rate": 8.701069066824194e-05,
|
| 1892 |
+
"loss": 0.347,
|
| 1893 |
+
"step": 268
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 1.7697368421052633,
|
| 1897 |
+
"grad_norm": 0.08845395362382234,
|
| 1898 |
+
"learning_rate": 8.625132890839706e-05,
|
| 1899 |
+
"loss": 0.3401,
|
| 1900 |
+
"step": 269
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 1.776315789473684,
|
| 1904 |
+
"grad_norm": 0.0890814727471271,
|
| 1905 |
+
"learning_rate": 8.549277436616551e-05,
|
| 1906 |
+
"loss": 0.3357,
|
| 1907 |
+
"step": 270
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 1.7828947368421053,
|
| 1911 |
+
"grad_norm": 0.08688996127676986,
|
| 1912 |
+
"learning_rate": 8.473507157811254e-05,
|
| 1913 |
+
"loss": 0.3255,
|
| 1914 |
+
"step": 271
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 1.7894736842105263,
|
| 1918 |
+
"grad_norm": 0.08839344810308368,
|
| 1919 |
+
"learning_rate": 8.397826503079489e-05,
|
| 1920 |
+
"loss": 0.3485,
|
| 1921 |
+
"step": 272
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 1.7960526315789473,
|
| 1925 |
+
"grad_norm": 0.08616082603732687,
|
| 1926 |
+
"learning_rate": 8.32223991581488e-05,
|
| 1927 |
+
"loss": 0.3333,
|
| 1928 |
+
"step": 273
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 1.8026315789473686,
|
| 1932 |
+
"grad_norm": 0.0873042236962117,
|
| 1933 |
+
"learning_rate": 8.246751833888122e-05,
|
| 1934 |
+
"loss": 0.3293,
|
| 1935 |
+
"step": 274
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 1.8092105263157894,
|
| 1939 |
+
"grad_norm": 0.08637643463516168,
|
| 1940 |
+
"learning_rate": 8.171366689386432e-05,
|
| 1941 |
+
"loss": 0.3366,
|
| 1942 |
+
"step": 275
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 1.8157894736842106,
|
| 1946 |
+
"grad_norm": 0.08702343026083556,
|
| 1947 |
+
"learning_rate": 8.096088908353315e-05,
|
| 1948 |
+
"loss": 0.317,
|
| 1949 |
+
"step": 276
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 1.8223684210526314,
|
| 1953 |
+
"grad_norm": 0.08671977794554171,
|
| 1954 |
+
"learning_rate": 8.020922910528717e-05,
|
| 1955 |
+
"loss": 0.3243,
|
| 1956 |
+
"step": 277
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 1.8289473684210527,
|
| 1960 |
+
"grad_norm": 0.08800970811383262,
|
| 1961 |
+
"learning_rate": 7.945873109089503e-05,
|
| 1962 |
+
"loss": 0.328,
|
| 1963 |
+
"step": 278
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 1.8355263157894737,
|
| 1967 |
+
"grad_norm": 0.08715460079863534,
|
| 1968 |
+
"learning_rate": 7.870943910390392e-05,
|
| 1969 |
+
"loss": 0.3393,
|
| 1970 |
+
"step": 279
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 1.8421052631578947,
|
| 1974 |
+
"grad_norm": 0.08522036841567814,
|
| 1975 |
+
"learning_rate": 7.796139713705213e-05,
|
| 1976 |
+
"loss": 0.3264,
|
| 1977 |
+
"step": 280
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 1.848684210526316,
|
| 1981 |
+
"grad_norm": 0.08867443005533418,
|
| 1982 |
+
"learning_rate": 7.721464910968627e-05,
|
| 1983 |
+
"loss": 0.3245,
|
| 1984 |
+
"step": 281
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 1.8552631578947367,
|
| 1988 |
+
"grad_norm": 0.085291540055969,
|
| 1989 |
+
"learning_rate": 7.646923886518264e-05,
|
| 1990 |
+
"loss": 0.3193,
|
| 1991 |
+
"step": 282
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 1.861842105263158,
|
| 1995 |
+
"grad_norm": 0.0865604586807815,
|
| 1996 |
+
"learning_rate": 7.57252101683731e-05,
|
| 1997 |
+
"loss": 0.3265,
|
| 1998 |
+
"step": 283
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 1.868421052631579,
|
| 2002 |
+
"grad_norm": 0.08796722757240084,
|
| 2003 |
+
"learning_rate": 7.49826067029755e-05,
|
| 2004 |
+
"loss": 0.3315,
|
| 2005 |
+
"step": 284
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 1.875,
|
| 2009 |
+
"grad_norm": 0.08552421405911567,
|
| 2010 |
+
"learning_rate": 7.424147206902891e-05,
|
| 2011 |
+
"loss": 0.3193,
|
| 2012 |
+
"step": 285
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 1.881578947368421,
|
| 2016 |
+
"grad_norm": 0.08359126397078406,
|
| 2017 |
+
"learning_rate": 7.350184978033386e-05,
|
| 2018 |
+
"loss": 0.3283,
|
| 2019 |
+
"step": 286
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 1.888157894736842,
|
| 2023 |
+
"grad_norm": 0.09295820408600056,
|
| 2024 |
+
"learning_rate": 7.276378326189728e-05,
|
| 2025 |
+
"loss": 0.34,
|
| 2026 |
+
"step": 287
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 1.8947368421052633,
|
| 2030 |
+
"grad_norm": 0.08607086022906524,
|
| 2031 |
+
"learning_rate": 7.202731584738322e-05,
|
| 2032 |
+
"loss": 0.3224,
|
| 2033 |
+
"step": 288
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 1.901315789473684,
|
| 2037 |
+
"grad_norm": 0.08511420128792765,
|
| 2038 |
+
"learning_rate": 7.129249077656843e-05,
|
| 2039 |
+
"loss": 0.3297,
|
| 2040 |
+
"step": 289
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 1.9078947368421053,
|
| 2044 |
+
"grad_norm": 0.08811788288471267,
|
| 2045 |
+
"learning_rate": 7.055935119280369e-05,
|
| 2046 |
+
"loss": 0.327,
|
| 2047 |
+
"step": 290
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 1.9144736842105263,
|
| 2051 |
+
"grad_norm": 0.08477145974314237,
|
| 2052 |
+
"learning_rate": 6.982794014048077e-05,
|
| 2053 |
+
"loss": 0.3161,
|
| 2054 |
+
"step": 291
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 1.9210526315789473,
|
| 2058 |
+
"grad_norm": 0.08372118058279919,
|
| 2059 |
+
"learning_rate": 6.909830056250527e-05,
|
| 2060 |
+
"loss": 0.3306,
|
| 2061 |
+
"step": 292
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 1.9276315789473686,
|
| 2065 |
+
"grad_norm": 0.08686106996534168,
|
| 2066 |
+
"learning_rate": 6.837047529777516e-05,
|
| 2067 |
+
"loss": 0.32,
|
| 2068 |
+
"step": 293
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 1.9342105263157894,
|
| 2072 |
+
"grad_norm": 0.08515118561718467,
|
| 2073 |
+
"learning_rate": 6.764450707866576e-05,
|
| 2074 |
+
"loss": 0.3252,
|
| 2075 |
+
"step": 294
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 1.9407894736842106,
|
| 2079 |
+
"grad_norm": 0.09265542100862634,
|
| 2080 |
+
"learning_rate": 6.692043852852079e-05,
|
| 2081 |
+
"loss": 0.3325,
|
| 2082 |
+
"step": 295
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 1.9473684210526314,
|
| 2086 |
+
"grad_norm": 0.09595543410539613,
|
| 2087 |
+
"learning_rate": 6.619831215914974e-05,
|
| 2088 |
+
"loss": 0.3333,
|
| 2089 |
+
"step": 296
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"epoch": 1.9539473684210527,
|
| 2093 |
+
"grad_norm": 0.08927193916247748,
|
| 2094 |
+
"learning_rate": 6.547817036833208e-05,
|
| 2095 |
+
"loss": 0.34,
|
| 2096 |
+
"step": 297
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 1.9605263157894737,
|
| 2100 |
+
"grad_norm": 0.08244175050863302,
|
| 2101 |
+
"learning_rate": 6.476005543732782e-05,
|
| 2102 |
+
"loss": 0.3206,
|
| 2103 |
+
"step": 298
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 1.9671052631578947,
|
| 2107 |
+
"grad_norm": 0.08541613997019516,
|
| 2108 |
+
"learning_rate": 6.404400952839522e-05,
|
| 2109 |
+
"loss": 0.3095,
|
| 2110 |
+
"step": 299
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 1.973684210526316,
|
| 2114 |
+
"grad_norm": 0.08585952116970799,
|
| 2115 |
+
"learning_rate": 6.333007468231521e-05,
|
| 2116 |
+
"loss": 0.3143,
|
| 2117 |
+
"step": 300
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 1.9802631578947367,
|
| 2121 |
+
"grad_norm": 0.09128909476408122,
|
| 2122 |
+
"learning_rate": 6.261829281592314e-05,
|
| 2123 |
+
"loss": 0.3345,
|
| 2124 |
+
"step": 301
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 1.986842105263158,
|
| 2128 |
+
"grad_norm": 0.09002004932775928,
|
| 2129 |
+
"learning_rate": 6.190870571964774e-05,
|
| 2130 |
+
"loss": 0.3257,
|
| 2131 |
+
"step": 302
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 1.993421052631579,
|
| 2135 |
+
"grad_norm": 0.08541319081987805,
|
| 2136 |
+
"learning_rate": 6.120135505505748e-05,
|
| 2137 |
+
"loss": 0.3269,
|
| 2138 |
+
"step": 303
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 2.0,
|
| 2142 |
+
"grad_norm": 0.08848658071924796,
|
| 2143 |
+
"learning_rate": 6.049628235241459e-05,
|
| 2144 |
+
"loss": 0.3233,
|
| 2145 |
+
"step": 304
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 2.0,
|
| 2149 |
+
"eval_loss": 0.33567601442337036,
|
| 2150 |
+
"eval_runtime": 160.3905,
|
| 2151 |
+
"eval_samples_per_second": 31.829,
|
| 2152 |
+
"eval_steps_per_second": 0.998,
|
| 2153 |
+
"step": 304
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 2.0065789473684212,
|
| 2157 |
+
"grad_norm": 0.08452043290010046,
|
| 2158 |
+
"learning_rate": 5.9793529008236625e-05,
|
| 2159 |
+
"loss": 0.308,
|
| 2160 |
+
"step": 305
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 2.013157894736842,
|
| 2164 |
+
"grad_norm": 0.084981364890352,
|
| 2165 |
+
"learning_rate": 5.909313628286601e-05,
|
| 2166 |
+
"loss": 0.3138,
|
| 2167 |
+
"step": 306
|
| 2168 |
+
},
|
| 2169 |
+
{
|
| 2170 |
+
"epoch": 2.0197368421052633,
|
| 2171 |
+
"grad_norm": 0.08373368193427122,
|
| 2172 |
+
"learning_rate": 5.839514529804757e-05,
|
| 2173 |
+
"loss": 0.3051,
|
| 2174 |
+
"step": 307
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 2.026315789473684,
|
| 2178 |
+
"grad_norm": 0.08658031473872292,
|
| 2179 |
+
"learning_rate": 5.769959703451413e-05,
|
| 2180 |
+
"loss": 0.3149,
|
| 2181 |
+
"step": 308
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 2.0328947368421053,
|
| 2185 |
+
"grad_norm": 0.0872119478961657,
|
| 2186 |
+
"learning_rate": 5.7006532329580464e-05,
|
| 2187 |
+
"loss": 0.2952,
|
| 2188 |
+
"step": 309
|
| 2189 |
+
},
|
| 2190 |
+
{
|
| 2191 |
+
"epoch": 2.039473684210526,
|
| 2192 |
+
"grad_norm": 0.08985312096854285,
|
| 2193 |
+
"learning_rate": 5.63159918747457e-05,
|
| 2194 |
+
"loss": 0.31,
|
| 2195 |
+
"step": 310
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 2.0460526315789473,
|
| 2199 |
+
"grad_norm": 0.08981442223402118,
|
| 2200 |
+
"learning_rate": 5.5628016213304025e-05,
|
| 2201 |
+
"loss": 0.2987,
|
| 2202 |
+
"step": 311
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 2.0526315789473686,
|
| 2206 |
+
"grad_norm": 0.08888884200687737,
|
| 2207 |
+
"learning_rate": 5.49426457379646e-05,
|
| 2208 |
+
"loss": 0.314,
|
| 2209 |
+
"step": 312
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 2.0592105263157894,
|
| 2213 |
+
"grad_norm": 0.09263884628501892,
|
| 2214 |
+
"learning_rate": 5.425992068847965e-05,
|
| 2215 |
+
"loss": 0.3105,
|
| 2216 |
+
"step": 313
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 2.0657894736842106,
|
| 2220 |
+
"grad_norm": 0.09695553478937406,
|
| 2221 |
+
"learning_rate": 5.3579881149282204e-05,
|
| 2222 |
+
"loss": 0.2961,
|
| 2223 |
+
"step": 314
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 2.0723684210526314,
|
| 2227 |
+
"grad_norm": 0.09301145489202128,
|
| 2228 |
+
"learning_rate": 5.29025670471325e-05,
|
| 2229 |
+
"loss": 0.3121,
|
| 2230 |
+
"step": 315
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 2.0789473684210527,
|
| 2234 |
+
"grad_norm": 0.09056322468061577,
|
| 2235 |
+
"learning_rate": 5.222801814877369e-05,
|
| 2236 |
+
"loss": 0.3084,
|
| 2237 |
+
"step": 316
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 2.085526315789474,
|
| 2241 |
+
"grad_norm": 0.09036800264131283,
|
| 2242 |
+
"learning_rate": 5.155627405859731e-05,
|
| 2243 |
+
"loss": 0.3068,
|
| 2244 |
+
"step": 317
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 2.0921052631578947,
|
| 2248 |
+
"grad_norm": 0.09181257128694434,
|
| 2249 |
+
"learning_rate": 5.088737421631767e-05,
|
| 2250 |
+
"loss": 0.3113,
|
| 2251 |
+
"step": 318
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 2.098684210526316,
|
| 2255 |
+
"grad_norm": 0.09316264301444047,
|
| 2256 |
+
"learning_rate": 5.0221357894656605e-05,
|
| 2257 |
+
"loss": 0.3067,
|
| 2258 |
+
"step": 319
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 2.1052631578947367,
|
| 2262 |
+
"grad_norm": 0.09246714604570534,
|
| 2263 |
+
"learning_rate": 4.955826419703735e-05,
|
| 2264 |
+
"loss": 0.3035,
|
| 2265 |
+
"step": 320
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 2.111842105263158,
|
| 2269 |
+
"grad_norm": 0.0933519192978327,
|
| 2270 |
+
"learning_rate": 4.889813205528895e-05,
|
| 2271 |
+
"loss": 0.3079,
|
| 2272 |
+
"step": 321
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 2.1184210526315788,
|
| 2276 |
+
"grad_norm": 0.08840630002939555,
|
| 2277 |
+
"learning_rate": 4.8241000227360354e-05,
|
| 2278 |
+
"loss": 0.3082,
|
| 2279 |
+
"step": 322
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 2.125,
|
| 2283 |
+
"grad_norm": 0.090468509851056,
|
| 2284 |
+
"learning_rate": 4.758690729504478e-05,
|
| 2285 |
+
"loss": 0.3013,
|
| 2286 |
+
"step": 323
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 2.1315789473684212,
|
| 2290 |
+
"grad_norm": 0.09269514982319728,
|
| 2291 |
+
"learning_rate": 4.693589166171466e-05,
|
| 2292 |
+
"loss": 0.3045,
|
| 2293 |
+
"step": 324
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 2.138157894736842,
|
| 2297 |
+
"grad_norm": 0.0944000183696063,
|
| 2298 |
+
"learning_rate": 4.628799155006669e-05,
|
| 2299 |
+
"loss": 0.3104,
|
| 2300 |
+
"step": 325
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 2.1447368421052633,
|
| 2304 |
+
"grad_norm": 0.09277066986876077,
|
| 2305 |
+
"learning_rate": 4.56432449998779e-05,
|
| 2306 |
+
"loss": 0.3061,
|
| 2307 |
+
"step": 326
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 2.151315789473684,
|
| 2311 |
+
"grad_norm": 0.09140609684482606,
|
| 2312 |
+
"learning_rate": 4.500168986577197e-05,
|
| 2313 |
+
"loss": 0.2987,
|
| 2314 |
+
"step": 327
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 2.1578947368421053,
|
| 2318 |
+
"grad_norm": 0.09135597016385583,
|
| 2319 |
+
"learning_rate": 4.4363363814997004e-05,
|
| 2320 |
+
"loss": 0.3138,
|
| 2321 |
+
"step": 328
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 2.1644736842105265,
|
| 2325 |
+
"grad_norm": 0.09228058446081117,
|
| 2326 |
+
"learning_rate": 4.3728304325213764e-05,
|
| 2327 |
+
"loss": 0.3093,
|
| 2328 |
+
"step": 329
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 2.1710526315789473,
|
| 2332 |
+
"grad_norm": 0.09444538193500554,
|
| 2333 |
+
"learning_rate": 4.30965486822953e-05,
|
| 2334 |
+
"loss": 0.3121,
|
| 2335 |
+
"step": 330
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 2.1776315789473686,
|
| 2339 |
+
"grad_norm": 0.09278379473856056,
|
| 2340 |
+
"learning_rate": 4.2468133978137945e-05,
|
| 2341 |
+
"loss": 0.3085,
|
| 2342 |
+
"step": 331
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 2.1842105263157894,
|
| 2346 |
+
"grad_norm": 0.09002068234283259,
|
| 2347 |
+
"learning_rate": 4.1843097108483465e-05,
|
| 2348 |
+
"loss": 0.3115,
|
| 2349 |
+
"step": 332
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 2.1907894736842106,
|
| 2353 |
+
"grad_norm": 0.09182734300425907,
|
| 2354 |
+
"learning_rate": 4.12214747707527e-05,
|
| 2355 |
+
"loss": 0.3041,
|
| 2356 |
+
"step": 333
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 2.1973684210526314,
|
| 2360 |
+
"grad_norm": 0.09480621705211878,
|
| 2361 |
+
"learning_rate": 4.0603303461891253e-05,
|
| 2362 |
+
"loss": 0.3119,
|
| 2363 |
+
"step": 334
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 2.2039473684210527,
|
| 2367 |
+
"grad_norm": 0.09462062961478308,
|
| 2368 |
+
"learning_rate": 3.9988619476226355e-05,
|
| 2369 |
+
"loss": 0.3097,
|
| 2370 |
+
"step": 335
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 2.2105263157894735,
|
| 2374 |
+
"grad_norm": 0.09047742114008103,
|
| 2375 |
+
"learning_rate": 3.937745890333623e-05,
|
| 2376 |
+
"loss": 0.2977,
|
| 2377 |
+
"step": 336
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 2.2171052631578947,
|
| 2381 |
+
"grad_norm": 0.09210517994431433,
|
| 2382 |
+
"learning_rate": 3.8769857625930906e-05,
|
| 2383 |
+
"loss": 0.3045,
|
| 2384 |
+
"step": 337
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 2.223684210526316,
|
| 2388 |
+
"grad_norm": 0.09051025798694169,
|
| 2389 |
+
"learning_rate": 3.816585131774571e-05,
|
| 2390 |
+
"loss": 0.3042,
|
| 2391 |
+
"step": 338
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 2.2302631578947367,
|
| 2395 |
+
"grad_norm": 0.0944091820926416,
|
| 2396 |
+
"learning_rate": 3.756547544144664e-05,
|
| 2397 |
+
"loss": 0.3146,
|
| 2398 |
+
"step": 339
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 2.236842105263158,
|
| 2402 |
+
"grad_norm": 0.09225272107547029,
|
| 2403 |
+
"learning_rate": 3.69687652465482e-05,
|
| 2404 |
+
"loss": 0.2979,
|
| 2405 |
+
"step": 340
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 2.2434210526315788,
|
| 2409 |
+
"grad_norm": 0.09466054564763007,
|
| 2410 |
+
"learning_rate": 3.637575576734404e-05,
|
| 2411 |
+
"loss": 0.3268,
|
| 2412 |
+
"step": 341
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 2.25,
|
| 2416 |
+
"grad_norm": 0.09415533738249948,
|
| 2417 |
+
"learning_rate": 3.578648182084975e-05,
|
| 2418 |
+
"loss": 0.3138,
|
| 2419 |
+
"step": 342
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 2.2565789473684212,
|
| 2423 |
+
"grad_norm": 0.09287732881109488,
|
| 2424 |
+
"learning_rate": 3.5200978004758886e-05,
|
| 2425 |
+
"loss": 0.3034,
|
| 2426 |
+
"step": 343
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 2.263157894736842,
|
| 2430 |
+
"grad_norm": 0.09052642673165286,
|
| 2431 |
+
"learning_rate": 3.46192786954115e-05,
|
| 2432 |
+
"loss": 0.3136,
|
| 2433 |
+
"step": 344
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 2.2697368421052633,
|
| 2437 |
+
"grad_norm": 0.0941549755498813,
|
| 2438 |
+
"learning_rate": 3.4041418045775895e-05,
|
| 2439 |
+
"loss": 0.3102,
|
| 2440 |
+
"step": 345
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 2.276315789473684,
|
| 2444 |
+
"grad_norm": 0.09081461979766672,
|
| 2445 |
+
"learning_rate": 3.346742998344348e-05,
|
| 2446 |
+
"loss": 0.3156,
|
| 2447 |
+
"step": 346
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 2.2828947368421053,
|
| 2451 |
+
"grad_norm": 0.09188038212201086,
|
| 2452 |
+
"learning_rate": 3.2897348208636614e-05,
|
| 2453 |
+
"loss": 0.2981,
|
| 2454 |
+
"step": 347
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 2.2894736842105265,
|
| 2458 |
+
"grad_norm": 0.09139958010871077,
|
| 2459 |
+
"learning_rate": 3.2331206192230214e-05,
|
| 2460 |
+
"loss": 0.3016,
|
| 2461 |
+
"step": 348
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 2.2960526315789473,
|
| 2465 |
+
"grad_norm": 0.09340863611831572,
|
| 2466 |
+
"learning_rate": 3.1769037173786375e-05,
|
| 2467 |
+
"loss": 0.2942,
|
| 2468 |
+
"step": 349
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 2.3026315789473686,
|
| 2472 |
+
"grad_norm": 0.09731529194622152,
|
| 2473 |
+
"learning_rate": 3.121087415960304e-05,
|
| 2474 |
+
"loss": 0.32,
|
| 2475 |
+
"step": 350
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 2.3092105263157894,
|
| 2479 |
+
"grad_norm": 0.09400113350744722,
|
| 2480 |
+
"learning_rate": 3.065674992077584e-05,
|
| 2481 |
+
"loss": 0.3121,
|
| 2482 |
+
"step": 351
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 2.3157894736842106,
|
| 2486 |
+
"grad_norm": 0.09370429535454715,
|
| 2487 |
+
"learning_rate": 3.0106696991274276e-05,
|
| 2488 |
+
"loss": 0.2998,
|
| 2489 |
+
"step": 352
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 2.3223684210526314,
|
| 2493 |
+
"grad_norm": 0.0931264318395488,
|
| 2494 |
+
"learning_rate": 2.9560747666031475e-05,
|
| 2495 |
+
"loss": 0.3045,
|
| 2496 |
+
"step": 353
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 2.3289473684210527,
|
| 2500 |
+
"grad_norm": 0.09457813125113766,
|
| 2501 |
+
"learning_rate": 2.901893399904797e-05,
|
| 2502 |
+
"loss": 0.3102,
|
| 2503 |
+
"step": 354
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 2.3355263157894735,
|
| 2507 |
+
"grad_norm": 0.09787356394017462,
|
| 2508 |
+
"learning_rate": 2.848128780150995e-05,
|
| 2509 |
+
"loss": 0.3124,
|
| 2510 |
+
"step": 355
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 2.3421052631578947,
|
| 2514 |
+
"grad_norm": 0.09294514140726695,
|
| 2515 |
+
"learning_rate": 2.794784063992131e-05,
|
| 2516 |
+
"loss": 0.3029,
|
| 2517 |
+
"step": 356
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 2.348684210526316,
|
| 2521 |
+
"grad_norm": 0.09472220447378009,
|
| 2522 |
+
"learning_rate": 2.74186238342505e-05,
|
| 2523 |
+
"loss": 0.3026,
|
| 2524 |
+
"step": 357
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 2.3552631578947367,
|
| 2528 |
+
"grad_norm": 0.09382282625627145,
|
| 2529 |
+
"learning_rate": 2.6893668456091625e-05,
|
| 2530 |
+
"loss": 0.3067,
|
| 2531 |
+
"step": 358
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 2.361842105263158,
|
| 2535 |
+
"grad_norm": 0.09371704952441416,
|
| 2536 |
+
"learning_rate": 2.6373005326839972e-05,
|
| 2537 |
+
"loss": 0.3065,
|
| 2538 |
+
"step": 359
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 2.3684210526315788,
|
| 2542 |
+
"grad_norm": 0.098774530422876,
|
| 2543 |
+
"learning_rate": 2.5856665015882685e-05,
|
| 2544 |
+
"loss": 0.3098,
|
| 2545 |
+
"step": 360
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 2.375,
|
| 2549 |
+
"grad_norm": 0.09126307975009218,
|
| 2550 |
+
"learning_rate": 2.5344677838803733e-05,
|
| 2551 |
+
"loss": 0.3044,
|
| 2552 |
+
"step": 361
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 2.3815789473684212,
|
| 2556 |
+
"grad_norm": 0.09144641630016763,
|
| 2557 |
+
"learning_rate": 2.4837073855604188e-05,
|
| 2558 |
+
"loss": 0.2908,
|
| 2559 |
+
"step": 362
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 2.388157894736842,
|
| 2563 |
+
"grad_norm": 0.09468863277888301,
|
| 2564 |
+
"learning_rate": 2.433388286893714e-05,
|
| 2565 |
+
"loss": 0.3101,
|
| 2566 |
+
"step": 363
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 2.3947368421052633,
|
| 2570 |
+
"grad_norm": 0.09354400587822048,
|
| 2571 |
+
"learning_rate": 2.383513442235812e-05,
|
| 2572 |
+
"loss": 0.3136,
|
| 2573 |
+
"step": 364
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 2.401315789473684,
|
| 2577 |
+
"grad_norm": 0.09390078122691951,
|
| 2578 |
+
"learning_rate": 2.334085779859041e-05,
|
| 2579 |
+
"loss": 0.3079,
|
| 2580 |
+
"step": 365
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 2.4078947368421053,
|
| 2584 |
+
"grad_norm": 0.09248111646485846,
|
| 2585 |
+
"learning_rate": 2.2851082017805703e-05,
|
| 2586 |
+
"loss": 0.305,
|
| 2587 |
+
"step": 366
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 2.4144736842105265,
|
| 2591 |
+
"grad_norm": 0.09244167723770774,
|
| 2592 |
+
"learning_rate": 2.236583583592049e-05,
|
| 2593 |
+
"loss": 0.302,
|
| 2594 |
+
"step": 367
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 2.4210526315789473,
|
| 2598 |
+
"grad_norm": 0.09129005795575806,
|
| 2599 |
+
"learning_rate": 2.1885147742907465e-05,
|
| 2600 |
+
"loss": 0.3027,
|
| 2601 |
+
"step": 368
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 2.4276315789473686,
|
| 2605 |
+
"grad_norm": 0.09432198973734413,
|
| 2606 |
+
"learning_rate": 2.1409045961123065e-05,
|
| 2607 |
+
"loss": 0.3217,
|
| 2608 |
+
"step": 369
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 2.4342105263157894,
|
| 2612 |
+
"grad_norm": 0.09309602245093605,
|
| 2613 |
+
"learning_rate": 2.0937558443650206e-05,
|
| 2614 |
+
"loss": 0.3063,
|
| 2615 |
+
"step": 370
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 2.4407894736842106,
|
| 2619 |
+
"grad_norm": 0.09208894600051501,
|
| 2620 |
+
"learning_rate": 2.047071287265735e-05,
|
| 2621 |
+
"loss": 0.3052,
|
| 2622 |
+
"step": 371
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 2.4473684210526314,
|
| 2626 |
+
"grad_norm": 0.09266552082076217,
|
| 2627 |
+
"learning_rate": 2.0008536657773046e-05,
|
| 2628 |
+
"loss": 0.3089,
|
| 2629 |
+
"step": 372
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 2.4539473684210527,
|
| 2633 |
+
"grad_norm": 0.09140806866621644,
|
| 2634 |
+
"learning_rate": 1.9551056934476654e-05,
|
| 2635 |
+
"loss": 0.3013,
|
| 2636 |
+
"step": 373
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 2.4605263157894735,
|
| 2640 |
+
"grad_norm": 0.09231696012858444,
|
| 2641 |
+
"learning_rate": 1.9098300562505266e-05,
|
| 2642 |
+
"loss": 0.3086,
|
| 2643 |
+
"step": 374
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 2.4671052631578947,
|
| 2647 |
+
"grad_norm": 0.08976907732352662,
|
| 2648 |
+
"learning_rate": 1.8650294124276556e-05,
|
| 2649 |
+
"loss": 0.2908,
|
| 2650 |
+
"step": 375
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 2.473684210526316,
|
| 2654 |
+
"grad_norm": 0.09089868074014618,
|
| 2655 |
+
"learning_rate": 1.8207063923328237e-05,
|
| 2656 |
+
"loss": 0.2934,
|
| 2657 |
+
"step": 376
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 2.4802631578947367,
|
| 2661 |
+
"grad_norm": 0.09124223709758555,
|
| 2662 |
+
"learning_rate": 1.7768635982773528e-05,
|
| 2663 |
+
"loss": 0.3025,
|
| 2664 |
+
"step": 377
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 2.486842105263158,
|
| 2668 |
+
"grad_norm": 0.09141228045415595,
|
| 2669 |
+
"learning_rate": 1.73350360437734e-05,
|
| 2670 |
+
"loss": 0.2984,
|
| 2671 |
+
"step": 378
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 2.4934210526315788,
|
| 2675 |
+
"grad_norm": 0.09338433503873096,
|
| 2676 |
+
"learning_rate": 1.6906289564025278e-05,
|
| 2677 |
+
"loss": 0.3015,
|
| 2678 |
+
"step": 379
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 2.5,
|
| 2682 |
+
"grad_norm": 0.09381766067303669,
|
| 2683 |
+
"learning_rate": 1.6482421716268215e-05,
|
| 2684 |
+
"loss": 0.3034,
|
| 2685 |
+
"step": 380
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 2.5065789473684212,
|
| 2689 |
+
"grad_norm": 0.09841665883274005,
|
| 2690 |
+
"learning_rate": 1.6063457386805004e-05,
|
| 2691 |
+
"loss": 0.3097,
|
| 2692 |
+
"step": 381
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 2.513157894736842,
|
| 2696 |
+
"grad_norm": 0.09562609257672962,
|
| 2697 |
+
"learning_rate": 1.564942117404119e-05,
|
| 2698 |
+
"loss": 0.3194,
|
| 2699 |
+
"step": 382
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 2.5197368421052633,
|
| 2703 |
+
"grad_norm": 0.09236226141715091,
|
| 2704 |
+
"learning_rate": 1.524033738704056e-05,
|
| 2705 |
+
"loss": 0.3116,
|
| 2706 |
+
"step": 383
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 2.526315789473684,
|
| 2710 |
+
"grad_norm": 0.09295111594898633,
|
| 2711 |
+
"learning_rate": 1.4836230044098166e-05,
|
| 2712 |
+
"loss": 0.2926,
|
| 2713 |
+
"step": 384
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 2.5328947368421053,
|
| 2717 |
+
"grad_norm": 0.09050178834061187,
|
| 2718 |
+
"learning_rate": 1.4437122871329955e-05,
|
| 2719 |
+
"loss": 0.3048,
|
| 2720 |
+
"step": 385
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 2.5394736842105265,
|
| 2724 |
+
"grad_norm": 0.09354446958969953,
|
| 2725 |
+
"learning_rate": 1.4043039301279903e-05,
|
| 2726 |
+
"loss": 0.3094,
|
| 2727 |
+
"step": 386
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 2.5460526315789473,
|
| 2731 |
+
"grad_norm": 0.094528346545483,
|
| 2732 |
+
"learning_rate": 1.3654002471544091e-05,
|
| 2733 |
+
"loss": 0.2976,
|
| 2734 |
+
"step": 387
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 2.5526315789473686,
|
| 2738 |
+
"grad_norm": 0.09437075440256357,
|
| 2739 |
+
"learning_rate": 1.3270035223412369e-05,
|
| 2740 |
+
"loss": 0.303,
|
| 2741 |
+
"step": 388
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 2.5592105263157894,
|
| 2745 |
+
"grad_norm": 0.09299672016467998,
|
| 2746 |
+
"learning_rate": 1.2891160100527222e-05,
|
| 2747 |
+
"loss": 0.2991,
|
| 2748 |
+
"step": 389
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 2.5657894736842106,
|
| 2752 |
+
"grad_norm": 0.09295508003371206,
|
| 2753 |
+
"learning_rate": 1.2517399347560167e-05,
|
| 2754 |
+
"loss": 0.3111,
|
| 2755 |
+
"step": 390
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 2.5723684210526314,
|
| 2759 |
+
"grad_norm": 0.09416195404382027,
|
| 2760 |
+
"learning_rate": 1.214877490890578e-05,
|
| 2761 |
+
"loss": 0.3175,
|
| 2762 |
+
"step": 391
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 2.5789473684210527,
|
| 2766 |
+
"grad_norm": 0.09276583875255047,
|
| 2767 |
+
"learning_rate": 1.1785308427393183e-05,
|
| 2768 |
+
"loss": 0.299,
|
| 2769 |
+
"step": 392
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 2.5855263157894735,
|
| 2773 |
+
"grad_norm": 0.09118856826069724,
|
| 2774 |
+
"learning_rate": 1.1427021243015467e-05,
|
| 2775 |
+
"loss": 0.3029,
|
| 2776 |
+
"step": 393
|
| 2777 |
+
},
|
| 2778 |
+
{
|
| 2779 |
+
"epoch": 2.5921052631578947,
|
| 2780 |
+
"grad_norm": 0.09153124872331597,
|
| 2781 |
+
"learning_rate": 1.1073934391676667e-05,
|
| 2782 |
+
"loss": 0.3053,
|
| 2783 |
+
"step": 394
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 2.598684210526316,
|
| 2787 |
+
"grad_norm": 0.0931741744031297,
|
| 2788 |
+
"learning_rate": 1.0726068603956741e-05,
|
| 2789 |
+
"loss": 0.2962,
|
| 2790 |
+
"step": 395
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 2.6052631578947367,
|
| 2794 |
+
"grad_norm": 0.0939187765998417,
|
| 2795 |
+
"learning_rate": 1.0383444303894452e-05,
|
| 2796 |
+
"loss": 0.3008,
|
| 2797 |
+
"step": 396
|
| 2798 |
+
},
|
| 2799 |
+
{
|
| 2800 |
+
"epoch": 2.611842105263158,
|
| 2801 |
+
"grad_norm": 0.09383471721688587,
|
| 2802 |
+
"learning_rate": 1.0046081607788127e-05,
|
| 2803 |
+
"loss": 0.312,
|
| 2804 |
+
"step": 397
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 2.6184210526315788,
|
| 2808 |
+
"grad_norm": 0.09081224477516317,
|
| 2809 |
+
"learning_rate": 9.714000323014705e-06,
|
| 2810 |
+
"loss": 0.2947,
|
| 2811 |
+
"step": 398
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 2.625,
|
| 2815 |
+
"grad_norm": 0.09356165649233147,
|
| 2816 |
+
"learning_rate": 9.3872199468667e-06,
|
| 2817 |
+
"loss": 0.3016,
|
| 2818 |
+
"step": 399
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 2.6315789473684212,
|
| 2822 |
+
"grad_norm": 0.09192904823893439,
|
| 2823 |
+
"learning_rate": 9.065759665407515e-06,
|
| 2824 |
+
"loss": 0.2971,
|
| 2825 |
+
"step": 400
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 2.638157894736842,
|
| 2829 |
+
"grad_norm": 0.09313138523215327,
|
| 2830 |
+
"learning_rate": 8.749638352345002e-06,
|
| 2831 |
+
"loss": 0.3042,
|
| 2832 |
+
"step": 401
|
| 2833 |
+
},
|
| 2834 |
+
{
|
| 2835 |
+
"epoch": 2.6447368421052633,
|
| 2836 |
+
"grad_norm": 0.0921977880339759,
|
| 2837 |
+
"learning_rate": 8.43887456792326e-06,
|
| 2838 |
+
"loss": 0.3016,
|
| 2839 |
+
"step": 402
|
| 2840 |
+
},
|
| 2841 |
+
{
|
| 2842 |
+
"epoch": 2.651315789473684,
|
| 2843 |
+
"grad_norm": 0.09055704810960578,
|
| 2844 |
+
"learning_rate": 8.133486557833015e-06,
|
| 2845 |
+
"loss": 0.3003,
|
| 2846 |
+
"step": 403
|
| 2847 |
+
},
|
| 2848 |
+
{
|
| 2849 |
+
"epoch": 2.6578947368421053,
|
| 2850 |
+
"grad_norm": 0.08952929321885197,
|
| 2851 |
+
"learning_rate": 7.833492252140284e-06,
|
| 2852 |
+
"loss": 0.2912,
|
| 2853 |
+
"step": 404
|
| 2854 |
+
},
|
| 2855 |
+
{
|
| 2856 |
+
"epoch": 2.6644736842105265,
|
| 2857 |
+
"grad_norm": 0.0917144195389881,
|
| 2858 |
+
"learning_rate": 7.538909264233751e-06,
|
| 2859 |
+
"loss": 0.2908,
|
| 2860 |
+
"step": 405
|
| 2861 |
+
},
|
| 2862 |
+
{
|
| 2863 |
+
"epoch": 2.6710526315789473,
|
| 2864 |
+
"grad_norm": 0.09112348602865261,
|
| 2865 |
+
"learning_rate": 7.249754889790539e-06,
|
| 2866 |
+
"loss": 0.2765,
|
| 2867 |
+
"step": 406
|
| 2868 |
+
},
|
| 2869 |
+
{
|
| 2870 |
+
"epoch": 2.6776315789473686,
|
| 2871 |
+
"grad_norm": 0.09188227591595896,
|
| 2872 |
+
"learning_rate": 6.966046105760826e-06,
|
| 2873 |
+
"loss": 0.2998,
|
| 2874 |
+
"step": 407
|
| 2875 |
+
},
|
| 2876 |
+
{
|
| 2877 |
+
"epoch": 2.6842105263157894,
|
| 2878 |
+
"grad_norm": 0.09193344799591113,
|
| 2879 |
+
"learning_rate": 6.6877995693710785e-06,
|
| 2880 |
+
"loss": 0.3073,
|
| 2881 |
+
"step": 408
|
| 2882 |
+
},
|
| 2883 |
+
{
|
| 2884 |
+
"epoch": 2.6907894736842106,
|
| 2885 |
+
"grad_norm": 0.09125510240984186,
|
| 2886 |
+
"learning_rate": 6.41503161714595e-06,
|
| 2887 |
+
"loss": 0.2856,
|
| 2888 |
+
"step": 409
|
| 2889 |
+
},
|
| 2890 |
+
{
|
| 2891 |
+
"epoch": 2.6973684210526314,
|
| 2892 |
+
"grad_norm": 0.09093722035003914,
|
| 2893 |
+
"learning_rate": 6.147758263949321e-06,
|
| 2894 |
+
"loss": 0.3029,
|
| 2895 |
+
"step": 410
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 2.7039473684210527,
|
| 2899 |
+
"grad_norm": 0.09720543266247605,
|
| 2900 |
+
"learning_rate": 5.885995202043848e-06,
|
| 2901 |
+
"loss": 0.3069,
|
| 2902 |
+
"step": 411
|
| 2903 |
+
},
|
| 2904 |
+
{
|
| 2905 |
+
"epoch": 2.7105263157894735,
|
| 2906 |
+
"grad_norm": 0.09427472370082618,
|
| 2907 |
+
"learning_rate": 5.629757800169733e-06,
|
| 2908 |
+
"loss": 0.3084,
|
| 2909 |
+
"step": 412
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 2.7171052631578947,
|
| 2913 |
+
"grad_norm": 0.089990452112005,
|
| 2914 |
+
"learning_rate": 5.379061102642358e-06,
|
| 2915 |
+
"loss": 0.2999,
|
| 2916 |
+
"step": 413
|
| 2917 |
+
},
|
| 2918 |
+
{
|
| 2919 |
+
"epoch": 2.723684210526316,
|
| 2920 |
+
"grad_norm": 0.09500774461582016,
|
| 2921 |
+
"learning_rate": 5.1339198284689916e-06,
|
| 2922 |
+
"loss": 0.3027,
|
| 2923 |
+
"step": 414
|
| 2924 |
+
},
|
| 2925 |
+
{
|
| 2926 |
+
"epoch": 2.7302631578947367,
|
| 2927 |
+
"grad_norm": 0.09140486608400628,
|
| 2928 |
+
"learning_rate": 4.8943483704846475e-06,
|
| 2929 |
+
"loss": 0.2943,
|
| 2930 |
+
"step": 415
|
| 2931 |
+
},
|
| 2932 |
+
{
|
| 2933 |
+
"epoch": 2.736842105263158,
|
| 2934 |
+
"grad_norm": 0.09254297353923395,
|
| 2935 |
+
"learning_rate": 4.660360794506946e-06,
|
| 2936 |
+
"loss": 0.2929,
|
| 2937 |
+
"step": 416
|
| 2938 |
+
},
|
| 2939 |
+
{
|
| 2940 |
+
"epoch": 2.7434210526315788,
|
| 2941 |
+
"grad_norm": 0.09259358584474216,
|
| 2942 |
+
"learning_rate": 4.431970838510391e-06,
|
| 2943 |
+
"loss": 0.3057,
|
| 2944 |
+
"step": 417
|
| 2945 |
+
},
|
| 2946 |
+
{
|
| 2947 |
+
"epoch": 2.75,
|
| 2948 |
+
"grad_norm": 0.0933006035630196,
|
| 2949 |
+
"learning_rate": 4.209191911819688e-06,
|
| 2950 |
+
"loss": 0.3066,
|
| 2951 |
+
"step": 418
|
| 2952 |
+
},
|
| 2953 |
+
{
|
| 2954 |
+
"epoch": 2.7565789473684212,
|
| 2955 |
+
"grad_norm": 0.09175027584175248,
|
| 2956 |
+
"learning_rate": 3.992037094322532e-06,
|
| 2957 |
+
"loss": 0.2974,
|
| 2958 |
+
"step": 419
|
| 2959 |
+
},
|
| 2960 |
+
{
|
| 2961 |
+
"epoch": 2.763157894736842,
|
| 2962 |
+
"grad_norm": 0.09612639615242942,
|
| 2963 |
+
"learning_rate": 3.7805191357015657e-06,
|
| 2964 |
+
"loss": 0.3152,
|
| 2965 |
+
"step": 420
|
| 2966 |
+
},
|
| 2967 |
+
{
|
| 2968 |
+
"epoch": 2.7697368421052633,
|
| 2969 |
+
"grad_norm": 0.09267277826656926,
|
| 2970 |
+
"learning_rate": 3.5746504546859018e-06,
|
| 2971 |
+
"loss": 0.3122,
|
| 2972 |
+
"step": 421
|
| 2973 |
+
},
|
| 2974 |
+
{
|
| 2975 |
+
"epoch": 2.776315789473684,
|
| 2976 |
+
"grad_norm": 0.09173196582903508,
|
| 2977 |
+
"learning_rate": 3.3744431383219367e-06,
|
| 2978 |
+
"loss": 0.3068,
|
| 2979 |
+
"step": 422
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 2.7828947368421053,
|
| 2983 |
+
"grad_norm": 0.09141455521971137,
|
| 2984 |
+
"learning_rate": 3.1799089412636762e-06,
|
| 2985 |
+
"loss": 0.3,
|
| 2986 |
+
"step": 423
|
| 2987 |
+
},
|
| 2988 |
+
{
|
| 2989 |
+
"epoch": 2.7894736842105265,
|
| 2990 |
+
"grad_norm": 0.09007921765376342,
|
| 2991 |
+
"learning_rate": 2.991059285082698e-06,
|
| 2992 |
+
"loss": 0.3009,
|
| 2993 |
+
"step": 424
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 2.7960526315789473,
|
| 2997 |
+
"grad_norm": 0.0953184723452783,
|
| 2998 |
+
"learning_rate": 2.8079052575973765e-06,
|
| 2999 |
+
"loss": 0.3043,
|
| 3000 |
+
"step": 425
|
| 3001 |
+
},
|
| 3002 |
+
{
|
| 3003 |
+
"epoch": 2.8026315789473686,
|
| 3004 |
+
"grad_norm": 0.09157757129707106,
|
| 3005 |
+
"learning_rate": 2.6304576122221035e-06,
|
| 3006 |
+
"loss": 0.3,
|
| 3007 |
+
"step": 426
|
| 3008 |
+
},
|
| 3009 |
+
{
|
| 3010 |
+
"epoch": 2.8092105263157894,
|
| 3011 |
+
"grad_norm": 0.09169344011676386,
|
| 3012 |
+
"learning_rate": 2.4587267673357394e-06,
|
| 3013 |
+
"loss": 0.2917,
|
| 3014 |
+
"step": 427
|
| 3015 |
+
},
|
| 3016 |
+
{
|
| 3017 |
+
"epoch": 2.8157894736842106,
|
| 3018 |
+
"grad_norm": 0.09140035394059028,
|
| 3019 |
+
"learning_rate": 2.2927228056700777e-06,
|
| 3020 |
+
"loss": 0.3086,
|
| 3021 |
+
"step": 428
|
| 3022 |
+
},
|
| 3023 |
+
{
|
| 3024 |
+
"epoch": 2.8223684210526314,
|
| 3025 |
+
"grad_norm": 0.08884580894240417,
|
| 3026 |
+
"learning_rate": 2.1324554737177648e-06,
|
| 3027 |
+
"loss": 0.2974,
|
| 3028 |
+
"step": 429
|
| 3029 |
+
},
|
| 3030 |
+
{
|
| 3031 |
+
"epoch": 2.8289473684210527,
|
| 3032 |
+
"grad_norm": 0.09227882207821884,
|
| 3033 |
+
"learning_rate": 1.9779341811600794e-06,
|
| 3034 |
+
"loss": 0.2993,
|
| 3035 |
+
"step": 430
|
| 3036 |
+
},
|
| 3037 |
+
{
|
| 3038 |
+
"epoch": 2.8355263157894735,
|
| 3039 |
+
"grad_norm": 0.09230636915969623,
|
| 3040 |
+
"learning_rate": 1.8291680003145073e-06,
|
| 3041 |
+
"loss": 0.3014,
|
| 3042 |
+
"step": 431
|
| 3043 |
+
},
|
| 3044 |
+
{
|
| 3045 |
+
"epoch": 2.8421052631578947,
|
| 3046 |
+
"grad_norm": 0.09126820223778219,
|
| 3047 |
+
"learning_rate": 1.6861656656020464e-06,
|
| 3048 |
+
"loss": 0.301,
|
| 3049 |
+
"step": 432
|
| 3050 |
+
},
|
| 3051 |
+
{
|
| 3052 |
+
"epoch": 2.848684210526316,
|
| 3053 |
+
"grad_norm": 0.09346572537156347,
|
| 3054 |
+
"learning_rate": 1.5489355730343934e-06,
|
| 3055 |
+
"loss": 0.3047,
|
| 3056 |
+
"step": 433
|
| 3057 |
+
},
|
| 3058 |
+
{
|
| 3059 |
+
"epoch": 2.8552631578947367,
|
| 3060 |
+
"grad_norm": 0.09159445394674477,
|
| 3061 |
+
"learning_rate": 1.417485779720995e-06,
|
| 3062 |
+
"loss": 0.3111,
|
| 3063 |
+
"step": 434
|
| 3064 |
+
},
|
| 3065 |
+
{
|
| 3066 |
+
"epoch": 2.861842105263158,
|
| 3067 |
+
"grad_norm": 0.09263259410862734,
|
| 3068 |
+
"learning_rate": 1.2918240033960027e-06,
|
| 3069 |
+
"loss": 0.3064,
|
| 3070 |
+
"step": 435
|
| 3071 |
+
},
|
| 3072 |
+
{
|
| 3073 |
+
"epoch": 2.8684210526315788,
|
| 3074 |
+
"grad_norm": 0.09195985713654573,
|
| 3075 |
+
"learning_rate": 1.1719576219651585e-06,
|
| 3076 |
+
"loss": 0.3085,
|
| 3077 |
+
"step": 436
|
| 3078 |
+
},
|
| 3079 |
+
{
|
| 3080 |
+
"epoch": 2.875,
|
| 3081 |
+
"grad_norm": 0.09287155646470045,
|
| 3082 |
+
"learning_rate": 1.057893673072563e-06,
|
| 3083 |
+
"loss": 0.3123,
|
| 3084 |
+
"step": 437
|
| 3085 |
+
},
|
| 3086 |
+
{
|
| 3087 |
+
"epoch": 2.8815789473684212,
|
| 3088 |
+
"grad_norm": 0.09261252125029774,
|
| 3089 |
+
"learning_rate": 9.496388536875623e-07,
|
| 3090 |
+
"loss": 0.3014,
|
| 3091 |
+
"step": 438
|
| 3092 |
+
},
|
| 3093 |
+
{
|
| 3094 |
+
"epoch": 2.888157894736842,
|
| 3095 |
+
"grad_norm": 0.09510242096519511,
|
| 3096 |
+
"learning_rate": 8.471995197114835e-07,
|
| 3097 |
+
"loss": 0.3117,
|
| 3098 |
+
"step": 439
|
| 3099 |
+
},
|
| 3100 |
+
{
|
| 3101 |
+
"epoch": 2.8947368421052633,
|
| 3102 |
+
"grad_norm": 0.09143211071538708,
|
| 3103 |
+
"learning_rate": 7.505816856045012e-07,
|
| 3104 |
+
"loss": 0.3021,
|
| 3105 |
+
"step": 440
|
| 3106 |
+
},
|
| 3107 |
+
{
|
| 3108 |
+
"epoch": 2.901315789473684,
|
| 3109 |
+
"grad_norm": 0.0920917445117736,
|
| 3110 |
+
"learning_rate": 6.597910240324967e-07,
|
| 3111 |
+
"loss": 0.3049,
|
| 3112 |
+
"step": 441
|
| 3113 |
+
},
|
| 3114 |
+
{
|
| 3115 |
+
"epoch": 2.9078947368421053,
|
| 3116 |
+
"grad_norm": 0.09323526996440369,
|
| 3117 |
+
"learning_rate": 5.74832865534014e-07,
|
| 3118 |
+
"loss": 0.3044,
|
| 3119 |
+
"step": 442
|
| 3120 |
+
},
|
| 3121 |
+
{
|
| 3122 |
+
"epoch": 2.9144736842105265,
|
| 3123 |
+
"grad_norm": 0.08995868186598065,
|
| 3124 |
+
"learning_rate": 4.957121982072765e-07,
|
| 3125 |
+
"loss": 0.3074,
|
| 3126 |
+
"step": 443
|
| 3127 |
+
},
|
| 3128 |
+
{
|
| 3129 |
+
"epoch": 2.9210526315789473,
|
| 3130 |
+
"grad_norm": 0.09113776795368703,
|
| 3131 |
+
"learning_rate": 4.224336674173546e-07,
|
| 3132 |
+
"loss": 0.3109,
|
| 3133 |
+
"step": 444
|
| 3134 |
+
},
|
| 3135 |
+
{
|
| 3136 |
+
"epoch": 2.9276315789473686,
|
| 3137 |
+
"grad_norm": 0.09224937503409146,
|
| 3138 |
+
"learning_rate": 3.5500157552337264e-07,
|
| 3139 |
+
"loss": 0.3017,
|
| 3140 |
+
"step": 445
|
| 3141 |
+
},
|
| 3142 |
+
{
|
| 3143 |
+
"epoch": 2.9342105263157894,
|
| 3144 |
+
"grad_norm": 0.09468784220174042,
|
| 3145 |
+
"learning_rate": 2.934198816259559e-07,
|
| 3146 |
+
"loss": 0.2992,
|
| 3147 |
+
"step": 446
|
| 3148 |
+
},
|
| 3149 |
+
{
|
| 3150 |
+
"epoch": 2.9407894736842106,
|
| 3151 |
+
"grad_norm": 0.09123611642303985,
|
| 3152 |
+
"learning_rate": 2.3769220133477156e-07,
|
| 3153 |
+
"loss": 0.3068,
|
| 3154 |
+
"step": 447
|
| 3155 |
+
},
|
| 3156 |
+
{
|
| 3157 |
+
"epoch": 2.9473684210526314,
|
| 3158 |
+
"grad_norm": 0.09217795191626804,
|
| 3159 |
+
"learning_rate": 1.8782180655622095e-07,
|
| 3160 |
+
"loss": 0.3032,
|
| 3161 |
+
"step": 448
|
| 3162 |
+
},
|
| 3163 |
+
{
|
| 3164 |
+
"epoch": 2.9539473684210527,
|
| 3165 |
+
"grad_norm": 0.09439928587738844,
|
| 3166 |
+
"learning_rate": 1.4381162530135993e-07,
|
| 3167 |
+
"loss": 0.3149,
|
| 3168 |
+
"step": 449
|
| 3169 |
+
},
|
| 3170 |
+
{
|
| 3171 |
+
"epoch": 2.9605263157894735,
|
| 3172 |
+
"grad_norm": 0.09155507167334377,
|
| 3173 |
+
"learning_rate": 1.0566424151401411e-07,
|
| 3174 |
+
"loss": 0.3064,
|
| 3175 |
+
"step": 450
|
| 3176 |
+
},
|
| 3177 |
+
{
|
| 3178 |
+
"epoch": 2.9671052631578947,
|
| 3179 |
+
"grad_norm": 0.09353543798817081,
|
| 3180 |
+
"learning_rate": 7.338189491900016e-08,
|
| 3181 |
+
"loss": 0.3079,
|
| 3182 |
+
"step": 451
|
| 3183 |
+
},
|
| 3184 |
+
{
|
| 3185 |
+
"epoch": 2.973684210526316,
|
| 3186 |
+
"grad_norm": 0.09071512231087696,
|
| 3187 |
+
"learning_rate": 4.696648089068667e-08,
|
| 3188 |
+
"loss": 0.3019,
|
| 3189 |
+
"step": 452
|
| 3190 |
+
},
|
| 3191 |
+
{
|
| 3192 |
+
"epoch": 2.9802631578947367,
|
| 3193 |
+
"grad_norm": 0.09100478976490751,
|
| 3194 |
+
"learning_rate": 2.641955034170529e-08,
|
| 3195 |
+
"loss": 0.3038,
|
| 3196 |
+
"step": 453
|
| 3197 |
+
},
|
| 3198 |
+
{
|
| 3199 |
+
"epoch": 2.986842105263158,
|
| 3200 |
+
"grad_norm": 0.09180608030721753,
|
| 3201 |
+
"learning_rate": 1.1742309631845861e-08,
|
| 3202 |
+
"loss": 0.305,
|
| 3203 |
+
"step": 454
|
| 3204 |
+
},
|
| 3205 |
+
{
|
| 3206 |
+
"epoch": 2.9934210526315788,
|
| 3207 |
+
"grad_norm": 0.09166087142524931,
|
| 3208 |
+
"learning_rate": 2.935620497301894e-09,
|
| 3209 |
+
"loss": 0.2995,
|
| 3210 |
+
"step": 455
|
| 3211 |
+
},
|
| 3212 |
+
{
|
| 3213 |
+
"epoch": 3.0,
|
| 3214 |
+
"grad_norm": 0.08984730795754493,
|
| 3215 |
+
"learning_rate": 0.0,
|
| 3216 |
+
"loss": 0.2989,
|
| 3217 |
+
"step": 456
|
| 3218 |
+
},
|
| 3219 |
+
{
|
| 3220 |
+
"epoch": 3.0,
|
| 3221 |
+
"eval_loss": 0.3327427804470062,
|
| 3222 |
+
"eval_runtime": 158.4336,
|
| 3223 |
+
"eval_samples_per_second": 32.222,
|
| 3224 |
+
"eval_steps_per_second": 1.01,
|
| 3225 |
+
"step": 456
|
| 3226 |
+
},
|
| 3227 |
+
{
|
| 3228 |
+
"epoch": 3.0,
|
| 3229 |
+
"step": 456,
|
| 3230 |
+
"total_flos": 1.6172409563591475e+17,
|
| 3231 |
+
"train_loss": 0.366262159849468,
|
| 3232 |
+
"train_runtime": 5417.5508,
|
| 3233 |
+
"train_samples_per_second": 10.741,
|
| 3234 |
+
"train_steps_per_second": 0.084
|
| 3235 |
+
}
|
| 3236 |
+
],
|
| 3237 |
+
"logging_steps": 1,
|
| 3238 |
+
"max_steps": 456,
|
| 3239 |
+
"num_input_tokens_seen": 0,
|
| 3240 |
+
"num_train_epochs": 3,
|
| 3241 |
+
"save_steps": 500,
|
| 3242 |
+
"stateful_callbacks": {
|
| 3243 |
+
"TrainerControl": {
|
| 3244 |
+
"args": {
|
| 3245 |
+
"should_epoch_stop": false,
|
| 3246 |
+
"should_evaluate": false,
|
| 3247 |
+
"should_log": false,
|
| 3248 |
+
"should_save": true,
|
| 3249 |
+
"should_training_stop": true
|
| 3250 |
+
},
|
| 3251 |
+
"attributes": {}
|
| 3252 |
+
}
|
| 3253 |
+
},
|
| 3254 |
+
"total_flos": 1.6172409563591475e+17,
|
| 3255 |
+
"train_batch_size": 8,
|
| 3256 |
+
"trial_name": null,
|
| 3257 |
+
"trial_params": null
|
| 3258 |
+
}
|