Model save
Browse files
README.md
ADDED
|
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
library_name: peft
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- sft
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
base_model: mistralai/Mistral-7B-v0.1
|
| 9 |
+
datasets:
|
| 10 |
+
- generator
|
| 11 |
+
model-index:
|
| 12 |
+
- name: zephyr-7b-sft-qlora
|
| 13 |
+
results: []
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 17 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 18 |
+
|
| 19 |
+
# zephyr-7b-sft-qlora
|
| 20 |
+
|
| 21 |
+
This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the generator dataset.
|
| 22 |
+
It achieves the following results on the evaluation set:
|
| 23 |
+
- Loss: 0.9438
|
| 24 |
+
|
| 25 |
+
## Model description
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Intended uses & limitations
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training and evaluation data
|
| 34 |
+
|
| 35 |
+
More information needed
|
| 36 |
+
|
| 37 |
+
## Training procedure
|
| 38 |
+
|
| 39 |
+
### Training hyperparameters
|
| 40 |
+
|
| 41 |
+
The following hyperparameters were used during training:
|
| 42 |
+
- learning_rate: 0.0002
|
| 43 |
+
- train_batch_size: 4
|
| 44 |
+
- eval_batch_size: 8
|
| 45 |
+
- seed: 42
|
| 46 |
+
- distributed_type: multi-GPU
|
| 47 |
+
- num_devices: 8
|
| 48 |
+
- gradient_accumulation_steps: 2
|
| 49 |
+
- total_train_batch_size: 64
|
| 50 |
+
- total_eval_batch_size: 64
|
| 51 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 52 |
+
- lr_scheduler_type: cosine
|
| 53 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 54 |
+
- num_epochs: 1
|
| 55 |
+
|
| 56 |
+
### Training results
|
| 57 |
+
|
| 58 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 59 |
+
|:-------------:|:-----:|:----:|:---------------:|
|
| 60 |
+
| 0.9452 | 1.0 | 2179 | 0.9438 |
|
| 61 |
+
|
| 62 |
+
|
| 63 |
+
### Framework versions
|
| 64 |
+
|
| 65 |
+
- PEFT 0.7.1
|
| 66 |
+
- Transformers 4.39.0.dev0
|
| 67 |
+
- Pytorch 2.2.2+cu121
|
| 68 |
+
- Datasets 2.14.6
|
| 69 |
+
- Tokenizers 0.15.2
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 83946192
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4f33af19f8496f7e6ee12809abd80ea89351814f105a8c6d723b399f17726287
|
| 3 |
size 83946192
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.951526911010322,
|
| 4 |
+
"train_runtime": 20629.7225,
|
| 5 |
+
"train_samples": 207864,
|
| 6 |
+
"train_samples_per_second": 6.758,
|
| 7 |
+
"train_steps_per_second": 0.106
|
| 8 |
+
}
|
runs/Apr12_14-58-15_autodl-container-879545b4da-32bd4804/events.out.tfevents.1712905170.autodl-container-879545b4da-32bd4804.38089.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a215db035d1ef3e6f30ff5f50bd8fc9c9c6d77be71d127548752556abc571c42
|
| 3 |
+
size 97215
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.951526911010322,
|
| 4 |
+
"train_runtime": 20629.7225,
|
| 5 |
+
"train_samples": 207864,
|
| 6 |
+
"train_samples_per_second": 6.758,
|
| 7 |
+
"train_steps_per_second": 0.106
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3090 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2179,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"grad_norm": 0.28782306759602416,
|
| 14 |
+
"learning_rate": 9.174311926605506e-07,
|
| 15 |
+
"loss": 1.1568,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0,
|
| 20 |
+
"grad_norm": 0.2765226643079522,
|
| 21 |
+
"learning_rate": 4.587155963302753e-06,
|
| 22 |
+
"loss": 1.1458,
|
| 23 |
+
"step": 5
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.0,
|
| 27 |
+
"grad_norm": 0.2707977544047907,
|
| 28 |
+
"learning_rate": 9.174311926605506e-06,
|
| 29 |
+
"loss": 1.1252,
|
| 30 |
+
"step": 10
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.01,
|
| 34 |
+
"grad_norm": 0.23184164475898333,
|
| 35 |
+
"learning_rate": 1.3761467889908258e-05,
|
| 36 |
+
"loss": 1.1309,
|
| 37 |
+
"step": 15
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.01,
|
| 41 |
+
"grad_norm": 0.207610763811708,
|
| 42 |
+
"learning_rate": 1.834862385321101e-05,
|
| 43 |
+
"loss": 1.1227,
|
| 44 |
+
"step": 20
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.01,
|
| 48 |
+
"grad_norm": 0.1816681264337633,
|
| 49 |
+
"learning_rate": 2.2935779816513765e-05,
|
| 50 |
+
"loss": 1.0656,
|
| 51 |
+
"step": 25
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.01,
|
| 55 |
+
"grad_norm": 0.1631492001334514,
|
| 56 |
+
"learning_rate": 2.7522935779816515e-05,
|
| 57 |
+
"loss": 1.0435,
|
| 58 |
+
"step": 30
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.02,
|
| 62 |
+
"grad_norm": 0.12091918318355398,
|
| 63 |
+
"learning_rate": 3.211009174311927e-05,
|
| 64 |
+
"loss": 1.0422,
|
| 65 |
+
"step": 35
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.02,
|
| 69 |
+
"grad_norm": 0.10894779898677374,
|
| 70 |
+
"learning_rate": 3.669724770642202e-05,
|
| 71 |
+
"loss": 1.0231,
|
| 72 |
+
"step": 40
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.02,
|
| 76 |
+
"grad_norm": 0.10540985367231842,
|
| 77 |
+
"learning_rate": 4.1284403669724776e-05,
|
| 78 |
+
"loss": 0.9966,
|
| 79 |
+
"step": 45
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.02,
|
| 83 |
+
"grad_norm": 0.09676232054101573,
|
| 84 |
+
"learning_rate": 4.587155963302753e-05,
|
| 85 |
+
"loss": 0.9798,
|
| 86 |
+
"step": 50
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.03,
|
| 90 |
+
"grad_norm": 0.11000697328127398,
|
| 91 |
+
"learning_rate": 5.0458715596330276e-05,
|
| 92 |
+
"loss": 1.0154,
|
| 93 |
+
"step": 55
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.03,
|
| 97 |
+
"grad_norm": 0.10373599911083506,
|
| 98 |
+
"learning_rate": 5.504587155963303e-05,
|
| 99 |
+
"loss": 0.9935,
|
| 100 |
+
"step": 60
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.03,
|
| 104 |
+
"grad_norm": 0.11059687218734753,
|
| 105 |
+
"learning_rate": 5.9633027522935784e-05,
|
| 106 |
+
"loss": 0.9899,
|
| 107 |
+
"step": 65
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.03,
|
| 111 |
+
"grad_norm": 0.11912414533134615,
|
| 112 |
+
"learning_rate": 6.422018348623854e-05,
|
| 113 |
+
"loss": 0.9856,
|
| 114 |
+
"step": 70
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.03,
|
| 118 |
+
"grad_norm": 0.1133131029182366,
|
| 119 |
+
"learning_rate": 6.880733944954129e-05,
|
| 120 |
+
"loss": 0.995,
|
| 121 |
+
"step": 75
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.04,
|
| 125 |
+
"grad_norm": 0.1285831542439206,
|
| 126 |
+
"learning_rate": 7.339449541284404e-05,
|
| 127 |
+
"loss": 0.9857,
|
| 128 |
+
"step": 80
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.04,
|
| 132 |
+
"grad_norm": 0.11413242308426012,
|
| 133 |
+
"learning_rate": 7.79816513761468e-05,
|
| 134 |
+
"loss": 0.9809,
|
| 135 |
+
"step": 85
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.04,
|
| 139 |
+
"grad_norm": 0.10876549432839086,
|
| 140 |
+
"learning_rate": 8.256880733944955e-05,
|
| 141 |
+
"loss": 0.9742,
|
| 142 |
+
"step": 90
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.04,
|
| 146 |
+
"grad_norm": 0.11916614500162052,
|
| 147 |
+
"learning_rate": 8.715596330275229e-05,
|
| 148 |
+
"loss": 0.9574,
|
| 149 |
+
"step": 95
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.05,
|
| 153 |
+
"grad_norm": 0.1370299381128718,
|
| 154 |
+
"learning_rate": 9.174311926605506e-05,
|
| 155 |
+
"loss": 0.9862,
|
| 156 |
+
"step": 100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.05,
|
| 160 |
+
"grad_norm": 0.137870839924706,
|
| 161 |
+
"learning_rate": 9.63302752293578e-05,
|
| 162 |
+
"loss": 1.0027,
|
| 163 |
+
"step": 105
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.05,
|
| 167 |
+
"grad_norm": 0.12566416676044165,
|
| 168 |
+
"learning_rate": 0.00010091743119266055,
|
| 169 |
+
"loss": 0.9572,
|
| 170 |
+
"step": 110
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.05,
|
| 174 |
+
"grad_norm": 0.11452270864650357,
|
| 175 |
+
"learning_rate": 0.00010550458715596329,
|
| 176 |
+
"loss": 0.9547,
|
| 177 |
+
"step": 115
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.06,
|
| 181 |
+
"grad_norm": 0.11788556058363543,
|
| 182 |
+
"learning_rate": 0.00011009174311926606,
|
| 183 |
+
"loss": 1.0131,
|
| 184 |
+
"step": 120
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.06,
|
| 188 |
+
"grad_norm": 0.11618026116590524,
|
| 189 |
+
"learning_rate": 0.00011467889908256881,
|
| 190 |
+
"loss": 0.9948,
|
| 191 |
+
"step": 125
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.06,
|
| 195 |
+
"grad_norm": 0.11293458291689443,
|
| 196 |
+
"learning_rate": 0.00011926605504587157,
|
| 197 |
+
"loss": 0.983,
|
| 198 |
+
"step": 130
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.06,
|
| 202 |
+
"grad_norm": 0.11157978831839803,
|
| 203 |
+
"learning_rate": 0.00012385321100917432,
|
| 204 |
+
"loss": 0.9708,
|
| 205 |
+
"step": 135
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.06,
|
| 209 |
+
"grad_norm": 0.1084963251618998,
|
| 210 |
+
"learning_rate": 0.00012844036697247707,
|
| 211 |
+
"loss": 0.9696,
|
| 212 |
+
"step": 140
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.07,
|
| 216 |
+
"grad_norm": 0.12208924093427911,
|
| 217 |
+
"learning_rate": 0.00013302752293577983,
|
| 218 |
+
"loss": 0.9589,
|
| 219 |
+
"step": 145
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.07,
|
| 223 |
+
"grad_norm": 0.12400327121666536,
|
| 224 |
+
"learning_rate": 0.00013761467889908258,
|
| 225 |
+
"loss": 0.9743,
|
| 226 |
+
"step": 150
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.07,
|
| 230 |
+
"grad_norm": 0.11683640502206159,
|
| 231 |
+
"learning_rate": 0.0001422018348623853,
|
| 232 |
+
"loss": 0.9561,
|
| 233 |
+
"step": 155
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.07,
|
| 237 |
+
"grad_norm": 0.10367694083396438,
|
| 238 |
+
"learning_rate": 0.0001467889908256881,
|
| 239 |
+
"loss": 0.9666,
|
| 240 |
+
"step": 160
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.08,
|
| 244 |
+
"grad_norm": 0.12707530495236583,
|
| 245 |
+
"learning_rate": 0.00015137614678899084,
|
| 246 |
+
"loss": 0.9865,
|
| 247 |
+
"step": 165
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.08,
|
| 251 |
+
"grad_norm": 0.10632460411245542,
|
| 252 |
+
"learning_rate": 0.0001559633027522936,
|
| 253 |
+
"loss": 0.9674,
|
| 254 |
+
"step": 170
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.08,
|
| 258 |
+
"grad_norm": 0.10377606105203888,
|
| 259 |
+
"learning_rate": 0.00016055045871559632,
|
| 260 |
+
"loss": 0.9868,
|
| 261 |
+
"step": 175
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.08,
|
| 265 |
+
"grad_norm": 0.10127959187939727,
|
| 266 |
+
"learning_rate": 0.0001651376146788991,
|
| 267 |
+
"loss": 0.964,
|
| 268 |
+
"step": 180
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.08,
|
| 272 |
+
"grad_norm": 0.1067977777725574,
|
| 273 |
+
"learning_rate": 0.00016972477064220186,
|
| 274 |
+
"loss": 0.9818,
|
| 275 |
+
"step": 185
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.09,
|
| 279 |
+
"grad_norm": 0.10097509908529916,
|
| 280 |
+
"learning_rate": 0.00017431192660550458,
|
| 281 |
+
"loss": 0.985,
|
| 282 |
+
"step": 190
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.09,
|
| 286 |
+
"grad_norm": 0.08983732625855309,
|
| 287 |
+
"learning_rate": 0.00017889908256880734,
|
| 288 |
+
"loss": 0.986,
|
| 289 |
+
"step": 195
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.09,
|
| 293 |
+
"grad_norm": 0.09879437209533858,
|
| 294 |
+
"learning_rate": 0.00018348623853211012,
|
| 295 |
+
"loss": 0.9575,
|
| 296 |
+
"step": 200
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.09,
|
| 300 |
+
"grad_norm": 0.09503268529031307,
|
| 301 |
+
"learning_rate": 0.00018807339449541284,
|
| 302 |
+
"loss": 0.9652,
|
| 303 |
+
"step": 205
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.1,
|
| 307 |
+
"grad_norm": 0.09280130396940776,
|
| 308 |
+
"learning_rate": 0.0001926605504587156,
|
| 309 |
+
"loss": 0.9546,
|
| 310 |
+
"step": 210
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.1,
|
| 314 |
+
"grad_norm": 0.09653028292096948,
|
| 315 |
+
"learning_rate": 0.00019724770642201835,
|
| 316 |
+
"loss": 0.9683,
|
| 317 |
+
"step": 215
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.1,
|
| 321 |
+
"grad_norm": 0.09433817156644378,
|
| 322 |
+
"learning_rate": 0.00019999948669655127,
|
| 323 |
+
"loss": 0.9658,
|
| 324 |
+
"step": 220
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.1,
|
| 328 |
+
"grad_norm": 0.09147287152333045,
|
| 329 |
+
"learning_rate": 0.00019999371209327089,
|
| 330 |
+
"loss": 0.9637,
|
| 331 |
+
"step": 225
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.11,
|
| 335 |
+
"grad_norm": 0.09512038122974721,
|
| 336 |
+
"learning_rate": 0.00019998152162914806,
|
| 337 |
+
"loss": 0.9459,
|
| 338 |
+
"step": 230
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.11,
|
| 342 |
+
"grad_norm": 0.0941352545868701,
|
| 343 |
+
"learning_rate": 0.00019996291608635525,
|
| 344 |
+
"loss": 0.9798,
|
| 345 |
+
"step": 235
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.11,
|
| 349 |
+
"grad_norm": 0.10048319682038799,
|
| 350 |
+
"learning_rate": 0.00019993789665867314,
|
| 351 |
+
"loss": 0.9477,
|
| 352 |
+
"step": 240
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.11,
|
| 356 |
+
"grad_norm": 0.0870804299235567,
|
| 357 |
+
"learning_rate": 0.00019990646495141444,
|
| 358 |
+
"loss": 0.9727,
|
| 359 |
+
"step": 245
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.11,
|
| 363 |
+
"grad_norm": 0.0911867604696636,
|
| 364 |
+
"learning_rate": 0.0001998686229813205,
|
| 365 |
+
"loss": 0.994,
|
| 366 |
+
"step": 250
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.12,
|
| 370 |
+
"grad_norm": 0.0910776617883196,
|
| 371 |
+
"learning_rate": 0.00019982437317643217,
|
| 372 |
+
"loss": 0.967,
|
| 373 |
+
"step": 255
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.12,
|
| 377 |
+
"grad_norm": 0.09364688409217624,
|
| 378 |
+
"learning_rate": 0.0001997737183759338,
|
| 379 |
+
"loss": 0.9521,
|
| 380 |
+
"step": 260
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.12,
|
| 384 |
+
"grad_norm": 0.09254538330109398,
|
| 385 |
+
"learning_rate": 0.00019971666182997137,
|
| 386 |
+
"loss": 0.986,
|
| 387 |
+
"step": 265
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.12,
|
| 391 |
+
"grad_norm": 0.10322422823108995,
|
| 392 |
+
"learning_rate": 0.00019965320719944366,
|
| 393 |
+
"loss": 0.9541,
|
| 394 |
+
"step": 270
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.13,
|
| 398 |
+
"grad_norm": 0.09173096996766032,
|
| 399 |
+
"learning_rate": 0.00019958335855576738,
|
| 400 |
+
"loss": 0.9546,
|
| 401 |
+
"step": 275
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.13,
|
| 405 |
+
"grad_norm": 0.0893743167592582,
|
| 406 |
+
"learning_rate": 0.00019950712038061617,
|
| 407 |
+
"loss": 0.9576,
|
| 408 |
+
"step": 280
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.13,
|
| 412 |
+
"grad_norm": 0.09477317623808881,
|
| 413 |
+
"learning_rate": 0.00019942449756563279,
|
| 414 |
+
"loss": 0.9677,
|
| 415 |
+
"step": 285
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.13,
|
| 419 |
+
"grad_norm": 0.0938118134559368,
|
| 420 |
+
"learning_rate": 0.00019933549541211552,
|
| 421 |
+
"loss": 0.9365,
|
| 422 |
+
"step": 290
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.14,
|
| 426 |
+
"grad_norm": 0.08783897990789609,
|
| 427 |
+
"learning_rate": 0.00019924011963067765,
|
| 428 |
+
"loss": 0.9683,
|
| 429 |
+
"step": 295
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.14,
|
| 433 |
+
"grad_norm": 0.09119730775090834,
|
| 434 |
+
"learning_rate": 0.00019913837634088144,
|
| 435 |
+
"loss": 0.9482,
|
| 436 |
+
"step": 300
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.14,
|
| 440 |
+
"grad_norm": 0.0923867205609476,
|
| 441 |
+
"learning_rate": 0.00019903027207084523,
|
| 442 |
+
"loss": 0.9352,
|
| 443 |
+
"step": 305
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.14,
|
| 447 |
+
"grad_norm": 0.08608823546934513,
|
| 448 |
+
"learning_rate": 0.00019891581375682473,
|
| 449 |
+
"loss": 0.9416,
|
| 450 |
+
"step": 310
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.14,
|
| 454 |
+
"grad_norm": 0.08755150309465999,
|
| 455 |
+
"learning_rate": 0.00019879500874276786,
|
| 456 |
+
"loss": 0.9572,
|
| 457 |
+
"step": 315
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.15,
|
| 461 |
+
"grad_norm": 0.08859531100037746,
|
| 462 |
+
"learning_rate": 0.00019866786477984357,
|
| 463 |
+
"loss": 0.9583,
|
| 464 |
+
"step": 320
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.15,
|
| 468 |
+
"grad_norm": 0.08989870433700839,
|
| 469 |
+
"learning_rate": 0.0001985343900259446,
|
| 470 |
+
"loss": 0.9556,
|
| 471 |
+
"step": 325
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.15,
|
| 475 |
+
"grad_norm": 0.08858019305307584,
|
| 476 |
+
"learning_rate": 0.0001983945930451639,
|
| 477 |
+
"loss": 0.9396,
|
| 478 |
+
"step": 330
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.15,
|
| 482 |
+
"grad_norm": 0.09002132618468012,
|
| 483 |
+
"learning_rate": 0.0001982484828072452,
|
| 484 |
+
"loss": 0.9484,
|
| 485 |
+
"step": 335
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.16,
|
| 489 |
+
"grad_norm": 0.09371174635475092,
|
| 490 |
+
"learning_rate": 0.00019809606868700756,
|
| 491 |
+
"loss": 0.9783,
|
| 492 |
+
"step": 340
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.16,
|
| 496 |
+
"grad_norm": 0.08957560235407128,
|
| 497 |
+
"learning_rate": 0.00019793736046374373,
|
| 498 |
+
"loss": 0.9705,
|
| 499 |
+
"step": 345
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.16,
|
| 503 |
+
"grad_norm": 0.09036101764586872,
|
| 504 |
+
"learning_rate": 0.0001977723683205928,
|
| 505 |
+
"loss": 0.9621,
|
| 506 |
+
"step": 350
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.16,
|
| 510 |
+
"grad_norm": 0.08511370022873634,
|
| 511 |
+
"learning_rate": 0.00019760110284388667,
|
| 512 |
+
"loss": 0.9496,
|
| 513 |
+
"step": 355
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.17,
|
| 517 |
+
"grad_norm": 0.08897578619547024,
|
| 518 |
+
"learning_rate": 0.00019742357502247103,
|
| 519 |
+
"loss": 0.956,
|
| 520 |
+
"step": 360
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.17,
|
| 524 |
+
"grad_norm": 0.08600773035072362,
|
| 525 |
+
"learning_rate": 0.00019723979624700004,
|
| 526 |
+
"loss": 0.9696,
|
| 527 |
+
"step": 365
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.17,
|
| 531 |
+
"grad_norm": 0.09055825979523356,
|
| 532 |
+
"learning_rate": 0.0001970497783092057,
|
| 533 |
+
"loss": 0.9367,
|
| 534 |
+
"step": 370
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.17,
|
| 538 |
+
"grad_norm": 0.0932727351645862,
|
| 539 |
+
"learning_rate": 0.00019685353340114103,
|
| 540 |
+
"loss": 0.9744,
|
| 541 |
+
"step": 375
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.17,
|
| 545 |
+
"grad_norm": 0.0854363406688081,
|
| 546 |
+
"learning_rate": 0.00019665107411439805,
|
| 547 |
+
"loss": 0.9484,
|
| 548 |
+
"step": 380
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.18,
|
| 552 |
+
"grad_norm": 0.09005207643389745,
|
| 553 |
+
"learning_rate": 0.00019644241343929966,
|
| 554 |
+
"loss": 0.9677,
|
| 555 |
+
"step": 385
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.18,
|
| 559 |
+
"grad_norm": 0.08649622389550082,
|
| 560 |
+
"learning_rate": 0.00019622756476406626,
|
| 561 |
+
"loss": 0.9712,
|
| 562 |
+
"step": 390
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.18,
|
| 566 |
+
"grad_norm": 0.09439150706615457,
|
| 567 |
+
"learning_rate": 0.00019600654187395663,
|
| 568 |
+
"loss": 0.9683,
|
| 569 |
+
"step": 395
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.18,
|
| 573 |
+
"grad_norm": 0.09297365586336294,
|
| 574 |
+
"learning_rate": 0.00019577935895038361,
|
| 575 |
+
"loss": 0.9597,
|
| 576 |
+
"step": 400
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.19,
|
| 580 |
+
"grad_norm": 0.08529999606035547,
|
| 581 |
+
"learning_rate": 0.00019554603057000397,
|
| 582 |
+
"loss": 0.955,
|
| 583 |
+
"step": 405
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.19,
|
| 587 |
+
"grad_norm": 0.09078999187809153,
|
| 588 |
+
"learning_rate": 0.0001953065717037832,
|
| 589 |
+
"loss": 0.982,
|
| 590 |
+
"step": 410
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.19,
|
| 594 |
+
"grad_norm": 0.08821581975562078,
|
| 595 |
+
"learning_rate": 0.00019506099771603513,
|
| 596 |
+
"loss": 0.9413,
|
| 597 |
+
"step": 415
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.19,
|
| 601 |
+
"grad_norm": 0.09158552500876148,
|
| 602 |
+
"learning_rate": 0.00019480932436343582,
|
| 603 |
+
"loss": 0.948,
|
| 604 |
+
"step": 420
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.2,
|
| 608 |
+
"grad_norm": 0.08959534367059374,
|
| 609 |
+
"learning_rate": 0.00019455156779401265,
|
| 610 |
+
"loss": 0.9427,
|
| 611 |
+
"step": 425
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.2,
|
| 615 |
+
"grad_norm": 0.08425686082734027,
|
| 616 |
+
"learning_rate": 0.00019428774454610843,
|
| 617 |
+
"loss": 0.9687,
|
| 618 |
+
"step": 430
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.2,
|
| 622 |
+
"grad_norm": 0.09345564412190348,
|
| 623 |
+
"learning_rate": 0.00019401787154731993,
|
| 624 |
+
"loss": 0.9518,
|
| 625 |
+
"step": 435
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.2,
|
| 629 |
+
"grad_norm": 0.09078358962410667,
|
| 630 |
+
"learning_rate": 0.0001937419661134121,
|
| 631 |
+
"loss": 0.9188,
|
| 632 |
+
"step": 440
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.2,
|
| 636 |
+
"grad_norm": 0.08705601046979651,
|
| 637 |
+
"learning_rate": 0.00019346004594720669,
|
| 638 |
+
"loss": 0.9368,
|
| 639 |
+
"step": 445
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.21,
|
| 643 |
+
"grad_norm": 0.08422228533676696,
|
| 644 |
+
"learning_rate": 0.0001931721291374467,
|
| 645 |
+
"loss": 0.9464,
|
| 646 |
+
"step": 450
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.21,
|
| 650 |
+
"grad_norm": 0.08632474019571335,
|
| 651 |
+
"learning_rate": 0.00019287823415763553,
|
| 652 |
+
"loss": 0.9729,
|
| 653 |
+
"step": 455
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.21,
|
| 657 |
+
"grad_norm": 0.08755073393921248,
|
| 658 |
+
"learning_rate": 0.00019257837986485187,
|
| 659 |
+
"loss": 0.9544,
|
| 660 |
+
"step": 460
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.21,
|
| 664 |
+
"grad_norm": 0.08687000103849005,
|
| 665 |
+
"learning_rate": 0.0001922725854985396,
|
| 666 |
+
"loss": 0.9545,
|
| 667 |
+
"step": 465
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.22,
|
| 671 |
+
"grad_norm": 0.09216723149978304,
|
| 672 |
+
"learning_rate": 0.0001919608706792735,
|
| 673 |
+
"loss": 0.9597,
|
| 674 |
+
"step": 470
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.22,
|
| 678 |
+
"grad_norm": 0.0923192619958272,
|
| 679 |
+
"learning_rate": 0.00019164325540750016,
|
| 680 |
+
"loss": 0.9633,
|
| 681 |
+
"step": 475
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.22,
|
| 685 |
+
"grad_norm": 0.08732884649919875,
|
| 686 |
+
"learning_rate": 0.00019131976006225488,
|
| 687 |
+
"loss": 0.9693,
|
| 688 |
+
"step": 480
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.22,
|
| 692 |
+
"grad_norm": 0.08589978089749792,
|
| 693 |
+
"learning_rate": 0.00019099040539985394,
|
| 694 |
+
"loss": 0.9493,
|
| 695 |
+
"step": 485
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.22,
|
| 699 |
+
"grad_norm": 0.08762407567201845,
|
| 700 |
+
"learning_rate": 0.00019065521255256298,
|
| 701 |
+
"loss": 0.9664,
|
| 702 |
+
"step": 490
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.23,
|
| 706 |
+
"grad_norm": 0.09213229196529499,
|
| 707 |
+
"learning_rate": 0.00019031420302724096,
|
| 708 |
+
"loss": 0.9778,
|
| 709 |
+
"step": 495
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.23,
|
| 713 |
+
"grad_norm": 0.08592356697919659,
|
| 714 |
+
"learning_rate": 0.00018996739870396026,
|
| 715 |
+
"loss": 0.9572,
|
| 716 |
+
"step": 500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.23,
|
| 720 |
+
"grad_norm": 0.10623494849251007,
|
| 721 |
+
"learning_rate": 0.00018961482183460282,
|
| 722 |
+
"loss": 0.964,
|
| 723 |
+
"step": 505
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.23,
|
| 727 |
+
"grad_norm": 0.08704350244157279,
|
| 728 |
+
"learning_rate": 0.00018925649504143244,
|
| 729 |
+
"loss": 0.9524,
|
| 730 |
+
"step": 510
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.24,
|
| 734 |
+
"grad_norm": 0.0897677422263576,
|
| 735 |
+
"learning_rate": 0.0001888924413156432,
|
| 736 |
+
"loss": 0.9496,
|
| 737 |
+
"step": 515
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.24,
|
| 741 |
+
"grad_norm": 0.08827247508424733,
|
| 742 |
+
"learning_rate": 0.0001885226840158843,
|
| 743 |
+
"loss": 0.9358,
|
| 744 |
+
"step": 520
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.24,
|
| 748 |
+
"grad_norm": 0.08453360810699566,
|
| 749 |
+
"learning_rate": 0.00018814724686676133,
|
| 750 |
+
"loss": 0.9374,
|
| 751 |
+
"step": 525
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.24,
|
| 755 |
+
"grad_norm": 0.09448297684027758,
|
| 756 |
+
"learning_rate": 0.00018776615395731398,
|
| 757 |
+
"loss": 0.958,
|
| 758 |
+
"step": 530
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.25,
|
| 762 |
+
"grad_norm": 0.08600018025356833,
|
| 763 |
+
"learning_rate": 0.0001873794297394706,
|
| 764 |
+
"loss": 0.9618,
|
| 765 |
+
"step": 535
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.25,
|
| 769 |
+
"grad_norm": 0.08675234848801924,
|
| 770 |
+
"learning_rate": 0.00018698709902647902,
|
| 771 |
+
"loss": 0.9438,
|
| 772 |
+
"step": 540
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.25,
|
| 776 |
+
"grad_norm": 0.09061001820394972,
|
| 777 |
+
"learning_rate": 0.00018658918699131468,
|
| 778 |
+
"loss": 0.9687,
|
| 779 |
+
"step": 545
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.25,
|
| 783 |
+
"grad_norm": 0.0918295658678201,
|
| 784 |
+
"learning_rate": 0.00018618571916506545,
|
| 785 |
+
"loss": 0.9472,
|
| 786 |
+
"step": 550
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.25,
|
| 790 |
+
"grad_norm": 0.08596947713972145,
|
| 791 |
+
"learning_rate": 0.00018577672143529336,
|
| 792 |
+
"loss": 0.9516,
|
| 793 |
+
"step": 555
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.26,
|
| 797 |
+
"grad_norm": 0.09019003738558351,
|
| 798 |
+
"learning_rate": 0.00018536222004437368,
|
| 799 |
+
"loss": 0.9487,
|
| 800 |
+
"step": 560
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.26,
|
| 804 |
+
"grad_norm": 0.0847736196421309,
|
| 805 |
+
"learning_rate": 0.0001849422415878112,
|
| 806 |
+
"loss": 0.9314,
|
| 807 |
+
"step": 565
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.26,
|
| 811 |
+
"grad_norm": 0.08712414636411857,
|
| 812 |
+
"learning_rate": 0.00018451681301253362,
|
| 813 |
+
"loss": 0.9562,
|
| 814 |
+
"step": 570
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.26,
|
| 818 |
+
"grad_norm": 0.09276684050446858,
|
| 819 |
+
"learning_rate": 0.00018408596161516267,
|
| 820 |
+
"loss": 0.94,
|
| 821 |
+
"step": 575
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.27,
|
| 825 |
+
"grad_norm": 0.08820606105057767,
|
| 826 |
+
"learning_rate": 0.00018364971504026273,
|
| 827 |
+
"loss": 0.9546,
|
| 828 |
+
"step": 580
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.27,
|
| 832 |
+
"grad_norm": 0.08793901876987452,
|
| 833 |
+
"learning_rate": 0.00018320810127856705,
|
| 834 |
+
"loss": 0.9556,
|
| 835 |
+
"step": 585
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.27,
|
| 839 |
+
"grad_norm": 0.08756027329406087,
|
| 840 |
+
"learning_rate": 0.00018276114866518168,
|
| 841 |
+
"loss": 0.9647,
|
| 842 |
+
"step": 590
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.27,
|
| 846 |
+
"grad_norm": 0.08612057766614822,
|
| 847 |
+
"learning_rate": 0.00018230888587776755,
|
| 848 |
+
"loss": 0.9348,
|
| 849 |
+
"step": 595
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.28,
|
| 853 |
+
"grad_norm": 0.08825824481464735,
|
| 854 |
+
"learning_rate": 0.00018185134193470043,
|
| 855 |
+
"loss": 0.9621,
|
| 856 |
+
"step": 600
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.28,
|
| 860 |
+
"grad_norm": 0.08855957551515957,
|
| 861 |
+
"learning_rate": 0.00018138854619320893,
|
| 862 |
+
"loss": 0.951,
|
| 863 |
+
"step": 605
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.28,
|
| 867 |
+
"grad_norm": 0.09202042855847319,
|
| 868 |
+
"learning_rate": 0.00018092052834749094,
|
| 869 |
+
"loss": 0.9524,
|
| 870 |
+
"step": 610
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.28,
|
| 874 |
+
"grad_norm": 0.0882422487809607,
|
| 875 |
+
"learning_rate": 0.0001804473184268084,
|
| 876 |
+
"loss": 0.9344,
|
| 877 |
+
"step": 615
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.28,
|
| 881 |
+
"grad_norm": 0.09188481277352664,
|
| 882 |
+
"learning_rate": 0.0001799689467935604,
|
| 883 |
+
"loss": 0.9575,
|
| 884 |
+
"step": 620
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.29,
|
| 888 |
+
"grad_norm": 0.08761250216320178,
|
| 889 |
+
"learning_rate": 0.00017948544414133534,
|
| 890 |
+
"loss": 0.9353,
|
| 891 |
+
"step": 625
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.29,
|
| 895 |
+
"grad_norm": 0.09089274409603867,
|
| 896 |
+
"learning_rate": 0.00017899684149294117,
|
| 897 |
+
"loss": 0.964,
|
| 898 |
+
"step": 630
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.29,
|
| 902 |
+
"grad_norm": 0.08766747210542415,
|
| 903 |
+
"learning_rate": 0.00017850317019841514,
|
| 904 |
+
"loss": 0.9525,
|
| 905 |
+
"step": 635
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.29,
|
| 909 |
+
"grad_norm": 0.08930600295491815,
|
| 910 |
+
"learning_rate": 0.00017800446193301227,
|
| 911 |
+
"loss": 0.96,
|
| 912 |
+
"step": 640
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.3,
|
| 916 |
+
"grad_norm": 0.08853887368794433,
|
| 917 |
+
"learning_rate": 0.00017750074869517284,
|
| 918 |
+
"loss": 0.9694,
|
| 919 |
+
"step": 645
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.3,
|
| 923 |
+
"grad_norm": 0.08442533232702532,
|
| 924 |
+
"learning_rate": 0.00017699206280446953,
|
| 925 |
+
"loss": 0.9208,
|
| 926 |
+
"step": 650
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.3,
|
| 930 |
+
"grad_norm": 0.08683714377922135,
|
| 931 |
+
"learning_rate": 0.00017647843689953352,
|
| 932 |
+
"loss": 0.9661,
|
| 933 |
+
"step": 655
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.3,
|
| 937 |
+
"grad_norm": 0.08674144789262467,
|
| 938 |
+
"learning_rate": 0.00017595990393596026,
|
| 939 |
+
"loss": 0.9586,
|
| 940 |
+
"step": 660
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.31,
|
| 944 |
+
"grad_norm": 0.08859684741151846,
|
| 945 |
+
"learning_rate": 0.0001754364971841952,
|
| 946 |
+
"loss": 0.9581,
|
| 947 |
+
"step": 665
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.31,
|
| 951 |
+
"grad_norm": 0.08859618436822166,
|
| 952 |
+
"learning_rate": 0.0001749082502273988,
|
| 953 |
+
"loss": 0.9396,
|
| 954 |
+
"step": 670
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.31,
|
| 958 |
+
"grad_norm": 0.08662436024506778,
|
| 959 |
+
"learning_rate": 0.00017437519695929194,
|
| 960 |
+
"loss": 0.9478,
|
| 961 |
+
"step": 675
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.31,
|
| 965 |
+
"grad_norm": 0.08702673421879521,
|
| 966 |
+
"learning_rate": 0.0001738373715819811,
|
| 967 |
+
"loss": 0.9659,
|
| 968 |
+
"step": 680
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.31,
|
| 972 |
+
"grad_norm": 0.08834975009894457,
|
| 973 |
+
"learning_rate": 0.00017329480860376391,
|
| 974 |
+
"loss": 0.9801,
|
| 975 |
+
"step": 685
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.32,
|
| 979 |
+
"grad_norm": 0.08907394268442888,
|
| 980 |
+
"learning_rate": 0.00017274754283691504,
|
| 981 |
+
"loss": 0.9527,
|
| 982 |
+
"step": 690
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.32,
|
| 986 |
+
"grad_norm": 0.09231370364779791,
|
| 987 |
+
"learning_rate": 0.00017219560939545246,
|
| 988 |
+
"loss": 0.9588,
|
| 989 |
+
"step": 695
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.32,
|
| 993 |
+
"grad_norm": 0.09009781412735576,
|
| 994 |
+
"learning_rate": 0.00017163904369288445,
|
| 995 |
+
"loss": 0.9419,
|
| 996 |
+
"step": 700
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.32,
|
| 1000 |
+
"grad_norm": 0.08650205473210101,
|
| 1001 |
+
"learning_rate": 0.00017107788143993743,
|
| 1002 |
+
"loss": 0.94,
|
| 1003 |
+
"step": 705
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.33,
|
| 1007 |
+
"grad_norm": 0.0919904852442932,
|
| 1008 |
+
"learning_rate": 0.00017051215864226469,
|
| 1009 |
+
"loss": 0.9506,
|
| 1010 |
+
"step": 710
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.33,
|
| 1014 |
+
"grad_norm": 0.08808529519659543,
|
| 1015 |
+
"learning_rate": 0.0001699419115981361,
|
| 1016 |
+
"loss": 0.957,
|
| 1017 |
+
"step": 715
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.33,
|
| 1021 |
+
"grad_norm": 0.09101444782991605,
|
| 1022 |
+
"learning_rate": 0.00016936717689610903,
|
| 1023 |
+
"loss": 0.9328,
|
| 1024 |
+
"step": 720
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.33,
|
| 1028 |
+
"grad_norm": 0.0865079634420763,
|
| 1029 |
+
"learning_rate": 0.00016878799141268106,
|
| 1030 |
+
"loss": 0.9421,
|
| 1031 |
+
"step": 725
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.34,
|
| 1035 |
+
"grad_norm": 0.0895608019365179,
|
| 1036 |
+
"learning_rate": 0.00016820439230992343,
|
| 1037 |
+
"loss": 0.9618,
|
| 1038 |
+
"step": 730
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.34,
|
| 1042 |
+
"grad_norm": 0.08619071121372446,
|
| 1043 |
+
"learning_rate": 0.00016761641703309704,
|
| 1044 |
+
"loss": 0.9265,
|
| 1045 |
+
"step": 735
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.34,
|
| 1049 |
+
"grad_norm": 0.0848234059929775,
|
| 1050 |
+
"learning_rate": 0.00016702410330824962,
|
| 1051 |
+
"loss": 0.9701,
|
| 1052 |
+
"step": 740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.34,
|
| 1056 |
+
"grad_norm": 0.08582375402337497,
|
| 1057 |
+
"learning_rate": 0.00016642748913979513,
|
| 1058 |
+
"loss": 0.9484,
|
| 1059 |
+
"step": 745
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.34,
|
| 1063 |
+
"grad_norm": 0.09152872800299702,
|
| 1064 |
+
"learning_rate": 0.00016582661280807552,
|
| 1065 |
+
"loss": 0.9639,
|
| 1066 |
+
"step": 750
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.35,
|
| 1070 |
+
"grad_norm": 0.0877108383812197,
|
| 1071 |
+
"learning_rate": 0.0001652215128669042,
|
| 1072 |
+
"loss": 0.9787,
|
| 1073 |
+
"step": 755
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.35,
|
| 1077 |
+
"grad_norm": 0.08689501691870347,
|
| 1078 |
+
"learning_rate": 0.00016461222814109268,
|
| 1079 |
+
"loss": 0.9608,
|
| 1080 |
+
"step": 760
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.35,
|
| 1084 |
+
"grad_norm": 0.09332436047860204,
|
| 1085 |
+
"learning_rate": 0.00016399879772395915,
|
| 1086 |
+
"loss": 0.9432,
|
| 1087 |
+
"step": 765
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.35,
|
| 1091 |
+
"grad_norm": 0.08516705345764235,
|
| 1092 |
+
"learning_rate": 0.00016338126097482057,
|
| 1093 |
+
"loss": 0.9348,
|
| 1094 |
+
"step": 770
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.36,
|
| 1098 |
+
"grad_norm": 0.08726474252968008,
|
| 1099 |
+
"learning_rate": 0.0001627596575164668,
|
| 1100 |
+
"loss": 0.9709,
|
| 1101 |
+
"step": 775
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.36,
|
| 1105 |
+
"grad_norm": 0.08888011508247526,
|
| 1106 |
+
"learning_rate": 0.0001621340272326185,
|
| 1107 |
+
"loss": 0.956,
|
| 1108 |
+
"step": 780
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.36,
|
| 1112 |
+
"grad_norm": 0.09139409277427421,
|
| 1113 |
+
"learning_rate": 0.00016150441026536823,
|
| 1114 |
+
"loss": 0.9483,
|
| 1115 |
+
"step": 785
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.36,
|
| 1119 |
+
"grad_norm": 0.09042448727858429,
|
| 1120 |
+
"learning_rate": 0.00016087084701260466,
|
| 1121 |
+
"loss": 0.9526,
|
| 1122 |
+
"step": 790
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.36,
|
| 1126 |
+
"grad_norm": 0.08642751941422035,
|
| 1127 |
+
"learning_rate": 0.00016023337812542048,
|
| 1128 |
+
"loss": 0.9442,
|
| 1129 |
+
"step": 795
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.37,
|
| 1133 |
+
"grad_norm": 0.08584355244570809,
|
| 1134 |
+
"learning_rate": 0.00015959204450550428,
|
| 1135 |
+
"loss": 0.9492,
|
| 1136 |
+
"step": 800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.37,
|
| 1140 |
+
"grad_norm": 0.08766167281479247,
|
| 1141 |
+
"learning_rate": 0.00015894688730251614,
|
| 1142 |
+
"loss": 0.9787,
|
| 1143 |
+
"step": 805
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.37,
|
| 1147 |
+
"grad_norm": 0.0883622912274688,
|
| 1148 |
+
"learning_rate": 0.0001582979479114472,
|
| 1149 |
+
"loss": 0.9466,
|
| 1150 |
+
"step": 810
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.37,
|
| 1154 |
+
"grad_norm": 0.09140030196293039,
|
| 1155 |
+
"learning_rate": 0.000157645267969964,
|
| 1156 |
+
"loss": 0.9495,
|
| 1157 |
+
"step": 815
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.38,
|
| 1161 |
+
"grad_norm": 0.08506975393435418,
|
| 1162 |
+
"learning_rate": 0.0001569888893557365,
|
| 1163 |
+
"loss": 0.9279,
|
| 1164 |
+
"step": 820
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.38,
|
| 1168 |
+
"grad_norm": 0.08519765362403448,
|
| 1169 |
+
"learning_rate": 0.00015632885418375136,
|
| 1170 |
+
"loss": 0.9509,
|
| 1171 |
+
"step": 825
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.38,
|
| 1175 |
+
"grad_norm": 0.08778850238010608,
|
| 1176 |
+
"learning_rate": 0.00015566520480360957,
|
| 1177 |
+
"loss": 0.9161,
|
| 1178 |
+
"step": 830
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.38,
|
| 1182 |
+
"grad_norm": 0.09158195757138242,
|
| 1183 |
+
"learning_rate": 0.0001549979837968094,
|
| 1184 |
+
"loss": 0.9245,
|
| 1185 |
+
"step": 835
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.39,
|
| 1189 |
+
"grad_norm": 0.08986630430709179,
|
| 1190 |
+
"learning_rate": 0.00015432723397401405,
|
| 1191 |
+
"loss": 0.9455,
|
| 1192 |
+
"step": 840
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.39,
|
| 1196 |
+
"grad_norm": 0.09336343830360828,
|
| 1197 |
+
"learning_rate": 0.00015365299837230483,
|
| 1198 |
+
"loss": 0.9584,
|
| 1199 |
+
"step": 845
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.39,
|
| 1203 |
+
"grad_norm": 0.08854767057712627,
|
| 1204 |
+
"learning_rate": 0.00015297532025241992,
|
| 1205 |
+
"loss": 0.9373,
|
| 1206 |
+
"step": 850
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.39,
|
| 1210 |
+
"grad_norm": 0.0861561244367331,
|
| 1211 |
+
"learning_rate": 0.00015229424309597852,
|
| 1212 |
+
"loss": 0.9337,
|
| 1213 |
+
"step": 855
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.39,
|
| 1217 |
+
"grad_norm": 0.09273027341749242,
|
| 1218 |
+
"learning_rate": 0.00015160981060269107,
|
| 1219 |
+
"loss": 0.9492,
|
| 1220 |
+
"step": 860
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.4,
|
| 1224 |
+
"grad_norm": 0.09378203137542973,
|
| 1225 |
+
"learning_rate": 0.00015092206668755517,
|
| 1226 |
+
"loss": 0.9463,
|
| 1227 |
+
"step": 865
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.4,
|
| 1231 |
+
"grad_norm": 0.0862644522566823,
|
| 1232 |
+
"learning_rate": 0.00015023105547803806,
|
| 1233 |
+
"loss": 0.9339,
|
| 1234 |
+
"step": 870
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.4,
|
| 1238 |
+
"grad_norm": 0.08790661308759892,
|
| 1239 |
+
"learning_rate": 0.00014953682131124528,
|
| 1240 |
+
"loss": 0.9474,
|
| 1241 |
+
"step": 875
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.4,
|
| 1245 |
+
"grad_norm": 0.08422068379445599,
|
| 1246 |
+
"learning_rate": 0.0001488394087310757,
|
| 1247 |
+
"loss": 0.932,
|
| 1248 |
+
"step": 880
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.41,
|
| 1252 |
+
"grad_norm": 0.08835895673531197,
|
| 1253 |
+
"learning_rate": 0.00014813886248536375,
|
| 1254 |
+
"loss": 0.9682,
|
| 1255 |
+
"step": 885
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.41,
|
| 1259 |
+
"grad_norm": 0.08812707061816671,
|
| 1260 |
+
"learning_rate": 0.00014743522752300793,
|
| 1261 |
+
"loss": 0.9494,
|
| 1262 |
+
"step": 890
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.41,
|
| 1266 |
+
"grad_norm": 0.08984312676721799,
|
| 1267 |
+
"learning_rate": 0.00014672854899108718,
|
| 1268 |
+
"loss": 0.968,
|
| 1269 |
+
"step": 895
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.41,
|
| 1273 |
+
"grad_norm": 0.08997766999205116,
|
| 1274 |
+
"learning_rate": 0.00014601887223196372,
|
| 1275 |
+
"loss": 0.9259,
|
| 1276 |
+
"step": 900
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.42,
|
| 1280 |
+
"grad_norm": 0.08554195694203166,
|
| 1281 |
+
"learning_rate": 0.00014530624278037406,
|
| 1282 |
+
"loss": 0.9325,
|
| 1283 |
+
"step": 905
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.42,
|
| 1287 |
+
"grad_norm": 0.08600975689157553,
|
| 1288 |
+
"learning_rate": 0.0001445907063605072,
|
| 1289 |
+
"loss": 0.9396,
|
| 1290 |
+
"step": 910
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.42,
|
| 1294 |
+
"grad_norm": 0.08549667691396576,
|
| 1295 |
+
"learning_rate": 0.00014387230888307097,
|
| 1296 |
+
"loss": 0.931,
|
| 1297 |
+
"step": 915
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.42,
|
| 1301 |
+
"grad_norm": 0.08633848719229598,
|
| 1302 |
+
"learning_rate": 0.0001431510964423462,
|
| 1303 |
+
"loss": 0.9736,
|
| 1304 |
+
"step": 920
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.42,
|
| 1308 |
+
"grad_norm": 0.08946502964232063,
|
| 1309 |
+
"learning_rate": 0.0001424271153132291,
|
| 1310 |
+
"loss": 0.9508,
|
| 1311 |
+
"step": 925
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.43,
|
| 1315 |
+
"grad_norm": 0.08644778095141836,
|
| 1316 |
+
"learning_rate": 0.00014170041194826248,
|
| 1317 |
+
"loss": 0.9291,
|
| 1318 |
+
"step": 930
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.43,
|
| 1322 |
+
"grad_norm": 0.08743477799274565,
|
| 1323 |
+
"learning_rate": 0.0001409710329746547,
|
| 1324 |
+
"loss": 0.9581,
|
| 1325 |
+
"step": 935
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.43,
|
| 1329 |
+
"grad_norm": 0.09102927350920745,
|
| 1330 |
+
"learning_rate": 0.0001402390251912885,
|
| 1331 |
+
"loss": 0.9641,
|
| 1332 |
+
"step": 940
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.43,
|
| 1336 |
+
"grad_norm": 0.08406401589929084,
|
| 1337 |
+
"learning_rate": 0.00013950443556571778,
|
| 1338 |
+
"loss": 0.928,
|
| 1339 |
+
"step": 945
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.44,
|
| 1343 |
+
"grad_norm": 0.08874042668156715,
|
| 1344 |
+
"learning_rate": 0.0001387673112311545,
|
| 1345 |
+
"loss": 0.911,
|
| 1346 |
+
"step": 950
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.44,
|
| 1350 |
+
"grad_norm": 0.0846652407845296,
|
| 1351 |
+
"learning_rate": 0.00013802769948344405,
|
| 1352 |
+
"loss": 0.93,
|
| 1353 |
+
"step": 955
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.44,
|
| 1357 |
+
"grad_norm": 0.08812982695434313,
|
| 1358 |
+
"learning_rate": 0.00013728564777803088,
|
| 1359 |
+
"loss": 0.9574,
|
| 1360 |
+
"step": 960
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.44,
|
| 1364 |
+
"grad_norm": 0.08779218610162595,
|
| 1365 |
+
"learning_rate": 0.0001365412037269136,
|
| 1366 |
+
"loss": 0.9263,
|
| 1367 |
+
"step": 965
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.45,
|
| 1371 |
+
"grad_norm": 0.08763397205182845,
|
| 1372 |
+
"learning_rate": 0.00013579441509559004,
|
| 1373 |
+
"loss": 0.9647,
|
| 1374 |
+
"step": 970
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.45,
|
| 1378 |
+
"grad_norm": 0.08771463604901869,
|
| 1379 |
+
"learning_rate": 0.0001350453297999925,
|
| 1380 |
+
"loss": 0.9422,
|
| 1381 |
+
"step": 975
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.45,
|
| 1385 |
+
"grad_norm": 0.09110171303652766,
|
| 1386 |
+
"learning_rate": 0.00013429399590341324,
|
| 1387 |
+
"loss": 0.947,
|
| 1388 |
+
"step": 980
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.45,
|
| 1392 |
+
"grad_norm": 0.0890607024116997,
|
| 1393 |
+
"learning_rate": 0.00013354046161342085,
|
| 1394 |
+
"loss": 0.9513,
|
| 1395 |
+
"step": 985
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.45,
|
| 1399 |
+
"grad_norm": 0.0859287161459074,
|
| 1400 |
+
"learning_rate": 0.0001327847752787669,
|
| 1401 |
+
"loss": 0.9706,
|
| 1402 |
+
"step": 990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.46,
|
| 1406 |
+
"grad_norm": 0.09247919961374537,
|
| 1407 |
+
"learning_rate": 0.00013202698538628376,
|
| 1408 |
+
"loss": 0.9657,
|
| 1409 |
+
"step": 995
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.46,
|
| 1413 |
+
"grad_norm": 0.09027056948690071,
|
| 1414 |
+
"learning_rate": 0.00013126714055777377,
|
| 1415 |
+
"loss": 0.9527,
|
| 1416 |
+
"step": 1000
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.46,
|
| 1420 |
+
"grad_norm": 0.08674017096568151,
|
| 1421 |
+
"learning_rate": 0.0001305052895468893,
|
| 1422 |
+
"loss": 0.9519,
|
| 1423 |
+
"step": 1005
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.46,
|
| 1427 |
+
"grad_norm": 0.08650629459230609,
|
| 1428 |
+
"learning_rate": 0.00012974148123600477,
|
| 1429 |
+
"loss": 0.9395,
|
| 1430 |
+
"step": 1010
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.47,
|
| 1434 |
+
"grad_norm": 0.08523687321318427,
|
| 1435 |
+
"learning_rate": 0.00012897576463307997,
|
| 1436 |
+
"loss": 0.9433,
|
| 1437 |
+
"step": 1015
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.47,
|
| 1441 |
+
"grad_norm": 0.08649388719980053,
|
| 1442 |
+
"learning_rate": 0.00012820818886851598,
|
| 1443 |
+
"loss": 0.9212,
|
| 1444 |
+
"step": 1020
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.47,
|
| 1448 |
+
"grad_norm": 0.08632252078335054,
|
| 1449 |
+
"learning_rate": 0.00012743880319200242,
|
| 1450 |
+
"loss": 0.9677,
|
| 1451 |
+
"step": 1025
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.47,
|
| 1455 |
+
"grad_norm": 0.08776100128200802,
|
| 1456 |
+
"learning_rate": 0.00012666765696935773,
|
| 1457 |
+
"loss": 0.9469,
|
| 1458 |
+
"step": 1030
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.47,
|
| 1462 |
+
"grad_norm": 0.08935277960845327,
|
| 1463 |
+
"learning_rate": 0.0001258947996793616,
|
| 1464 |
+
"loss": 0.9354,
|
| 1465 |
+
"step": 1035
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.48,
|
| 1469 |
+
"grad_norm": 0.09034950318804204,
|
| 1470 |
+
"learning_rate": 0.00012512028091058044,
|
| 1471 |
+
"loss": 0.9541,
|
| 1472 |
+
"step": 1040
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.48,
|
| 1476 |
+
"grad_norm": 0.08871578654302656,
|
| 1477 |
+
"learning_rate": 0.00012434415035818535,
|
| 1478 |
+
"loss": 0.9435,
|
| 1479 |
+
"step": 1045
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.48,
|
| 1483 |
+
"grad_norm": 0.08906304568486471,
|
| 1484 |
+
"learning_rate": 0.00012356645782076383,
|
| 1485 |
+
"loss": 0.9673,
|
| 1486 |
+
"step": 1050
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.48,
|
| 1490 |
+
"grad_norm": 0.08552583500630938,
|
| 1491 |
+
"learning_rate": 0.00012278725319712447,
|
| 1492 |
+
"loss": 0.9545,
|
| 1493 |
+
"step": 1055
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.49,
|
| 1497 |
+
"grad_norm": 0.08555288016156151,
|
| 1498 |
+
"learning_rate": 0.00012200658648309531,
|
| 1499 |
+
"loss": 0.9552,
|
| 1500 |
+
"step": 1060
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.49,
|
| 1504 |
+
"grad_norm": 0.08664206961589904,
|
| 1505 |
+
"learning_rate": 0.00012122450776831594,
|
| 1506 |
+
"loss": 0.9463,
|
| 1507 |
+
"step": 1065
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.49,
|
| 1511 |
+
"grad_norm": 0.08904216070092703,
|
| 1512 |
+
"learning_rate": 0.00012044106723302364,
|
| 1513 |
+
"loss": 0.9489,
|
| 1514 |
+
"step": 1070
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.49,
|
| 1518 |
+
"grad_norm": 0.08709170070909589,
|
| 1519 |
+
"learning_rate": 0.00011965631514483375,
|
| 1520 |
+
"loss": 0.9475,
|
| 1521 |
+
"step": 1075
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.5,
|
| 1525 |
+
"grad_norm": 0.08790044008622408,
|
| 1526 |
+
"learning_rate": 0.00011887030185551426,
|
| 1527 |
+
"loss": 0.9446,
|
| 1528 |
+
"step": 1080
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.5,
|
| 1532 |
+
"grad_norm": 0.08658875670552062,
|
| 1533 |
+
"learning_rate": 0.00011808307779775518,
|
| 1534 |
+
"loss": 0.9411,
|
| 1535 |
+
"step": 1085
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.5,
|
| 1539 |
+
"grad_norm": 0.0977471754063036,
|
| 1540 |
+
"learning_rate": 0.00011729469348193262,
|
| 1541 |
+
"loss": 0.9255,
|
| 1542 |
+
"step": 1090
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.5,
|
| 1546 |
+
"grad_norm": 0.08804945539452566,
|
| 1547 |
+
"learning_rate": 0.00011650519949286795,
|
| 1548 |
+
"loss": 0.9802,
|
| 1549 |
+
"step": 1095
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.5,
|
| 1553 |
+
"grad_norm": 0.08581976314554501,
|
| 1554 |
+
"learning_rate": 0.00011571464648658201,
|
| 1555 |
+
"loss": 0.9383,
|
| 1556 |
+
"step": 1100
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.51,
|
| 1560 |
+
"grad_norm": 0.08839470223710093,
|
| 1561 |
+
"learning_rate": 0.00011492308518704506,
|
| 1562 |
+
"loss": 0.9441,
|
| 1563 |
+
"step": 1105
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.51,
|
| 1567 |
+
"grad_norm": 0.08941383297565787,
|
| 1568 |
+
"learning_rate": 0.00011413056638292215,
|
| 1569 |
+
"loss": 0.9515,
|
| 1570 |
+
"step": 1110
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.51,
|
| 1574 |
+
"grad_norm": 0.08584986185665876,
|
| 1575 |
+
"learning_rate": 0.00011333714092431423,
|
| 1576 |
+
"loss": 0.9309,
|
| 1577 |
+
"step": 1115
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.51,
|
| 1581 |
+
"grad_norm": 0.0865352790700451,
|
| 1582 |
+
"learning_rate": 0.00011254285971949574,
|
| 1583 |
+
"loss": 0.9335,
|
| 1584 |
+
"step": 1120
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.52,
|
| 1588 |
+
"grad_norm": 0.0863789933194355,
|
| 1589 |
+
"learning_rate": 0.00011174777373164798,
|
| 1590 |
+
"loss": 0.9487,
|
| 1591 |
+
"step": 1125
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.52,
|
| 1595 |
+
"grad_norm": 0.08729028323407131,
|
| 1596 |
+
"learning_rate": 0.0001109519339755893,
|
| 1597 |
+
"loss": 0.9381,
|
| 1598 |
+
"step": 1130
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.52,
|
| 1602 |
+
"grad_norm": 0.08674945320440842,
|
| 1603 |
+
"learning_rate": 0.00011015539151450172,
|
| 1604 |
+
"loss": 0.9398,
|
| 1605 |
+
"step": 1135
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.52,
|
| 1609 |
+
"grad_norm": 0.09452409408655012,
|
| 1610 |
+
"learning_rate": 0.00010935819745665477,
|
| 1611 |
+
"loss": 0.9709,
|
| 1612 |
+
"step": 1140
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.53,
|
| 1616 |
+
"grad_norm": 0.0862575634432857,
|
| 1617 |
+
"learning_rate": 0.00010856040295212612,
|
| 1618 |
+
"loss": 0.9306,
|
| 1619 |
+
"step": 1145
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.53,
|
| 1623 |
+
"grad_norm": 0.08968733351995441,
|
| 1624 |
+
"learning_rate": 0.00010776205918951969,
|
| 1625 |
+
"loss": 0.9573,
|
| 1626 |
+
"step": 1150
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.53,
|
| 1630 |
+
"grad_norm": 0.0901921037984152,
|
| 1631 |
+
"learning_rate": 0.00010696321739268121,
|
| 1632 |
+
"loss": 0.964,
|
| 1633 |
+
"step": 1155
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.53,
|
| 1637 |
+
"grad_norm": 0.08958751346326631,
|
| 1638 |
+
"learning_rate": 0.00010616392881741166,
|
| 1639 |
+
"loss": 0.9563,
|
| 1640 |
+
"step": 1160
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.53,
|
| 1644 |
+
"grad_norm": 0.08935150733851532,
|
| 1645 |
+
"learning_rate": 0.00010536424474817847,
|
| 1646 |
+
"loss": 0.9281,
|
| 1647 |
+
"step": 1165
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.54,
|
| 1651 |
+
"grad_norm": 0.08505482136944548,
|
| 1652 |
+
"learning_rate": 0.00010456421649482502,
|
| 1653 |
+
"loss": 0.9522,
|
| 1654 |
+
"step": 1170
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.54,
|
| 1658 |
+
"grad_norm": 0.08977467093508296,
|
| 1659 |
+
"learning_rate": 0.0001037638953892784,
|
| 1660 |
+
"loss": 0.9515,
|
| 1661 |
+
"step": 1175
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.54,
|
| 1665 |
+
"grad_norm": 0.09103794206490902,
|
| 1666 |
+
"learning_rate": 0.00010296333278225599,
|
| 1667 |
+
"loss": 0.9488,
|
| 1668 |
+
"step": 1180
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.54,
|
| 1672 |
+
"grad_norm": 0.08855592795266524,
|
| 1673 |
+
"learning_rate": 0.00010216258003997043,
|
| 1674 |
+
"loss": 0.9382,
|
| 1675 |
+
"step": 1185
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.55,
|
| 1679 |
+
"grad_norm": 0.08551252341505816,
|
| 1680 |
+
"learning_rate": 0.00010136168854083402,
|
| 1681 |
+
"loss": 0.9592,
|
| 1682 |
+
"step": 1190
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.55,
|
| 1686 |
+
"grad_norm": 0.08711479138102324,
|
| 1687 |
+
"learning_rate": 0.00010056070967216198,
|
| 1688 |
+
"loss": 0.959,
|
| 1689 |
+
"step": 1195
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.55,
|
| 1693 |
+
"grad_norm": 0.08549275625721742,
|
| 1694 |
+
"learning_rate": 9.975969482687547e-05,
|
| 1695 |
+
"loss": 0.9449,
|
| 1696 |
+
"step": 1200
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.55,
|
| 1700 |
+
"grad_norm": 0.08888803602947658,
|
| 1701 |
+
"learning_rate": 9.8958695400204e-05,
|
| 1702 |
+
"loss": 0.9287,
|
| 1703 |
+
"step": 1205
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.56,
|
| 1707 |
+
"grad_norm": 0.08689172899821032,
|
| 1708 |
+
"learning_rate": 9.815776278638771e-05,
|
| 1709 |
+
"loss": 0.94,
|
| 1710 |
+
"step": 1210
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.56,
|
| 1714 |
+
"grad_norm": 0.09137637622059057,
|
| 1715 |
+
"learning_rate": 9.735694837537993e-05,
|
| 1716 |
+
"loss": 0.967,
|
| 1717 |
+
"step": 1215
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.56,
|
| 1721 |
+
"grad_norm": 0.08486290429115295,
|
| 1722 |
+
"learning_rate": 9.655630354954974e-05,
|
| 1723 |
+
"loss": 0.9324,
|
| 1724 |
+
"step": 1220
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.56,
|
| 1728 |
+
"grad_norm": 0.08503393222349716,
|
| 1729 |
+
"learning_rate": 9.57558796803852e-05,
|
| 1730 |
+
"loss": 0.9494,
|
| 1731 |
+
"step": 1225
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.56,
|
| 1735 |
+
"grad_norm": 0.0873443558033662,
|
| 1736 |
+
"learning_rate": 9.495572812519718e-05,
|
| 1737 |
+
"loss": 0.9506,
|
| 1738 |
+
"step": 1230
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.57,
|
| 1742 |
+
"grad_norm": 0.09121014808349219,
|
| 1743 |
+
"learning_rate": 9.415590022382419e-05,
|
| 1744 |
+
"loss": 0.9252,
|
| 1745 |
+
"step": 1235
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.57,
|
| 1749 |
+
"grad_norm": 0.086584430175237,
|
| 1750 |
+
"learning_rate": 9.33564472953383e-05,
|
| 1751 |
+
"loss": 0.9439,
|
| 1752 |
+
"step": 1240
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.57,
|
| 1756 |
+
"grad_norm": 0.08692310082632433,
|
| 1757 |
+
"learning_rate": 9.255742063475228e-05,
|
| 1758 |
+
"loss": 0.9297,
|
| 1759 |
+
"step": 1245
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.57,
|
| 1763 |
+
"grad_norm": 0.08990215639751623,
|
| 1764 |
+
"learning_rate": 9.175887150972841e-05,
|
| 1765 |
+
"loss": 0.9629,
|
| 1766 |
+
"step": 1250
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.58,
|
| 1770 |
+
"grad_norm": 0.08963253963139663,
|
| 1771 |
+
"learning_rate": 9.096085115728902e-05,
|
| 1772 |
+
"loss": 0.9642,
|
| 1773 |
+
"step": 1255
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.58,
|
| 1777 |
+
"grad_norm": 0.09194536883220446,
|
| 1778 |
+
"learning_rate": 9.016341078052908e-05,
|
| 1779 |
+
"loss": 0.9308,
|
| 1780 |
+
"step": 1260
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.58,
|
| 1784 |
+
"grad_norm": 0.0911811600503115,
|
| 1785 |
+
"learning_rate": 8.936660154533069e-05,
|
| 1786 |
+
"loss": 0.952,
|
| 1787 |
+
"step": 1265
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.58,
|
| 1791 |
+
"grad_norm": 0.08845398994221279,
|
| 1792 |
+
"learning_rate": 8.857047457708023e-05,
|
| 1793 |
+
"loss": 0.9676,
|
| 1794 |
+
"step": 1270
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.59,
|
| 1798 |
+
"grad_norm": 0.08623567977555162,
|
| 1799 |
+
"learning_rate": 8.777508095738818e-05,
|
| 1800 |
+
"loss": 0.9277,
|
| 1801 |
+
"step": 1275
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.59,
|
| 1805 |
+
"grad_norm": 0.0857316423044204,
|
| 1806 |
+
"learning_rate": 8.698047172081128e-05,
|
| 1807 |
+
"loss": 0.9472,
|
| 1808 |
+
"step": 1280
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.59,
|
| 1812 |
+
"grad_norm": 0.09001530302196627,
|
| 1813 |
+
"learning_rate": 8.618669785157825e-05,
|
| 1814 |
+
"loss": 0.9304,
|
| 1815 |
+
"step": 1285
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.59,
|
| 1819 |
+
"grad_norm": 0.09938874508667282,
|
| 1820 |
+
"learning_rate": 8.539381028031837e-05,
|
| 1821 |
+
"loss": 0.9644,
|
| 1822 |
+
"step": 1290
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.59,
|
| 1826 |
+
"grad_norm": 0.08956519091936009,
|
| 1827 |
+
"learning_rate": 8.460185988079379e-05,
|
| 1828 |
+
"loss": 0.9244,
|
| 1829 |
+
"step": 1295
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.6,
|
| 1833 |
+
"grad_norm": 0.09301102302473901,
|
| 1834 |
+
"learning_rate": 8.381089746663517e-05,
|
| 1835 |
+
"loss": 0.9469,
|
| 1836 |
+
"step": 1300
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.6,
|
| 1840 |
+
"grad_norm": 0.08856088507182096,
|
| 1841 |
+
"learning_rate": 8.302097378808146e-05,
|
| 1842 |
+
"loss": 0.9421,
|
| 1843 |
+
"step": 1305
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.6,
|
| 1847 |
+
"grad_norm": 0.08829957085883325,
|
| 1848 |
+
"learning_rate": 8.223213952872353e-05,
|
| 1849 |
+
"loss": 0.9752,
|
| 1850 |
+
"step": 1310
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.6,
|
| 1854 |
+
"grad_norm": 0.09355532885262087,
|
| 1855 |
+
"learning_rate": 8.144444530225236e-05,
|
| 1856 |
+
"loss": 0.9201,
|
| 1857 |
+
"step": 1315
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.61,
|
| 1861 |
+
"grad_norm": 0.08851734452569436,
|
| 1862 |
+
"learning_rate": 8.065794164921128e-05,
|
| 1863 |
+
"loss": 0.9556,
|
| 1864 |
+
"step": 1320
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.61,
|
| 1868 |
+
"grad_norm": 0.08801571017475394,
|
| 1869 |
+
"learning_rate": 7.987267903375332e-05,
|
| 1870 |
+
"loss": 0.9679,
|
| 1871 |
+
"step": 1325
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.61,
|
| 1875 |
+
"grad_norm": 0.08789034709036309,
|
| 1876 |
+
"learning_rate": 7.90887078404033e-05,
|
| 1877 |
+
"loss": 0.9704,
|
| 1878 |
+
"step": 1330
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.61,
|
| 1882 |
+
"grad_norm": 0.08766717856159563,
|
| 1883 |
+
"learning_rate": 7.830607837082493e-05,
|
| 1884 |
+
"loss": 0.9537,
|
| 1885 |
+
"step": 1335
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.61,
|
| 1889 |
+
"grad_norm": 0.08796194617668124,
|
| 1890 |
+
"learning_rate": 7.75248408405934e-05,
|
| 1891 |
+
"loss": 0.9295,
|
| 1892 |
+
"step": 1340
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.62,
|
| 1896 |
+
"grad_norm": 0.08644181932736543,
|
| 1897 |
+
"learning_rate": 7.674504537597336e-05,
|
| 1898 |
+
"loss": 0.9106,
|
| 1899 |
+
"step": 1345
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.62,
|
| 1903 |
+
"grad_norm": 0.08730172492521628,
|
| 1904 |
+
"learning_rate": 7.596674201070282e-05,
|
| 1905 |
+
"loss": 0.9247,
|
| 1906 |
+
"step": 1350
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.62,
|
| 1910 |
+
"grad_norm": 0.08810264581282028,
|
| 1911 |
+
"learning_rate": 7.518998068278266e-05,
|
| 1912 |
+
"loss": 0.941,
|
| 1913 |
+
"step": 1355
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.62,
|
| 1917 |
+
"grad_norm": 0.08798895320349515,
|
| 1918 |
+
"learning_rate": 7.441481123127257e-05,
|
| 1919 |
+
"loss": 0.9242,
|
| 1920 |
+
"step": 1360
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.63,
|
| 1924 |
+
"grad_norm": 0.09819188555964814,
|
| 1925 |
+
"learning_rate": 7.364128339309326e-05,
|
| 1926 |
+
"loss": 0.9216,
|
| 1927 |
+
"step": 1365
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.63,
|
| 1931 |
+
"grad_norm": 0.08880903672413525,
|
| 1932 |
+
"learning_rate": 7.28694467998352e-05,
|
| 1933 |
+
"loss": 0.9386,
|
| 1934 |
+
"step": 1370
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.63,
|
| 1938 |
+
"grad_norm": 0.08594152206617324,
|
| 1939 |
+
"learning_rate": 7.209935097457413e-05,
|
| 1940 |
+
"loss": 0.9365,
|
| 1941 |
+
"step": 1375
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.63,
|
| 1945 |
+
"grad_norm": 0.08716534615610971,
|
| 1946 |
+
"learning_rate": 7.133104532869342e-05,
|
| 1947 |
+
"loss": 0.9823,
|
| 1948 |
+
"step": 1380
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.64,
|
| 1952 |
+
"grad_norm": 0.08772796875372083,
|
| 1953 |
+
"learning_rate": 7.056457915871399e-05,
|
| 1954 |
+
"loss": 0.9526,
|
| 1955 |
+
"step": 1385
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.64,
|
| 1959 |
+
"grad_norm": 0.10720913898902203,
|
| 1960 |
+
"learning_rate": 6.980000164313093e-05,
|
| 1961 |
+
"loss": 0.9445,
|
| 1962 |
+
"step": 1390
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.64,
|
| 1966 |
+
"grad_norm": 0.09018703484669713,
|
| 1967 |
+
"learning_rate": 6.903736183925835e-05,
|
| 1968 |
+
"loss": 0.9479,
|
| 1969 |
+
"step": 1395
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.64,
|
| 1973 |
+
"grad_norm": 0.08733086798681411,
|
| 1974 |
+
"learning_rate": 6.827670868008171e-05,
|
| 1975 |
+
"loss": 0.9194,
|
| 1976 |
+
"step": 1400
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.64,
|
| 1980 |
+
"grad_norm": 0.0863528369966269,
|
| 1981 |
+
"learning_rate": 6.751809097111798e-05,
|
| 1982 |
+
"loss": 0.9483,
|
| 1983 |
+
"step": 1405
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.65,
|
| 1987 |
+
"grad_norm": 0.08631912201267505,
|
| 1988 |
+
"learning_rate": 6.676155738728438e-05,
|
| 1989 |
+
"loss": 0.9455,
|
| 1990 |
+
"step": 1410
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.65,
|
| 1994 |
+
"grad_norm": 0.08609745370001315,
|
| 1995 |
+
"learning_rate": 6.600715646977502e-05,
|
| 1996 |
+
"loss": 0.9413,
|
| 1997 |
+
"step": 1415
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.65,
|
| 2001 |
+
"grad_norm": 0.08629402800348958,
|
| 2002 |
+
"learning_rate": 6.525493662294668e-05,
|
| 2003 |
+
"loss": 0.9609,
|
| 2004 |
+
"step": 1420
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.65,
|
| 2008 |
+
"grad_norm": 0.08751936032063028,
|
| 2009 |
+
"learning_rate": 6.450494611121273e-05,
|
| 2010 |
+
"loss": 0.914,
|
| 2011 |
+
"step": 1425
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.66,
|
| 2015 |
+
"grad_norm": 0.08745706909629332,
|
| 2016 |
+
"learning_rate": 6.375723305594658e-05,
|
| 2017 |
+
"loss": 0.9217,
|
| 2018 |
+
"step": 1430
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.66,
|
| 2022 |
+
"grad_norm": 0.08641611663046964,
|
| 2023 |
+
"learning_rate": 6.301184543239398e-05,
|
| 2024 |
+
"loss": 0.9189,
|
| 2025 |
+
"step": 1435
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.66,
|
| 2029 |
+
"grad_norm": 0.08722147871097277,
|
| 2030 |
+
"learning_rate": 6.226883106659485e-05,
|
| 2031 |
+
"loss": 0.9371,
|
| 2032 |
+
"step": 1440
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.66,
|
| 2036 |
+
"grad_norm": 0.0899356065834259,
|
| 2037 |
+
"learning_rate": 6.152823763231463e-05,
|
| 2038 |
+
"loss": 0.9558,
|
| 2039 |
+
"step": 1445
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.67,
|
| 2043 |
+
"grad_norm": 0.08955527969440581,
|
| 2044 |
+
"learning_rate": 6.079011264798534e-05,
|
| 2045 |
+
"loss": 0.9291,
|
| 2046 |
+
"step": 1450
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.67,
|
| 2050 |
+
"grad_norm": 0.08576837808957845,
|
| 2051 |
+
"learning_rate": 6.005450347365687e-05,
|
| 2052 |
+
"loss": 0.9274,
|
| 2053 |
+
"step": 1455
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.67,
|
| 2057 |
+
"grad_norm": 0.08645095620943792,
|
| 2058 |
+
"learning_rate": 5.932145730795793e-05,
|
| 2059 |
+
"loss": 0.9598,
|
| 2060 |
+
"step": 1460
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.67,
|
| 2064 |
+
"grad_norm": 0.0869911444874417,
|
| 2065 |
+
"learning_rate": 5.8591021185067876e-05,
|
| 2066 |
+
"loss": 0.9363,
|
| 2067 |
+
"step": 1465
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.67,
|
| 2071 |
+
"grad_norm": 0.0873524141173332,
|
| 2072 |
+
"learning_rate": 5.786324197169887e-05,
|
| 2073 |
+
"loss": 0.9422,
|
| 2074 |
+
"step": 1470
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.68,
|
| 2078 |
+
"grad_norm": 0.08911905819028881,
|
| 2079 |
+
"learning_rate": 5.7138166364088705e-05,
|
| 2080 |
+
"loss": 0.9087,
|
| 2081 |
+
"step": 1475
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.68,
|
| 2085 |
+
"grad_norm": 0.08693990530615593,
|
| 2086 |
+
"learning_rate": 5.641584088500461e-05,
|
| 2087 |
+
"loss": 0.9429,
|
| 2088 |
+
"step": 1480
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.68,
|
| 2092 |
+
"grad_norm": 0.08924602848888696,
|
| 2093 |
+
"learning_rate": 5.569631188075841e-05,
|
| 2094 |
+
"loss": 0.9371,
|
| 2095 |
+
"step": 1485
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.68,
|
| 2099 |
+
"grad_norm": 0.08914866066749584,
|
| 2100 |
+
"learning_rate": 5.497962551823266e-05,
|
| 2101 |
+
"loss": 0.9531,
|
| 2102 |
+
"step": 1490
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.69,
|
| 2106 |
+
"grad_norm": 0.08529396225999064,
|
| 2107 |
+
"learning_rate": 5.4265827781918576e-05,
|
| 2108 |
+
"loss": 0.9566,
|
| 2109 |
+
"step": 1495
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.69,
|
| 2113 |
+
"grad_norm": 0.08823829670606588,
|
| 2114 |
+
"learning_rate": 5.355496447096533e-05,
|
| 2115 |
+
"loss": 0.9364,
|
| 2116 |
+
"step": 1500
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.69,
|
| 2120 |
+
"grad_norm": 0.08772222657507125,
|
| 2121 |
+
"learning_rate": 5.284708119624173e-05,
|
| 2122 |
+
"loss": 0.9432,
|
| 2123 |
+
"step": 1505
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.69,
|
| 2127 |
+
"grad_norm": 0.08557062679793,
|
| 2128 |
+
"learning_rate": 5.214222337740962e-05,
|
| 2129 |
+
"loss": 0.9262,
|
| 2130 |
+
"step": 1510
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.7,
|
| 2134 |
+
"grad_norm": 0.08659945848240291,
|
| 2135 |
+
"learning_rate": 5.144043624000944e-05,
|
| 2136 |
+
"loss": 0.9345,
|
| 2137 |
+
"step": 1515
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.7,
|
| 2141 |
+
"grad_norm": 0.08848309633261352,
|
| 2142 |
+
"learning_rate": 5.0741764812558724e-05,
|
| 2143 |
+
"loss": 0.9383,
|
| 2144 |
+
"step": 1520
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.7,
|
| 2148 |
+
"grad_norm": 0.08846095794944839,
|
| 2149 |
+
"learning_rate": 5.00462539236628e-05,
|
| 2150 |
+
"loss": 0.9187,
|
| 2151 |
+
"step": 1525
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.7,
|
| 2155 |
+
"grad_norm": 0.08813900803993555,
|
| 2156 |
+
"learning_rate": 4.935394819913849e-05,
|
| 2157 |
+
"loss": 0.9312,
|
| 2158 |
+
"step": 1530
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.7,
|
| 2162 |
+
"grad_norm": 0.09209963741468702,
|
| 2163 |
+
"learning_rate": 4.8664892059150723e-05,
|
| 2164 |
+
"loss": 0.9598,
|
| 2165 |
+
"step": 1535
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.71,
|
| 2169 |
+
"grad_norm": 0.08771530518180069,
|
| 2170 |
+
"learning_rate": 4.7979129715362626e-05,
|
| 2171 |
+
"loss": 0.9474,
|
| 2172 |
+
"step": 1540
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.71,
|
| 2176 |
+
"grad_norm": 0.08792286254368877,
|
| 2177 |
+
"learning_rate": 4.72967051680985e-05,
|
| 2178 |
+
"loss": 0.957,
|
| 2179 |
+
"step": 1545
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.71,
|
| 2183 |
+
"grad_norm": 0.08941277401969798,
|
| 2184 |
+
"learning_rate": 4.661766220352097e-05,
|
| 2185 |
+
"loss": 0.9378,
|
| 2186 |
+
"step": 1550
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.71,
|
| 2190 |
+
"grad_norm": 0.09121672960781078,
|
| 2191 |
+
"learning_rate": 4.5942044390821214e-05,
|
| 2192 |
+
"loss": 0.9483,
|
| 2193 |
+
"step": 1555
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.72,
|
| 2197 |
+
"grad_norm": 0.08879176023561161,
|
| 2198 |
+
"learning_rate": 4.526989507942374e-05,
|
| 2199 |
+
"loss": 0.931,
|
| 2200 |
+
"step": 1560
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.72,
|
| 2204 |
+
"grad_norm": 0.0883812911577774,
|
| 2205 |
+
"learning_rate": 4.460125739620479e-05,
|
| 2206 |
+
"loss": 0.9406,
|
| 2207 |
+
"step": 1565
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.72,
|
| 2211 |
+
"grad_norm": 0.09091739732819643,
|
| 2212 |
+
"learning_rate": 4.393617424272527e-05,
|
| 2213 |
+
"loss": 0.943,
|
| 2214 |
+
"step": 1570
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.72,
|
| 2218 |
+
"grad_norm": 0.08712923953000383,
|
| 2219 |
+
"learning_rate": 4.3274688292478106e-05,
|
| 2220 |
+
"loss": 0.924,
|
| 2221 |
+
"step": 1575
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.73,
|
| 2225 |
+
"grad_norm": 0.08764495954523147,
|
| 2226 |
+
"learning_rate": 4.261684198815003e-05,
|
| 2227 |
+
"loss": 0.9421,
|
| 2228 |
+
"step": 1580
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.73,
|
| 2232 |
+
"grad_norm": 0.08942249793670395,
|
| 2233 |
+
"learning_rate": 4.1962677538898645e-05,
|
| 2234 |
+
"loss": 0.9285,
|
| 2235 |
+
"step": 1585
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.73,
|
| 2239 |
+
"grad_norm": 0.0912127042783993,
|
| 2240 |
+
"learning_rate": 4.131223691764383e-05,
|
| 2241 |
+
"loss": 0.9279,
|
| 2242 |
+
"step": 1590
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.73,
|
| 2246 |
+
"grad_norm": 0.08585042904329478,
|
| 2247 |
+
"learning_rate": 4.0665561858374934e-05,
|
| 2248 |
+
"loss": 0.9624,
|
| 2249 |
+
"step": 1595
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.73,
|
| 2253 |
+
"grad_norm": 0.0878958377245175,
|
| 2254 |
+
"learning_rate": 4.0022693853472884e-05,
|
| 2255 |
+
"loss": 0.9198,
|
| 2256 |
+
"step": 1600
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.74,
|
| 2260 |
+
"grad_norm": 0.09047554656699612,
|
| 2261 |
+
"learning_rate": 3.938367415104793e-05,
|
| 2262 |
+
"loss": 0.9482,
|
| 2263 |
+
"step": 1605
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.74,
|
| 2267 |
+
"grad_norm": 0.08814416840573022,
|
| 2268 |
+
"learning_rate": 3.8748543752293e-05,
|
| 2269 |
+
"loss": 0.9356,
|
| 2270 |
+
"step": 1610
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.74,
|
| 2274 |
+
"grad_norm": 0.08723692471240356,
|
| 2275 |
+
"learning_rate": 3.8117343408853127e-05,
|
| 2276 |
+
"loss": 0.9278,
|
| 2277 |
+
"step": 1615
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.74,
|
| 2281 |
+
"grad_norm": 0.08717485150452346,
|
| 2282 |
+
"learning_rate": 3.7490113620210485e-05,
|
| 2283 |
+
"loss": 0.9524,
|
| 2284 |
+
"step": 1620
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.75,
|
| 2288 |
+
"grad_norm": 0.08692486607533598,
|
| 2289 |
+
"learning_rate": 3.686689463108608e-05,
|
| 2290 |
+
"loss": 0.9643,
|
| 2291 |
+
"step": 1625
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.75,
|
| 2295 |
+
"grad_norm": 0.0884000414734674,
|
| 2296 |
+
"learning_rate": 3.624772642885734e-05,
|
| 2297 |
+
"loss": 0.9325,
|
| 2298 |
+
"step": 1630
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.75,
|
| 2302 |
+
"grad_norm": 0.08869428008386021,
|
| 2303 |
+
"learning_rate": 3.563264874099258e-05,
|
| 2304 |
+
"loss": 0.9287,
|
| 2305 |
+
"step": 1635
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.75,
|
| 2309 |
+
"grad_norm": 0.08820910205821329,
|
| 2310 |
+
"learning_rate": 3.502170103250177e-05,
|
| 2311 |
+
"loss": 0.9302,
|
| 2312 |
+
"step": 1640
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.75,
|
| 2316 |
+
"grad_norm": 0.08861971632371925,
|
| 2317 |
+
"learning_rate": 3.441492250340461e-05,
|
| 2318 |
+
"loss": 0.9585,
|
| 2319 |
+
"step": 1645
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.76,
|
| 2323 |
+
"grad_norm": 0.0894936060024211,
|
| 2324 |
+
"learning_rate": 3.381235208621522e-05,
|
| 2325 |
+
"loss": 0.9308,
|
| 2326 |
+
"step": 1650
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.76,
|
| 2330 |
+
"grad_norm": 0.08873703732670399,
|
| 2331 |
+
"learning_rate": 3.3214028443444036e-05,
|
| 2332 |
+
"loss": 0.9566,
|
| 2333 |
+
"step": 1655
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.76,
|
| 2337 |
+
"grad_norm": 0.08737818209354478,
|
| 2338 |
+
"learning_rate": 3.2619989965117356e-05,
|
| 2339 |
+
"loss": 0.9498,
|
| 2340 |
+
"step": 1660
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.76,
|
| 2344 |
+
"grad_norm": 0.08766867092475095,
|
| 2345 |
+
"learning_rate": 3.2030274766313865e-05,
|
| 2346 |
+
"loss": 0.967,
|
| 2347 |
+
"step": 1665
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.77,
|
| 2351 |
+
"grad_norm": 0.08557302682938876,
|
| 2352 |
+
"learning_rate": 3.1444920684719394e-05,
|
| 2353 |
+
"loss": 0.9263,
|
| 2354 |
+
"step": 1670
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.77,
|
| 2358 |
+
"grad_norm": 0.08700645088876224,
|
| 2359 |
+
"learning_rate": 3.086396527819876e-05,
|
| 2360 |
+
"loss": 0.9468,
|
| 2361 |
+
"step": 1675
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.77,
|
| 2365 |
+
"grad_norm": 0.08747938373180429,
|
| 2366 |
+
"learning_rate": 3.028744582238633e-05,
|
| 2367 |
+
"loss": 0.9313,
|
| 2368 |
+
"step": 1680
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.77,
|
| 2372 |
+
"grad_norm": 0.08729419364535573,
|
| 2373 |
+
"learning_rate": 2.9715399308294e-05,
|
| 2374 |
+
"loss": 0.9397,
|
| 2375 |
+
"step": 1685
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.78,
|
| 2379 |
+
"grad_norm": 0.08469658948344076,
|
| 2380 |
+
"learning_rate": 2.914786243993808e-05,
|
| 2381 |
+
"loss": 0.9289,
|
| 2382 |
+
"step": 1690
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.78,
|
| 2386 |
+
"grad_norm": 0.08761596719724722,
|
| 2387 |
+
"learning_rate": 2.8584871631983888e-05,
|
| 2388 |
+
"loss": 0.9445,
|
| 2389 |
+
"step": 1695
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.78,
|
| 2393 |
+
"grad_norm": 0.08867345772765582,
|
| 2394 |
+
"learning_rate": 2.8026463007409664e-05,
|
| 2395 |
+
"loss": 0.9411,
|
| 2396 |
+
"step": 1700
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.78,
|
| 2400 |
+
"grad_norm": 0.08862291602018178,
|
| 2401 |
+
"learning_rate": 2.747267239518857e-05,
|
| 2402 |
+
"loss": 0.9211,
|
| 2403 |
+
"step": 1705
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.78,
|
| 2407 |
+
"grad_norm": 0.0880959695459882,
|
| 2408 |
+
"learning_rate": 2.6923535327989924e-05,
|
| 2409 |
+
"loss": 0.9366,
|
| 2410 |
+
"step": 1710
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.79,
|
| 2414 |
+
"grad_norm": 0.08567801939104099,
|
| 2415 |
+
"learning_rate": 2.637908703989924e-05,
|
| 2416 |
+
"loss": 0.923,
|
| 2417 |
+
"step": 1715
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.79,
|
| 2421 |
+
"grad_norm": 0.08783102422275266,
|
| 2422 |
+
"learning_rate": 2.5839362464157634e-05,
|
| 2423 |
+
"loss": 0.9446,
|
| 2424 |
+
"step": 1720
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.79,
|
| 2428 |
+
"grad_norm": 0.08738566610637465,
|
| 2429 |
+
"learning_rate": 2.5304396230920345e-05,
|
| 2430 |
+
"loss": 0.9494,
|
| 2431 |
+
"step": 1725
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.79,
|
| 2435 |
+
"grad_norm": 0.08754716332558513,
|
| 2436 |
+
"learning_rate": 2.477422266503473e-05,
|
| 2437 |
+
"loss": 0.9261,
|
| 2438 |
+
"step": 1730
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.8,
|
| 2442 |
+
"grad_norm": 0.08938747413217578,
|
| 2443 |
+
"learning_rate": 2.4248875783837987e-05,
|
| 2444 |
+
"loss": 0.9522,
|
| 2445 |
+
"step": 1735
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.8,
|
| 2449 |
+
"grad_norm": 0.09156561613392297,
|
| 2450 |
+
"learning_rate": 2.3728389294974475e-05,
|
| 2451 |
+
"loss": 0.949,
|
| 2452 |
+
"step": 1740
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.8,
|
| 2456 |
+
"grad_norm": 0.09048390614208174,
|
| 2457 |
+
"learning_rate": 2.3212796594232943e-05,
|
| 2458 |
+
"loss": 0.9302,
|
| 2459 |
+
"step": 1745
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.8,
|
| 2463 |
+
"grad_norm": 0.08934599478933157,
|
| 2464 |
+
"learning_rate": 2.2702130763403673e-05,
|
| 2465 |
+
"loss": 0.9688,
|
| 2466 |
+
"step": 1750
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.81,
|
| 2470 |
+
"grad_norm": 0.08845439903331347,
|
| 2471 |
+
"learning_rate": 2.2196424568156073e-05,
|
| 2472 |
+
"loss": 0.9303,
|
| 2473 |
+
"step": 1755
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.81,
|
| 2477 |
+
"grad_norm": 0.08832857380862127,
|
| 2478 |
+
"learning_rate": 2.1695710455936115e-05,
|
| 2479 |
+
"loss": 0.9416,
|
| 2480 |
+
"step": 1760
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.81,
|
| 2484 |
+
"grad_norm": 0.08892222595846298,
|
| 2485 |
+
"learning_rate": 2.1200020553884604e-05,
|
| 2486 |
+
"loss": 0.9518,
|
| 2487 |
+
"step": 1765
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.81,
|
| 2491 |
+
"grad_norm": 0.08649233810401519,
|
| 2492 |
+
"learning_rate": 2.070938666677573e-05,
|
| 2493 |
+
"loss": 0.942,
|
| 2494 |
+
"step": 1770
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.81,
|
| 2498 |
+
"grad_norm": 0.0900743762289211,
|
| 2499 |
+
"learning_rate": 2.0223840274976413e-05,
|
| 2500 |
+
"loss": 0.9576,
|
| 2501 |
+
"step": 1775
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.82,
|
| 2505 |
+
"grad_norm": 0.08695584735679976,
|
| 2506 |
+
"learning_rate": 1.974341253242635e-05,
|
| 2507 |
+
"loss": 0.9542,
|
| 2508 |
+
"step": 1780
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.82,
|
| 2512 |
+
"grad_norm": 0.08744798511286825,
|
| 2513 |
+
"learning_rate": 1.9268134264639272e-05,
|
| 2514 |
+
"loss": 0.959,
|
| 2515 |
+
"step": 1785
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.82,
|
| 2519 |
+
"grad_norm": 0.08747274450074151,
|
| 2520 |
+
"learning_rate": 1.879803596672497e-05,
|
| 2521 |
+
"loss": 0.9173,
|
| 2522 |
+
"step": 1790
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.82,
|
| 2526 |
+
"grad_norm": 0.0918571830857398,
|
| 2527 |
+
"learning_rate": 1.8333147801432614e-05,
|
| 2528 |
+
"loss": 0.9476,
|
| 2529 |
+
"step": 1795
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.83,
|
| 2533 |
+
"grad_norm": 0.08982466572383123,
|
| 2534 |
+
"learning_rate": 1.7873499597215604e-05,
|
| 2535 |
+
"loss": 0.9323,
|
| 2536 |
+
"step": 1800
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.83,
|
| 2540 |
+
"grad_norm": 0.08759844398854984,
|
| 2541 |
+
"learning_rate": 1.741912084631746e-05,
|
| 2542 |
+
"loss": 0.97,
|
| 2543 |
+
"step": 1805
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.83,
|
| 2547 |
+
"grad_norm": 0.09055340730791506,
|
| 2548 |
+
"learning_rate": 1.697004070287982e-05,
|
| 2549 |
+
"loss": 0.939,
|
| 2550 |
+
"step": 1810
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.83,
|
| 2554 |
+
"grad_norm": 0.09056007317649489,
|
| 2555 |
+
"learning_rate": 1.6526287981071475e-05,
|
| 2556 |
+
"loss": 0.9629,
|
| 2557 |
+
"step": 1815
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.84,
|
| 2561 |
+
"grad_norm": 0.08589739963435682,
|
| 2562 |
+
"learning_rate": 1.608789115323993e-05,
|
| 2563 |
+
"loss": 0.9451,
|
| 2564 |
+
"step": 1820
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.84,
|
| 2568 |
+
"grad_norm": 0.08640941835070298,
|
| 2569 |
+
"learning_rate": 1.5654878348084244e-05,
|
| 2570 |
+
"loss": 0.9197,
|
| 2571 |
+
"step": 1825
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.84,
|
| 2575 |
+
"grad_norm": 0.08910530693806586,
|
| 2576 |
+
"learning_rate": 1.5227277348850466e-05,
|
| 2577 |
+
"loss": 0.9494,
|
| 2578 |
+
"step": 1830
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.84,
|
| 2582 |
+
"grad_norm": 0.08850326334733846,
|
| 2583 |
+
"learning_rate": 1.4805115591548746e-05,
|
| 2584 |
+
"loss": 0.9228,
|
| 2585 |
+
"step": 1835
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.84,
|
| 2589 |
+
"grad_norm": 0.08689898885121261,
|
| 2590 |
+
"learning_rate": 1.4388420163193217e-05,
|
| 2591 |
+
"loss": 0.9432,
|
| 2592 |
+
"step": 1840
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.85,
|
| 2596 |
+
"grad_norm": 0.08505242590609849,
|
| 2597 |
+
"learning_rate": 1.3977217800063846e-05,
|
| 2598 |
+
"loss": 0.9347,
|
| 2599 |
+
"step": 1845
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.85,
|
| 2603 |
+
"grad_norm": 0.09147654680555943,
|
| 2604 |
+
"learning_rate": 1.3571534885991044e-05,
|
| 2605 |
+
"loss": 0.9308,
|
| 2606 |
+
"step": 1850
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.85,
|
| 2610 |
+
"grad_norm": 0.09022868476385788,
|
| 2611 |
+
"learning_rate": 1.3171397450662715e-05,
|
| 2612 |
+
"loss": 0.948,
|
| 2613 |
+
"step": 1855
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.85,
|
| 2617 |
+
"grad_norm": 0.08685234829482098,
|
| 2618 |
+
"learning_rate": 1.277683116795425e-05,
|
| 2619 |
+
"loss": 0.92,
|
| 2620 |
+
"step": 1860
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.86,
|
| 2624 |
+
"grad_norm": 0.08814308202101184,
|
| 2625 |
+
"learning_rate": 1.2387861354281194e-05,
|
| 2626 |
+
"loss": 0.9362,
|
| 2627 |
+
"step": 1865
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.86,
|
| 2631 |
+
"grad_norm": 0.0851637412899579,
|
| 2632 |
+
"learning_rate": 1.2004512966974745e-05,
|
| 2633 |
+
"loss": 0.9366,
|
| 2634 |
+
"step": 1870
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.86,
|
| 2638 |
+
"grad_norm": 0.08894109019360302,
|
| 2639 |
+
"learning_rate": 1.162681060268065e-05,
|
| 2640 |
+
"loss": 0.9439,
|
| 2641 |
+
"step": 1875
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.86,
|
| 2645 |
+
"grad_norm": 0.08621574297233676,
|
| 2646 |
+
"learning_rate": 1.1254778495780748e-05,
|
| 2647 |
+
"loss": 0.917,
|
| 2648 |
+
"step": 1880
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.87,
|
| 2652 |
+
"grad_norm": 0.09245233519670913,
|
| 2653 |
+
"learning_rate": 1.0888440516838373e-05,
|
| 2654 |
+
"loss": 0.9506,
|
| 2655 |
+
"step": 1885
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.87,
|
| 2659 |
+
"grad_norm": 0.09044369763554148,
|
| 2660 |
+
"learning_rate": 1.0527820171066372e-05,
|
| 2661 |
+
"loss": 0.9303,
|
| 2662 |
+
"step": 1890
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.87,
|
| 2666 |
+
"grad_norm": 0.09008186958818373,
|
| 2667 |
+
"learning_rate": 1.0172940596819259e-05,
|
| 2668 |
+
"loss": 0.9411,
|
| 2669 |
+
"step": 1895
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.87,
|
| 2673 |
+
"grad_norm": 0.0904602032108916,
|
| 2674 |
+
"learning_rate": 9.823824564108408e-06,
|
| 2675 |
+
"loss": 0.94,
|
| 2676 |
+
"step": 1900
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.87,
|
| 2680 |
+
"grad_norm": 0.08715850828856053,
|
| 2681 |
+
"learning_rate": 9.480494473141188e-06,
|
| 2682 |
+
"loss": 0.9474,
|
| 2683 |
+
"step": 1905
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.88,
|
| 2687 |
+
"grad_norm": 0.08874174744461717,
|
| 2688 |
+
"learning_rate": 9.142972352883594e-06,
|
| 2689 |
+
"loss": 0.9604,
|
| 2690 |
+
"step": 1910
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.88,
|
| 2694 |
+
"grad_norm": 0.08626507990923932,
|
| 2695 |
+
"learning_rate": 8.811279859646915e-06,
|
| 2696 |
+
"loss": 0.9371,
|
| 2697 |
+
"step": 1915
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.88,
|
| 2701 |
+
"grad_norm": 0.08536401487047512,
|
| 2702 |
+
"learning_rate": 8.485438275698154e-06,
|
| 2703 |
+
"loss": 0.9478,
|
| 2704 |
+
"step": 1920
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.88,
|
| 2708 |
+
"grad_norm": 0.08762760613175481,
|
| 2709 |
+
"learning_rate": 8.165468507894514e-06,
|
| 2710 |
+
"loss": 0.9619,
|
| 2711 |
+
"step": 1925
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.89,
|
| 2715 |
+
"grad_norm": 0.08649803369527974,
|
| 2716 |
+
"learning_rate": 7.851391086341952e-06,
|
| 2717 |
+
"loss": 0.9213,
|
| 2718 |
+
"step": 1930
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.89,
|
| 2722 |
+
"grad_norm": 0.0863185912090572,
|
| 2723 |
+
"learning_rate": 7.543226163077899e-06,
|
| 2724 |
+
"loss": 0.9398,
|
| 2725 |
+
"step": 1935
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.89,
|
| 2729 |
+
"grad_norm": 0.08832161267172574,
|
| 2730 |
+
"learning_rate": 7.240993510778305e-06,
|
| 2731 |
+
"loss": 0.9355,
|
| 2732 |
+
"step": 1940
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.89,
|
| 2736 |
+
"grad_norm": 0.08872195570007042,
|
| 2737 |
+
"learning_rate": 6.9447125214888835e-06,
|
| 2738 |
+
"loss": 0.9484,
|
| 2739 |
+
"step": 1945
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.89,
|
| 2743 |
+
"grad_norm": 0.08537888004129789,
|
| 2744 |
+
"learning_rate": 6.65440220538096e-06,
|
| 2745 |
+
"loss": 0.937,
|
| 2746 |
+
"step": 1950
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.9,
|
| 2750 |
+
"grad_norm": 0.08646668763284765,
|
| 2751 |
+
"learning_rate": 6.370081189531707e-06,
|
| 2752 |
+
"loss": 0.9266,
|
| 2753 |
+
"step": 1955
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.9,
|
| 2757 |
+
"grad_norm": 0.09151612686918877,
|
| 2758 |
+
"learning_rate": 6.091767716728924e-06,
|
| 2759 |
+
"loss": 0.9529,
|
| 2760 |
+
"step": 1960
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.9,
|
| 2764 |
+
"grad_norm": 0.08625391503125818,
|
| 2765 |
+
"learning_rate": 5.819479644300563e-06,
|
| 2766 |
+
"loss": 0.9623,
|
| 2767 |
+
"step": 1965
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.9,
|
| 2771 |
+
"grad_norm": 0.09219046541541549,
|
| 2772 |
+
"learning_rate": 5.553234442969013e-06,
|
| 2773 |
+
"loss": 0.9421,
|
| 2774 |
+
"step": 1970
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.91,
|
| 2778 |
+
"grad_norm": 0.08774543131766023,
|
| 2779 |
+
"learning_rate": 5.293049195730038e-06,
|
| 2780 |
+
"loss": 0.9419,
|
| 2781 |
+
"step": 1975
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.91,
|
| 2785 |
+
"grad_norm": 0.08654274395119196,
|
| 2786 |
+
"learning_rate": 5.038940596756747e-06,
|
| 2787 |
+
"loss": 0.9575,
|
| 2788 |
+
"step": 1980
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.91,
|
| 2792 |
+
"grad_norm": 0.0876337850241222,
|
| 2793 |
+
"learning_rate": 4.790924950328435e-06,
|
| 2794 |
+
"loss": 0.9391,
|
| 2795 |
+
"step": 1985
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.91,
|
| 2799 |
+
"grad_norm": 0.08639018592656753,
|
| 2800 |
+
"learning_rate": 4.5490181697844916e-06,
|
| 2801 |
+
"loss": 0.9333,
|
| 2802 |
+
"step": 1990
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.92,
|
| 2806 |
+
"grad_norm": 0.0873104746399997,
|
| 2807 |
+
"learning_rate": 4.313235776503244e-06,
|
| 2808 |
+
"loss": 0.9793,
|
| 2809 |
+
"step": 1995
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.92,
|
| 2813 |
+
"grad_norm": 0.09228098479128997,
|
| 2814 |
+
"learning_rate": 4.08359289890623e-06,
|
| 2815 |
+
"loss": 0.9282,
|
| 2816 |
+
"step": 2000
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.92,
|
| 2820 |
+
"grad_norm": 0.08822164948945242,
|
| 2821 |
+
"learning_rate": 3.860104271487397e-06,
|
| 2822 |
+
"loss": 0.9332,
|
| 2823 |
+
"step": 2005
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.92,
|
| 2827 |
+
"grad_norm": 0.08678259126429717,
|
| 2828 |
+
"learning_rate": 3.6427842338677352e-06,
|
| 2829 |
+
"loss": 0.936,
|
| 2830 |
+
"step": 2010
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.92,
|
| 2834 |
+
"grad_norm": 0.0881737292442483,
|
| 2835 |
+
"learning_rate": 3.4316467298752263e-06,
|
| 2836 |
+
"loss": 0.9625,
|
| 2837 |
+
"step": 2015
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.93,
|
| 2841 |
+
"grad_norm": 0.09213889711272016,
|
| 2842 |
+
"learning_rate": 3.226705306650113e-06,
|
| 2843 |
+
"loss": 0.9649,
|
| 2844 |
+
"step": 2020
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.93,
|
| 2848 |
+
"grad_norm": 0.08904039321814257,
|
| 2849 |
+
"learning_rate": 3.0279731137757948e-06,
|
| 2850 |
+
"loss": 0.9177,
|
| 2851 |
+
"step": 2025
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.93,
|
| 2855 |
+
"grad_norm": 0.08671758256460199,
|
| 2856 |
+
"learning_rate": 2.835462902434971e-06,
|
| 2857 |
+
"loss": 0.9362,
|
| 2858 |
+
"step": 2030
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.93,
|
| 2862 |
+
"grad_norm": 0.08617851604946536,
|
| 2863 |
+
"learning_rate": 2.649187024591604e-06,
|
| 2864 |
+
"loss": 0.9233,
|
| 2865 |
+
"step": 2035
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.94,
|
| 2869 |
+
"grad_norm": 0.08788489883371325,
|
| 2870 |
+
"learning_rate": 2.4691574321983215e-06,
|
| 2871 |
+
"loss": 0.9627,
|
| 2872 |
+
"step": 2040
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.94,
|
| 2876 |
+
"grad_norm": 0.08882492486063351,
|
| 2877 |
+
"learning_rate": 2.2953856764295624e-06,
|
| 2878 |
+
"loss": 0.9539,
|
| 2879 |
+
"step": 2045
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.94,
|
| 2883 |
+
"grad_norm": 0.08894914791155306,
|
| 2884 |
+
"learning_rate": 2.1278829069404484e-06,
|
| 2885 |
+
"loss": 0.9547,
|
| 2886 |
+
"step": 2050
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.94,
|
| 2890 |
+
"grad_norm": 0.08942232005808806,
|
| 2891 |
+
"learning_rate": 1.966659871151366e-06,
|
| 2892 |
+
"loss": 0.9457,
|
| 2893 |
+
"step": 2055
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.95,
|
| 2897 |
+
"grad_norm": 0.0878596216919505,
|
| 2898 |
+
"learning_rate": 1.811726913558387e-06,
|
| 2899 |
+
"loss": 0.9401,
|
| 2900 |
+
"step": 2060
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.95,
|
| 2904 |
+
"grad_norm": 0.08624565992036248,
|
| 2905 |
+
"learning_rate": 1.6630939750695519e-06,
|
| 2906 |
+
"loss": 0.9283,
|
| 2907 |
+
"step": 2065
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.95,
|
| 2911 |
+
"grad_norm": 0.08751905040079597,
|
| 2912 |
+
"learning_rate": 1.5207705923670157e-06,
|
| 2913 |
+
"loss": 0.9361,
|
| 2914 |
+
"step": 2070
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.95,
|
| 2918 |
+
"grad_norm": 0.08655512741877175,
|
| 2919 |
+
"learning_rate": 1.3847658972951483e-06,
|
| 2920 |
+
"loss": 0.9351,
|
| 2921 |
+
"step": 2075
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.95,
|
| 2925 |
+
"grad_norm": 0.08957313404563644,
|
| 2926 |
+
"learning_rate": 1.2550886162746467e-06,
|
| 2927 |
+
"loss": 0.952,
|
| 2928 |
+
"step": 2080
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.96,
|
| 2932 |
+
"grad_norm": 0.08660110151814926,
|
| 2933 |
+
"learning_rate": 1.1317470697425837e-06,
|
| 2934 |
+
"loss": 0.952,
|
| 2935 |
+
"step": 2085
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.96,
|
| 2939 |
+
"grad_norm": 0.08972626387226099,
|
| 2940 |
+
"learning_rate": 1.0147491716185676e-06,
|
| 2941 |
+
"loss": 0.9009,
|
| 2942 |
+
"step": 2090
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.96,
|
| 2946 |
+
"grad_norm": 0.08970864022551972,
|
| 2947 |
+
"learning_rate": 9.04102428796949e-07,
|
| 2948 |
+
"loss": 0.9323,
|
| 2949 |
+
"step": 2095
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.96,
|
| 2953 |
+
"grad_norm": 0.08435882797271965,
|
| 2954 |
+
"learning_rate": 7.99813940665195e-07,
|
| 2955 |
+
"loss": 0.9112,
|
| 2956 |
+
"step": 2100
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.97,
|
| 2960 |
+
"grad_norm": 0.09018316452815832,
|
| 2961 |
+
"learning_rate": 7.018903986483083e-07,
|
| 2962 |
+
"loss": 0.9663,
|
| 2963 |
+
"step": 2105
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.97,
|
| 2967 |
+
"grad_norm": 0.08893420874426604,
|
| 2968 |
+
"learning_rate": 6.103380857795604e-07,
|
| 2969 |
+
"loss": 0.9607,
|
| 2970 |
+
"step": 2110
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.97,
|
| 2974 |
+
"grad_norm": 0.08572861346513169,
|
| 2975 |
+
"learning_rate": 5.251628762972916e-07,
|
| 2976 |
+
"loss": 0.9447,
|
| 2977 |
+
"step": 2115
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.97,
|
| 2981 |
+
"grad_norm": 0.08774975368376736,
|
| 2982 |
+
"learning_rate": 4.463702352680787e-07,
|
| 2983 |
+
"loss": 0.9385,
|
| 2984 |
+
"step": 2120
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.98,
|
| 2988 |
+
"grad_norm": 0.08689461932505746,
|
| 2989 |
+
"learning_rate": 3.7396521823600537e-07,
|
| 2990 |
+
"loss": 0.9274,
|
| 2991 |
+
"step": 2125
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.98,
|
| 2995 |
+
"grad_norm": 0.09004858281896364,
|
| 2996 |
+
"learning_rate": 3.079524708983095e-07,
|
| 2997 |
+
"loss": 0.913,
|
| 2998 |
+
"step": 2130
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.98,
|
| 3002 |
+
"grad_norm": 0.08569955277802684,
|
| 3003 |
+
"learning_rate": 2.483362288073443e-07,
|
| 3004 |
+
"loss": 0.9154,
|
| 3005 |
+
"step": 2135
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.98,
|
| 3009 |
+
"grad_norm": 0.0871518933503058,
|
| 3010 |
+
"learning_rate": 1.9512031709874035e-07,
|
| 3011 |
+
"loss": 0.9185,
|
| 3012 |
+
"step": 2140
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.98,
|
| 3016 |
+
"grad_norm": 0.08625979911743939,
|
| 3017 |
+
"learning_rate": 1.4830815024606815e-07,
|
| 3018 |
+
"loss": 0.9531,
|
| 3019 |
+
"step": 2145
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.99,
|
| 3023 |
+
"grad_norm": 0.08834947806834018,
|
| 3024 |
+
"learning_rate": 1.0790273184164701e-07,
|
| 3025 |
+
"loss": 0.9524,
|
| 3026 |
+
"step": 2150
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.99,
|
| 3030 |
+
"grad_norm": 0.09003565411089794,
|
| 3031 |
+
"learning_rate": 7.390665440393241e-08,
|
| 3032 |
+
"loss": 0.9328,
|
| 3033 |
+
"step": 2155
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.99,
|
| 3037 |
+
"grad_norm": 0.08858802189320047,
|
| 3038 |
+
"learning_rate": 4.632209921107134e-08,
|
| 3039 |
+
"loss": 0.9394,
|
| 3040 |
+
"step": 2160
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.99,
|
| 3044 |
+
"grad_norm": 0.08654302372571418,
|
| 3045 |
+
"learning_rate": 2.5150836161058622e-08,
|
| 3046 |
+
"loss": 0.9371,
|
| 3047 |
+
"step": 2165
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 1.0,
|
| 3051 |
+
"grad_norm": 0.08900054318893043,
|
| 3052 |
+
"learning_rate": 1.03942236580723e-08,
|
| 3053 |
+
"loss": 0.9539,
|
| 3054 |
+
"step": 2170
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 1.0,
|
| 3058 |
+
"grad_norm": 0.08775869737859891,
|
| 3059 |
+
"learning_rate": 2.053208525365502e-09,
|
| 3060 |
+
"loss": 0.9452,
|
| 3061 |
+
"step": 2175
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 1.0,
|
| 3065 |
+
"eval_loss": 0.9438490867614746,
|
| 3066 |
+
"eval_runtime": 602.9143,
|
| 3067 |
+
"eval_samples_per_second": 25.594,
|
| 3068 |
+
"eval_steps_per_second": 0.401,
|
| 3069 |
+
"step": 2179
|
| 3070 |
+
},
|
| 3071 |
+
{
|
| 3072 |
+
"epoch": 1.0,
|
| 3073 |
+
"step": 2179,
|
| 3074 |
+
"total_flos": 3.4589037545127936e+16,
|
| 3075 |
+
"train_loss": 0.951526911010322,
|
| 3076 |
+
"train_runtime": 20629.7225,
|
| 3077 |
+
"train_samples_per_second": 6.758,
|
| 3078 |
+
"train_steps_per_second": 0.106
|
| 3079 |
+
}
|
| 3080 |
+
],
|
| 3081 |
+
"logging_steps": 5,
|
| 3082 |
+
"max_steps": 2179,
|
| 3083 |
+
"num_input_tokens_seen": 0,
|
| 3084 |
+
"num_train_epochs": 1,
|
| 3085 |
+
"save_steps": 100,
|
| 3086 |
+
"total_flos": 3.4589037545127936e+16,
|
| 3087 |
+
"train_batch_size": 4,
|
| 3088 |
+
"trial_name": null,
|
| 3089 |
+
"trial_params": null
|
| 3090 |
+
}
|