End of training
Browse files- README.md +2 -1
- all_results.json +12 -0
- eval_results.json +7 -0
- train_results.json +8 -0
- trainer_state.json +3727 -0
- training_eval_loss.png +0 -0
- training_loss.png +0 -0
README.md
CHANGED
|
@@ -4,6 +4,7 @@ license: apache-2.0
|
|
| 4 |
base_model: Qwen/Qwen2.5-7B
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
|
|
|
| 7 |
- generated_from_trainer
|
| 8 |
model-index:
|
| 9 |
- name: oh-dcft-v3.1-claude-3-5-haiku-20241022-qwen
|
|
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
# oh-dcft-v3.1-claude-3-5-haiku-20241022-qwen
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [Qwen/Qwen2.5-7B](https://huggingface.co/Qwen/Qwen2.5-7B) on
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
- Loss: 0.4279
|
| 21 |
|
|
|
|
| 4 |
base_model: Qwen/Qwen2.5-7B
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
| 7 |
+
- full
|
| 8 |
- generated_from_trainer
|
| 9 |
model-index:
|
| 10 |
- name: oh-dcft-v3.1-claude-3-5-haiku-20241022-qwen
|
|
|
|
| 16 |
|
| 17 |
# oh-dcft-v3.1-claude-3-5-haiku-20241022-qwen
|
| 18 |
|
| 19 |
+
This model is a fine-tuned version of [Qwen/Qwen2.5-7B](https://huggingface.co/Qwen/Qwen2.5-7B) on the mlfoundations-dev/oh-dcft-v3.1-claude-3-5-haiku-20241022 dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
- Loss: 0.4279
|
| 22 |
|
all_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_loss": 0.42792943120002747,
|
| 4 |
+
"eval_runtime": 443.4778,
|
| 5 |
+
"eval_samples_per_second": 26.522,
|
| 6 |
+
"eval_steps_per_second": 0.415,
|
| 7 |
+
"total_flos": 2745782658662400.0,
|
| 8 |
+
"train_loss": 0.3825414228548572,
|
| 9 |
+
"train_runtime": 71090.721,
|
| 10 |
+
"train_samples_per_second": 9.431,
|
| 11 |
+
"train_steps_per_second": 0.074
|
| 12 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_loss": 0.42792943120002747,
|
| 4 |
+
"eval_runtime": 443.4778,
|
| 5 |
+
"eval_samples_per_second": 26.522,
|
| 6 |
+
"eval_steps_per_second": 0.415
|
| 7 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 2745782658662400.0,
|
| 4 |
+
"train_loss": 0.3825414228548572,
|
| 5 |
+
"train_runtime": 71090.721,
|
| 6 |
+
"train_samples_per_second": 9.431,
|
| 7 |
+
"train_steps_per_second": 0.074
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3727 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 5238,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0057273768613974796,
|
| 13 |
+
"grad_norm": 1.1893006680700424,
|
| 14 |
+
"learning_rate": 5e-06,
|
| 15 |
+
"loss": 0.7164,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.011454753722794959,
|
| 20 |
+
"grad_norm": 1.04026748119074,
|
| 21 |
+
"learning_rate": 5e-06,
|
| 22 |
+
"loss": 0.6034,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.01718213058419244,
|
| 27 |
+
"grad_norm": 0.947476798322524,
|
| 28 |
+
"learning_rate": 5e-06,
|
| 29 |
+
"loss": 0.5621,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.022909507445589918,
|
| 34 |
+
"grad_norm": 0.8781018637437837,
|
| 35 |
+
"learning_rate": 5e-06,
|
| 36 |
+
"loss": 0.5341,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.0286368843069874,
|
| 41 |
+
"grad_norm": 0.9580875402301294,
|
| 42 |
+
"learning_rate": 5e-06,
|
| 43 |
+
"loss": 0.526,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.03436426116838488,
|
| 48 |
+
"grad_norm": 0.8781914311552192,
|
| 49 |
+
"learning_rate": 5e-06,
|
| 50 |
+
"loss": 0.5078,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.04009163802978236,
|
| 55 |
+
"grad_norm": 0.671566687192882,
|
| 56 |
+
"learning_rate": 5e-06,
|
| 57 |
+
"loss": 0.5056,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.045819014891179836,
|
| 62 |
+
"grad_norm": 0.6015476511437352,
|
| 63 |
+
"learning_rate": 5e-06,
|
| 64 |
+
"loss": 0.5,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.05154639175257732,
|
| 69 |
+
"grad_norm": 0.6233828429107251,
|
| 70 |
+
"learning_rate": 5e-06,
|
| 71 |
+
"loss": 0.5062,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.0572737686139748,
|
| 76 |
+
"grad_norm": 0.5502218388168,
|
| 77 |
+
"learning_rate": 5e-06,
|
| 78 |
+
"loss": 0.4995,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.06300114547537228,
|
| 83 |
+
"grad_norm": 0.5727700188508118,
|
| 84 |
+
"learning_rate": 5e-06,
|
| 85 |
+
"loss": 0.5018,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.06872852233676977,
|
| 90 |
+
"grad_norm": 0.5825399902463676,
|
| 91 |
+
"learning_rate": 5e-06,
|
| 92 |
+
"loss": 0.4897,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.07445589919816724,
|
| 97 |
+
"grad_norm": 0.5794095605103436,
|
| 98 |
+
"learning_rate": 5e-06,
|
| 99 |
+
"loss": 0.4921,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.08018327605956473,
|
| 104 |
+
"grad_norm": 0.5905275386374104,
|
| 105 |
+
"learning_rate": 5e-06,
|
| 106 |
+
"loss": 0.4872,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.0859106529209622,
|
| 111 |
+
"grad_norm": 0.5323846340490436,
|
| 112 |
+
"learning_rate": 5e-06,
|
| 113 |
+
"loss": 0.4771,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.09163802978235967,
|
| 118 |
+
"grad_norm": 0.5827507453646251,
|
| 119 |
+
"learning_rate": 5e-06,
|
| 120 |
+
"loss": 0.4923,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.09736540664375716,
|
| 125 |
+
"grad_norm": 0.6052222320099511,
|
| 126 |
+
"learning_rate": 5e-06,
|
| 127 |
+
"loss": 0.4889,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.10309278350515463,
|
| 132 |
+
"grad_norm": 0.5461900186435344,
|
| 133 |
+
"learning_rate": 5e-06,
|
| 134 |
+
"loss": 0.4906,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.10882016036655212,
|
| 139 |
+
"grad_norm": 0.5446312759280234,
|
| 140 |
+
"learning_rate": 5e-06,
|
| 141 |
+
"loss": 0.4792,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.1145475372279496,
|
| 146 |
+
"grad_norm": 0.5282278298362526,
|
| 147 |
+
"learning_rate": 5e-06,
|
| 148 |
+
"loss": 0.4749,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.12027491408934708,
|
| 153 |
+
"grad_norm": 0.5564949237418149,
|
| 154 |
+
"learning_rate": 5e-06,
|
| 155 |
+
"loss": 0.476,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.12600229095074456,
|
| 160 |
+
"grad_norm": 0.5511775080408522,
|
| 161 |
+
"learning_rate": 5e-06,
|
| 162 |
+
"loss": 0.4795,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.13172966781214204,
|
| 167 |
+
"grad_norm": 0.5328672926489202,
|
| 168 |
+
"learning_rate": 5e-06,
|
| 169 |
+
"loss": 0.4768,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.13745704467353953,
|
| 174 |
+
"grad_norm": 0.5278974935096795,
|
| 175 |
+
"learning_rate": 5e-06,
|
| 176 |
+
"loss": 0.471,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.143184421534937,
|
| 181 |
+
"grad_norm": 0.5354440170377812,
|
| 182 |
+
"learning_rate": 5e-06,
|
| 183 |
+
"loss": 0.4713,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.14891179839633448,
|
| 188 |
+
"grad_norm": 0.5438446163693844,
|
| 189 |
+
"learning_rate": 5e-06,
|
| 190 |
+
"loss": 0.4696,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.15463917525773196,
|
| 195 |
+
"grad_norm": 0.5496174748297591,
|
| 196 |
+
"learning_rate": 5e-06,
|
| 197 |
+
"loss": 0.4655,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.16036655211912945,
|
| 202 |
+
"grad_norm": 0.5926866056184342,
|
| 203 |
+
"learning_rate": 5e-06,
|
| 204 |
+
"loss": 0.4747,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.1660939289805269,
|
| 209 |
+
"grad_norm": 0.5141247582508096,
|
| 210 |
+
"learning_rate": 5e-06,
|
| 211 |
+
"loss": 0.4648,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.1718213058419244,
|
| 216 |
+
"grad_norm": 0.5536073278932672,
|
| 217 |
+
"learning_rate": 5e-06,
|
| 218 |
+
"loss": 0.4659,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.1775486827033219,
|
| 223 |
+
"grad_norm": 0.5234564239929281,
|
| 224 |
+
"learning_rate": 5e-06,
|
| 225 |
+
"loss": 0.4581,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.18327605956471935,
|
| 230 |
+
"grad_norm": 0.5219512867012481,
|
| 231 |
+
"learning_rate": 5e-06,
|
| 232 |
+
"loss": 0.4577,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.18900343642611683,
|
| 237 |
+
"grad_norm": 0.5474679242801981,
|
| 238 |
+
"learning_rate": 5e-06,
|
| 239 |
+
"loss": 0.463,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.19473081328751432,
|
| 244 |
+
"grad_norm": 0.516308993975195,
|
| 245 |
+
"learning_rate": 5e-06,
|
| 246 |
+
"loss": 0.4606,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.2004581901489118,
|
| 251 |
+
"grad_norm": 0.5401317271229032,
|
| 252 |
+
"learning_rate": 5e-06,
|
| 253 |
+
"loss": 0.4722,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.20618556701030927,
|
| 258 |
+
"grad_norm": 0.5234223381679171,
|
| 259 |
+
"learning_rate": 5e-06,
|
| 260 |
+
"loss": 0.461,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.21191294387170675,
|
| 265 |
+
"grad_norm": 0.5475015406952914,
|
| 266 |
+
"learning_rate": 5e-06,
|
| 267 |
+
"loss": 0.4527,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.21764032073310424,
|
| 272 |
+
"grad_norm": 0.5376680018928803,
|
| 273 |
+
"learning_rate": 5e-06,
|
| 274 |
+
"loss": 0.4618,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.22336769759450173,
|
| 279 |
+
"grad_norm": 0.5410529332865766,
|
| 280 |
+
"learning_rate": 5e-06,
|
| 281 |
+
"loss": 0.454,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.2290950744558992,
|
| 286 |
+
"grad_norm": 0.5652526379991017,
|
| 287 |
+
"learning_rate": 5e-06,
|
| 288 |
+
"loss": 0.4656,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.23482245131729668,
|
| 293 |
+
"grad_norm": 0.5555688465804705,
|
| 294 |
+
"learning_rate": 5e-06,
|
| 295 |
+
"loss": 0.4594,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.24054982817869416,
|
| 300 |
+
"grad_norm": 0.5701732999281934,
|
| 301 |
+
"learning_rate": 5e-06,
|
| 302 |
+
"loss": 0.4662,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.24627720504009165,
|
| 307 |
+
"grad_norm": 0.5512026939695381,
|
| 308 |
+
"learning_rate": 5e-06,
|
| 309 |
+
"loss": 0.4683,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.2520045819014891,
|
| 314 |
+
"grad_norm": 0.553814123316092,
|
| 315 |
+
"learning_rate": 5e-06,
|
| 316 |
+
"loss": 0.463,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.25773195876288657,
|
| 321 |
+
"grad_norm": 0.5200195587347526,
|
| 322 |
+
"learning_rate": 5e-06,
|
| 323 |
+
"loss": 0.4609,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.2634593356242841,
|
| 328 |
+
"grad_norm": 0.550688588791354,
|
| 329 |
+
"learning_rate": 5e-06,
|
| 330 |
+
"loss": 0.4457,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.26918671248568155,
|
| 335 |
+
"grad_norm": 0.5863844516267137,
|
| 336 |
+
"learning_rate": 5e-06,
|
| 337 |
+
"loss": 0.4632,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.27491408934707906,
|
| 342 |
+
"grad_norm": 0.5729329474663841,
|
| 343 |
+
"learning_rate": 5e-06,
|
| 344 |
+
"loss": 0.4434,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.2806414662084765,
|
| 349 |
+
"grad_norm": 0.5523269014403068,
|
| 350 |
+
"learning_rate": 5e-06,
|
| 351 |
+
"loss": 0.4433,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.286368843069874,
|
| 356 |
+
"grad_norm": 0.5116583829791455,
|
| 357 |
+
"learning_rate": 5e-06,
|
| 358 |
+
"loss": 0.4562,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.2920962199312715,
|
| 363 |
+
"grad_norm": 0.5787963201989073,
|
| 364 |
+
"learning_rate": 5e-06,
|
| 365 |
+
"loss": 0.4654,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.29782359679266895,
|
| 370 |
+
"grad_norm": 0.5447902232207782,
|
| 371 |
+
"learning_rate": 5e-06,
|
| 372 |
+
"loss": 0.4298,
|
| 373 |
+
"step": 520
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.3035509736540664,
|
| 377 |
+
"grad_norm": 0.5423741210811989,
|
| 378 |
+
"learning_rate": 5e-06,
|
| 379 |
+
"loss": 0.463,
|
| 380 |
+
"step": 530
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.30927835051546393,
|
| 384 |
+
"grad_norm": 0.5232608979728346,
|
| 385 |
+
"learning_rate": 5e-06,
|
| 386 |
+
"loss": 0.4534,
|
| 387 |
+
"step": 540
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.3150057273768614,
|
| 391 |
+
"grad_norm": 0.5745290999609551,
|
| 392 |
+
"learning_rate": 5e-06,
|
| 393 |
+
"loss": 0.4514,
|
| 394 |
+
"step": 550
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.3207331042382589,
|
| 398 |
+
"grad_norm": 0.5974645168072564,
|
| 399 |
+
"learning_rate": 5e-06,
|
| 400 |
+
"loss": 0.4606,
|
| 401 |
+
"step": 560
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.32646048109965636,
|
| 405 |
+
"grad_norm": 0.5487734496746278,
|
| 406 |
+
"learning_rate": 5e-06,
|
| 407 |
+
"loss": 0.4607,
|
| 408 |
+
"step": 570
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.3321878579610538,
|
| 412 |
+
"grad_norm": 0.5441562004543612,
|
| 413 |
+
"learning_rate": 5e-06,
|
| 414 |
+
"loss": 0.4507,
|
| 415 |
+
"step": 580
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.33791523482245134,
|
| 419 |
+
"grad_norm": 0.5573672899202607,
|
| 420 |
+
"learning_rate": 5e-06,
|
| 421 |
+
"loss": 0.4441,
|
| 422 |
+
"step": 590
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.3436426116838488,
|
| 426 |
+
"grad_norm": 0.5308370142895599,
|
| 427 |
+
"learning_rate": 5e-06,
|
| 428 |
+
"loss": 0.4518,
|
| 429 |
+
"step": 600
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.34936998854524626,
|
| 433 |
+
"grad_norm": 0.5117603002146386,
|
| 434 |
+
"learning_rate": 5e-06,
|
| 435 |
+
"loss": 0.4583,
|
| 436 |
+
"step": 610
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.3550973654066438,
|
| 440 |
+
"grad_norm": 0.5574954390885013,
|
| 441 |
+
"learning_rate": 5e-06,
|
| 442 |
+
"loss": 0.4603,
|
| 443 |
+
"step": 620
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.36082474226804123,
|
| 447 |
+
"grad_norm": 0.5606001506682383,
|
| 448 |
+
"learning_rate": 5e-06,
|
| 449 |
+
"loss": 0.4514,
|
| 450 |
+
"step": 630
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.3665521191294387,
|
| 454 |
+
"grad_norm": 0.5569042632837081,
|
| 455 |
+
"learning_rate": 5e-06,
|
| 456 |
+
"loss": 0.4537,
|
| 457 |
+
"step": 640
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.3722794959908362,
|
| 461 |
+
"grad_norm": 0.5585555273085546,
|
| 462 |
+
"learning_rate": 5e-06,
|
| 463 |
+
"loss": 0.4626,
|
| 464 |
+
"step": 650
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.37800687285223367,
|
| 468 |
+
"grad_norm": 0.5029368495012907,
|
| 469 |
+
"learning_rate": 5e-06,
|
| 470 |
+
"loss": 0.4438,
|
| 471 |
+
"step": 660
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.3837342497136312,
|
| 475 |
+
"grad_norm": 0.5901403440941212,
|
| 476 |
+
"learning_rate": 5e-06,
|
| 477 |
+
"loss": 0.456,
|
| 478 |
+
"step": 670
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.38946162657502864,
|
| 482 |
+
"grad_norm": 0.5195926126583429,
|
| 483 |
+
"learning_rate": 5e-06,
|
| 484 |
+
"loss": 0.4445,
|
| 485 |
+
"step": 680
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.3951890034364261,
|
| 489 |
+
"grad_norm": 0.5225881930388002,
|
| 490 |
+
"learning_rate": 5e-06,
|
| 491 |
+
"loss": 0.4439,
|
| 492 |
+
"step": 690
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.4009163802978236,
|
| 496 |
+
"grad_norm": 0.5279076911326633,
|
| 497 |
+
"learning_rate": 5e-06,
|
| 498 |
+
"loss": 0.4518,
|
| 499 |
+
"step": 700
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.4066437571592211,
|
| 503 |
+
"grad_norm": 0.5178025129238295,
|
| 504 |
+
"learning_rate": 5e-06,
|
| 505 |
+
"loss": 0.4444,
|
| 506 |
+
"step": 710
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.41237113402061853,
|
| 510 |
+
"grad_norm": 0.5391197058842394,
|
| 511 |
+
"learning_rate": 5e-06,
|
| 512 |
+
"loss": 0.4417,
|
| 513 |
+
"step": 720
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.41809851088201605,
|
| 517 |
+
"grad_norm": 0.5693350216341284,
|
| 518 |
+
"learning_rate": 5e-06,
|
| 519 |
+
"loss": 0.4531,
|
| 520 |
+
"step": 730
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.4238258877434135,
|
| 524 |
+
"grad_norm": 0.5272168668986746,
|
| 525 |
+
"learning_rate": 5e-06,
|
| 526 |
+
"loss": 0.4507,
|
| 527 |
+
"step": 740
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.42955326460481097,
|
| 531 |
+
"grad_norm": 0.5601282100265634,
|
| 532 |
+
"learning_rate": 5e-06,
|
| 533 |
+
"loss": 0.4395,
|
| 534 |
+
"step": 750
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.4352806414662085,
|
| 538 |
+
"grad_norm": 0.5283833355102217,
|
| 539 |
+
"learning_rate": 5e-06,
|
| 540 |
+
"loss": 0.4392,
|
| 541 |
+
"step": 760
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.44100801832760594,
|
| 545 |
+
"grad_norm": 0.5348435308559183,
|
| 546 |
+
"learning_rate": 5e-06,
|
| 547 |
+
"loss": 0.4571,
|
| 548 |
+
"step": 770
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.44673539518900346,
|
| 552 |
+
"grad_norm": 0.5431552168715813,
|
| 553 |
+
"learning_rate": 5e-06,
|
| 554 |
+
"loss": 0.4532,
|
| 555 |
+
"step": 780
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.4524627720504009,
|
| 559 |
+
"grad_norm": 0.545120281271903,
|
| 560 |
+
"learning_rate": 5e-06,
|
| 561 |
+
"loss": 0.4534,
|
| 562 |
+
"step": 790
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.4581901489117984,
|
| 566 |
+
"grad_norm": 0.5387987858480611,
|
| 567 |
+
"learning_rate": 5e-06,
|
| 568 |
+
"loss": 0.4536,
|
| 569 |
+
"step": 800
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.4639175257731959,
|
| 573 |
+
"grad_norm": 0.5430847301042485,
|
| 574 |
+
"learning_rate": 5e-06,
|
| 575 |
+
"loss": 0.4469,
|
| 576 |
+
"step": 810
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.46964490263459335,
|
| 580 |
+
"grad_norm": 0.5311236600330211,
|
| 581 |
+
"learning_rate": 5e-06,
|
| 582 |
+
"loss": 0.4428,
|
| 583 |
+
"step": 820
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.4753722794959908,
|
| 587 |
+
"grad_norm": 0.5560451338696027,
|
| 588 |
+
"learning_rate": 5e-06,
|
| 589 |
+
"loss": 0.4502,
|
| 590 |
+
"step": 830
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.48109965635738833,
|
| 594 |
+
"grad_norm": 0.5470174916746855,
|
| 595 |
+
"learning_rate": 5e-06,
|
| 596 |
+
"loss": 0.4482,
|
| 597 |
+
"step": 840
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.4868270332187858,
|
| 601 |
+
"grad_norm": 0.563263919253685,
|
| 602 |
+
"learning_rate": 5e-06,
|
| 603 |
+
"loss": 0.4446,
|
| 604 |
+
"step": 850
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.4925544100801833,
|
| 608 |
+
"grad_norm": 0.573113515782645,
|
| 609 |
+
"learning_rate": 5e-06,
|
| 610 |
+
"loss": 0.4508,
|
| 611 |
+
"step": 860
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.49828178694158076,
|
| 615 |
+
"grad_norm": 0.5231438549151345,
|
| 616 |
+
"learning_rate": 5e-06,
|
| 617 |
+
"loss": 0.45,
|
| 618 |
+
"step": 870
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.5040091638029782,
|
| 622 |
+
"grad_norm": 0.5176148074624103,
|
| 623 |
+
"learning_rate": 5e-06,
|
| 624 |
+
"loss": 0.4348,
|
| 625 |
+
"step": 880
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.5097365406643757,
|
| 629 |
+
"grad_norm": 0.5309016268362335,
|
| 630 |
+
"learning_rate": 5e-06,
|
| 631 |
+
"loss": 0.443,
|
| 632 |
+
"step": 890
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.5154639175257731,
|
| 636 |
+
"grad_norm": 0.5021579023128154,
|
| 637 |
+
"learning_rate": 5e-06,
|
| 638 |
+
"loss": 0.448,
|
| 639 |
+
"step": 900
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.5211912943871707,
|
| 643 |
+
"grad_norm": 0.559779127609092,
|
| 644 |
+
"learning_rate": 5e-06,
|
| 645 |
+
"loss": 0.444,
|
| 646 |
+
"step": 910
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.5269186712485682,
|
| 650 |
+
"grad_norm": 0.5789927290923392,
|
| 651 |
+
"learning_rate": 5e-06,
|
| 652 |
+
"loss": 0.4493,
|
| 653 |
+
"step": 920
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.5326460481099656,
|
| 657 |
+
"grad_norm": 0.5690148664602116,
|
| 658 |
+
"learning_rate": 5e-06,
|
| 659 |
+
"loss": 0.4541,
|
| 660 |
+
"step": 930
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.5383734249713631,
|
| 664 |
+
"grad_norm": 0.5018200538904565,
|
| 665 |
+
"learning_rate": 5e-06,
|
| 666 |
+
"loss": 0.4437,
|
| 667 |
+
"step": 940
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.5441008018327605,
|
| 671 |
+
"grad_norm": 0.5626094146626068,
|
| 672 |
+
"learning_rate": 5e-06,
|
| 673 |
+
"loss": 0.438,
|
| 674 |
+
"step": 950
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.5498281786941581,
|
| 678 |
+
"grad_norm": 0.5453326758465812,
|
| 679 |
+
"learning_rate": 5e-06,
|
| 680 |
+
"loss": 0.4382,
|
| 681 |
+
"step": 960
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.5555555555555556,
|
| 685 |
+
"grad_norm": 0.5731487204099626,
|
| 686 |
+
"learning_rate": 5e-06,
|
| 687 |
+
"loss": 0.443,
|
| 688 |
+
"step": 970
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.561282932416953,
|
| 692 |
+
"grad_norm": 0.5469415847599124,
|
| 693 |
+
"learning_rate": 5e-06,
|
| 694 |
+
"loss": 0.4541,
|
| 695 |
+
"step": 980
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.5670103092783505,
|
| 699 |
+
"grad_norm": 0.5216017929376741,
|
| 700 |
+
"learning_rate": 5e-06,
|
| 701 |
+
"loss": 0.447,
|
| 702 |
+
"step": 990
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.572737686139748,
|
| 706 |
+
"grad_norm": 0.5633524684965274,
|
| 707 |
+
"learning_rate": 5e-06,
|
| 708 |
+
"loss": 0.4371,
|
| 709 |
+
"step": 1000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.5784650630011455,
|
| 713 |
+
"grad_norm": 0.5293660880153032,
|
| 714 |
+
"learning_rate": 5e-06,
|
| 715 |
+
"loss": 0.4383,
|
| 716 |
+
"step": 1010
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.584192439862543,
|
| 720 |
+
"grad_norm": 0.5399877214818077,
|
| 721 |
+
"learning_rate": 5e-06,
|
| 722 |
+
"loss": 0.456,
|
| 723 |
+
"step": 1020
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.5899198167239404,
|
| 727 |
+
"grad_norm": 0.5249271107308078,
|
| 728 |
+
"learning_rate": 5e-06,
|
| 729 |
+
"loss": 0.459,
|
| 730 |
+
"step": 1030
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.5956471935853379,
|
| 734 |
+
"grad_norm": 0.5341902718265518,
|
| 735 |
+
"learning_rate": 5e-06,
|
| 736 |
+
"loss": 0.4379,
|
| 737 |
+
"step": 1040
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.6013745704467354,
|
| 741 |
+
"grad_norm": 0.5681490785135558,
|
| 742 |
+
"learning_rate": 5e-06,
|
| 743 |
+
"loss": 0.445,
|
| 744 |
+
"step": 1050
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.6071019473081328,
|
| 748 |
+
"grad_norm": 0.5894093830458618,
|
| 749 |
+
"learning_rate": 5e-06,
|
| 750 |
+
"loss": 0.4353,
|
| 751 |
+
"step": 1060
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.6128293241695304,
|
| 755 |
+
"grad_norm": 0.5221738673506199,
|
| 756 |
+
"learning_rate": 5e-06,
|
| 757 |
+
"loss": 0.4412,
|
| 758 |
+
"step": 1070
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.6185567010309279,
|
| 762 |
+
"grad_norm": 0.5797474198144454,
|
| 763 |
+
"learning_rate": 5e-06,
|
| 764 |
+
"loss": 0.443,
|
| 765 |
+
"step": 1080
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.6242840778923253,
|
| 769 |
+
"grad_norm": 0.5717734893270968,
|
| 770 |
+
"learning_rate": 5e-06,
|
| 771 |
+
"loss": 0.4435,
|
| 772 |
+
"step": 1090
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.6300114547537228,
|
| 776 |
+
"grad_norm": 0.4981452424201392,
|
| 777 |
+
"learning_rate": 5e-06,
|
| 778 |
+
"loss": 0.4294,
|
| 779 |
+
"step": 1100
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.6357388316151202,
|
| 783 |
+
"grad_norm": 0.5341189709831258,
|
| 784 |
+
"learning_rate": 5e-06,
|
| 785 |
+
"loss": 0.4443,
|
| 786 |
+
"step": 1110
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.6414662084765178,
|
| 790 |
+
"grad_norm": 0.5337916570686927,
|
| 791 |
+
"learning_rate": 5e-06,
|
| 792 |
+
"loss": 0.4388,
|
| 793 |
+
"step": 1120
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.6471935853379153,
|
| 797 |
+
"grad_norm": 0.5191012616320769,
|
| 798 |
+
"learning_rate": 5e-06,
|
| 799 |
+
"loss": 0.4278,
|
| 800 |
+
"step": 1130
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.6529209621993127,
|
| 804 |
+
"grad_norm": 0.5532648229910325,
|
| 805 |
+
"learning_rate": 5e-06,
|
| 806 |
+
"loss": 0.4445,
|
| 807 |
+
"step": 1140
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.6586483390607102,
|
| 811 |
+
"grad_norm": 0.5376131317705378,
|
| 812 |
+
"learning_rate": 5e-06,
|
| 813 |
+
"loss": 0.4443,
|
| 814 |
+
"step": 1150
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.6643757159221076,
|
| 818 |
+
"grad_norm": 0.5163417459667795,
|
| 819 |
+
"learning_rate": 5e-06,
|
| 820 |
+
"loss": 0.449,
|
| 821 |
+
"step": 1160
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.6701030927835051,
|
| 825 |
+
"grad_norm": 0.5750744468716225,
|
| 826 |
+
"learning_rate": 5e-06,
|
| 827 |
+
"loss": 0.4424,
|
| 828 |
+
"step": 1170
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.6758304696449027,
|
| 832 |
+
"grad_norm": 0.5691426147058902,
|
| 833 |
+
"learning_rate": 5e-06,
|
| 834 |
+
"loss": 0.4409,
|
| 835 |
+
"step": 1180
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.6815578465063001,
|
| 839 |
+
"grad_norm": 0.572169838625891,
|
| 840 |
+
"learning_rate": 5e-06,
|
| 841 |
+
"loss": 0.4408,
|
| 842 |
+
"step": 1190
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.6872852233676976,
|
| 846 |
+
"grad_norm": 0.562085920947936,
|
| 847 |
+
"learning_rate": 5e-06,
|
| 848 |
+
"loss": 0.4539,
|
| 849 |
+
"step": 1200
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.693012600229095,
|
| 853 |
+
"grad_norm": 0.549680800307771,
|
| 854 |
+
"learning_rate": 5e-06,
|
| 855 |
+
"loss": 0.4408,
|
| 856 |
+
"step": 1210
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.6987399770904925,
|
| 860 |
+
"grad_norm": 0.5268961102121411,
|
| 861 |
+
"learning_rate": 5e-06,
|
| 862 |
+
"loss": 0.4318,
|
| 863 |
+
"step": 1220
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.7044673539518901,
|
| 867 |
+
"grad_norm": 0.546084380380316,
|
| 868 |
+
"learning_rate": 5e-06,
|
| 869 |
+
"loss": 0.4274,
|
| 870 |
+
"step": 1230
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.7101947308132875,
|
| 874 |
+
"grad_norm": 0.5789125996714688,
|
| 875 |
+
"learning_rate": 5e-06,
|
| 876 |
+
"loss": 0.426,
|
| 877 |
+
"step": 1240
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.715922107674685,
|
| 881 |
+
"grad_norm": 0.5552812309484606,
|
| 882 |
+
"learning_rate": 5e-06,
|
| 883 |
+
"loss": 0.4433,
|
| 884 |
+
"step": 1250
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.7216494845360825,
|
| 888 |
+
"grad_norm": 0.5490841525640856,
|
| 889 |
+
"learning_rate": 5e-06,
|
| 890 |
+
"loss": 0.4347,
|
| 891 |
+
"step": 1260
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.7273768613974799,
|
| 895 |
+
"grad_norm": 0.5287976784408353,
|
| 896 |
+
"learning_rate": 5e-06,
|
| 897 |
+
"loss": 0.4404,
|
| 898 |
+
"step": 1270
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.7331042382588774,
|
| 902 |
+
"grad_norm": 0.5303023254262377,
|
| 903 |
+
"learning_rate": 5e-06,
|
| 904 |
+
"loss": 0.4195,
|
| 905 |
+
"step": 1280
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.738831615120275,
|
| 909 |
+
"grad_norm": 0.5615661953108583,
|
| 910 |
+
"learning_rate": 5e-06,
|
| 911 |
+
"loss": 0.4399,
|
| 912 |
+
"step": 1290
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.7445589919816724,
|
| 916 |
+
"grad_norm": 0.5392244469382096,
|
| 917 |
+
"learning_rate": 5e-06,
|
| 918 |
+
"loss": 0.4416,
|
| 919 |
+
"step": 1300
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.7502863688430699,
|
| 923 |
+
"grad_norm": 0.5631926544247716,
|
| 924 |
+
"learning_rate": 5e-06,
|
| 925 |
+
"loss": 0.4458,
|
| 926 |
+
"step": 1310
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.7560137457044673,
|
| 930 |
+
"grad_norm": 0.5321981965200037,
|
| 931 |
+
"learning_rate": 5e-06,
|
| 932 |
+
"loss": 0.4425,
|
| 933 |
+
"step": 1320
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.7617411225658648,
|
| 937 |
+
"grad_norm": 0.5472839772157344,
|
| 938 |
+
"learning_rate": 5e-06,
|
| 939 |
+
"loss": 0.438,
|
| 940 |
+
"step": 1330
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.7674684994272624,
|
| 944 |
+
"grad_norm": 0.5148846097662794,
|
| 945 |
+
"learning_rate": 5e-06,
|
| 946 |
+
"loss": 0.441,
|
| 947 |
+
"step": 1340
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.7731958762886598,
|
| 951 |
+
"grad_norm": 0.5814136125437906,
|
| 952 |
+
"learning_rate": 5e-06,
|
| 953 |
+
"loss": 0.4422,
|
| 954 |
+
"step": 1350
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.7789232531500573,
|
| 958 |
+
"grad_norm": 0.4865290153302818,
|
| 959 |
+
"learning_rate": 5e-06,
|
| 960 |
+
"loss": 0.4195,
|
| 961 |
+
"step": 1360
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.7846506300114547,
|
| 965 |
+
"grad_norm": 0.5729362641564091,
|
| 966 |
+
"learning_rate": 5e-06,
|
| 967 |
+
"loss": 0.4252,
|
| 968 |
+
"step": 1370
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.7903780068728522,
|
| 972 |
+
"grad_norm": 0.513721699658904,
|
| 973 |
+
"learning_rate": 5e-06,
|
| 974 |
+
"loss": 0.428,
|
| 975 |
+
"step": 1380
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.7961053837342497,
|
| 979 |
+
"grad_norm": 0.5404966717471676,
|
| 980 |
+
"learning_rate": 5e-06,
|
| 981 |
+
"loss": 0.4321,
|
| 982 |
+
"step": 1390
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.8018327605956472,
|
| 986 |
+
"grad_norm": 0.5705888817719007,
|
| 987 |
+
"learning_rate": 5e-06,
|
| 988 |
+
"loss": 0.4292,
|
| 989 |
+
"step": 1400
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.8075601374570447,
|
| 993 |
+
"grad_norm": 0.5433900974318504,
|
| 994 |
+
"learning_rate": 5e-06,
|
| 995 |
+
"loss": 0.436,
|
| 996 |
+
"step": 1410
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.8132875143184422,
|
| 1000 |
+
"grad_norm": 0.5233288735400002,
|
| 1001 |
+
"learning_rate": 5e-06,
|
| 1002 |
+
"loss": 0.4347,
|
| 1003 |
+
"step": 1420
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.8190148911798396,
|
| 1007 |
+
"grad_norm": 0.5315754725482877,
|
| 1008 |
+
"learning_rate": 5e-06,
|
| 1009 |
+
"loss": 0.4343,
|
| 1010 |
+
"step": 1430
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.8247422680412371,
|
| 1014 |
+
"grad_norm": 0.540898996728165,
|
| 1015 |
+
"learning_rate": 5e-06,
|
| 1016 |
+
"loss": 0.4385,
|
| 1017 |
+
"step": 1440
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.8304696449026346,
|
| 1021 |
+
"grad_norm": 0.4918837151161211,
|
| 1022 |
+
"learning_rate": 5e-06,
|
| 1023 |
+
"loss": 0.4187,
|
| 1024 |
+
"step": 1450
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.8361970217640321,
|
| 1028 |
+
"grad_norm": 0.5167559153442192,
|
| 1029 |
+
"learning_rate": 5e-06,
|
| 1030 |
+
"loss": 0.4354,
|
| 1031 |
+
"step": 1460
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.8419243986254296,
|
| 1035 |
+
"grad_norm": 0.5388428750370872,
|
| 1036 |
+
"learning_rate": 5e-06,
|
| 1037 |
+
"loss": 0.442,
|
| 1038 |
+
"step": 1470
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.847651775486827,
|
| 1042 |
+
"grad_norm": 0.5478670963710769,
|
| 1043 |
+
"learning_rate": 5e-06,
|
| 1044 |
+
"loss": 0.432,
|
| 1045 |
+
"step": 1480
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.8533791523482245,
|
| 1049 |
+
"grad_norm": 0.5551238913425003,
|
| 1050 |
+
"learning_rate": 5e-06,
|
| 1051 |
+
"loss": 0.4349,
|
| 1052 |
+
"step": 1490
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.8591065292096219,
|
| 1056 |
+
"grad_norm": 0.5096809451223056,
|
| 1057 |
+
"learning_rate": 5e-06,
|
| 1058 |
+
"loss": 0.4337,
|
| 1059 |
+
"step": 1500
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.8648339060710195,
|
| 1063 |
+
"grad_norm": 0.5485551648513652,
|
| 1064 |
+
"learning_rate": 5e-06,
|
| 1065 |
+
"loss": 0.4304,
|
| 1066 |
+
"step": 1510
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.870561282932417,
|
| 1070 |
+
"grad_norm": 0.5439471352377651,
|
| 1071 |
+
"learning_rate": 5e-06,
|
| 1072 |
+
"loss": 0.4291,
|
| 1073 |
+
"step": 1520
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.8762886597938144,
|
| 1077 |
+
"grad_norm": 0.5541220552945509,
|
| 1078 |
+
"learning_rate": 5e-06,
|
| 1079 |
+
"loss": 0.435,
|
| 1080 |
+
"step": 1530
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.8820160366552119,
|
| 1084 |
+
"grad_norm": 0.5549824280277013,
|
| 1085 |
+
"learning_rate": 5e-06,
|
| 1086 |
+
"loss": 0.4455,
|
| 1087 |
+
"step": 1540
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.8877434135166093,
|
| 1091 |
+
"grad_norm": 0.5160198263139852,
|
| 1092 |
+
"learning_rate": 5e-06,
|
| 1093 |
+
"loss": 0.4246,
|
| 1094 |
+
"step": 1550
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.8934707903780069,
|
| 1098 |
+
"grad_norm": 0.5789101227791154,
|
| 1099 |
+
"learning_rate": 5e-06,
|
| 1100 |
+
"loss": 0.4459,
|
| 1101 |
+
"step": 1560
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.8991981672394044,
|
| 1105 |
+
"grad_norm": 0.5619060606532335,
|
| 1106 |
+
"learning_rate": 5e-06,
|
| 1107 |
+
"loss": 0.4227,
|
| 1108 |
+
"step": 1570
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.9049255441008018,
|
| 1112 |
+
"grad_norm": 0.5131288466424669,
|
| 1113 |
+
"learning_rate": 5e-06,
|
| 1114 |
+
"loss": 0.425,
|
| 1115 |
+
"step": 1580
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.9106529209621993,
|
| 1119 |
+
"grad_norm": 0.5378774111576103,
|
| 1120 |
+
"learning_rate": 5e-06,
|
| 1121 |
+
"loss": 0.4339,
|
| 1122 |
+
"step": 1590
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.9163802978235968,
|
| 1126 |
+
"grad_norm": 0.5496777521069192,
|
| 1127 |
+
"learning_rate": 5e-06,
|
| 1128 |
+
"loss": 0.4351,
|
| 1129 |
+
"step": 1600
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.9221076746849943,
|
| 1133 |
+
"grad_norm": 0.5090132737788183,
|
| 1134 |
+
"learning_rate": 5e-06,
|
| 1135 |
+
"loss": 0.4439,
|
| 1136 |
+
"step": 1610
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.9278350515463918,
|
| 1140 |
+
"grad_norm": 0.5537237167714347,
|
| 1141 |
+
"learning_rate": 5e-06,
|
| 1142 |
+
"loss": 0.4315,
|
| 1143 |
+
"step": 1620
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.9335624284077892,
|
| 1147 |
+
"grad_norm": 0.5083065954607388,
|
| 1148 |
+
"learning_rate": 5e-06,
|
| 1149 |
+
"loss": 0.4209,
|
| 1150 |
+
"step": 1630
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.9392898052691867,
|
| 1154 |
+
"grad_norm": 0.5504634632628532,
|
| 1155 |
+
"learning_rate": 5e-06,
|
| 1156 |
+
"loss": 0.4255,
|
| 1157 |
+
"step": 1640
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.9450171821305842,
|
| 1161 |
+
"grad_norm": 0.497466661980522,
|
| 1162 |
+
"learning_rate": 5e-06,
|
| 1163 |
+
"loss": 0.4359,
|
| 1164 |
+
"step": 1650
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.9507445589919816,
|
| 1168 |
+
"grad_norm": 0.5728710662746097,
|
| 1169 |
+
"learning_rate": 5e-06,
|
| 1170 |
+
"loss": 0.4298,
|
| 1171 |
+
"step": 1660
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.9564719358533792,
|
| 1175 |
+
"grad_norm": 0.5382940747387163,
|
| 1176 |
+
"learning_rate": 5e-06,
|
| 1177 |
+
"loss": 0.4321,
|
| 1178 |
+
"step": 1670
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.9621993127147767,
|
| 1182 |
+
"grad_norm": 0.5602645110237611,
|
| 1183 |
+
"learning_rate": 5e-06,
|
| 1184 |
+
"loss": 0.4376,
|
| 1185 |
+
"step": 1680
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.9679266895761741,
|
| 1189 |
+
"grad_norm": 0.500936047855792,
|
| 1190 |
+
"learning_rate": 5e-06,
|
| 1191 |
+
"loss": 0.4138,
|
| 1192 |
+
"step": 1690
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.9736540664375716,
|
| 1196 |
+
"grad_norm": 0.5586135425680717,
|
| 1197 |
+
"learning_rate": 5e-06,
|
| 1198 |
+
"loss": 0.4348,
|
| 1199 |
+
"step": 1700
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.979381443298969,
|
| 1203 |
+
"grad_norm": 0.5493075698232323,
|
| 1204 |
+
"learning_rate": 5e-06,
|
| 1205 |
+
"loss": 0.4235,
|
| 1206 |
+
"step": 1710
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.9851088201603666,
|
| 1210 |
+
"grad_norm": 0.5618879613810053,
|
| 1211 |
+
"learning_rate": 5e-06,
|
| 1212 |
+
"loss": 0.4331,
|
| 1213 |
+
"step": 1720
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.9908361970217641,
|
| 1217 |
+
"grad_norm": 0.5518636911221633,
|
| 1218 |
+
"learning_rate": 5e-06,
|
| 1219 |
+
"loss": 0.4324,
|
| 1220 |
+
"step": 1730
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.9965635738831615,
|
| 1224 |
+
"grad_norm": 0.540041363098471,
|
| 1225 |
+
"learning_rate": 5e-06,
|
| 1226 |
+
"loss": 0.4292,
|
| 1227 |
+
"step": 1740
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 1.0,
|
| 1231 |
+
"eval_loss": 0.4308567941188812,
|
| 1232 |
+
"eval_runtime": 442.3894,
|
| 1233 |
+
"eval_samples_per_second": 26.587,
|
| 1234 |
+
"eval_steps_per_second": 0.416,
|
| 1235 |
+
"step": 1746
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 1.002290950744559,
|
| 1239 |
+
"grad_norm": 0.5315755979624516,
|
| 1240 |
+
"learning_rate": 5e-06,
|
| 1241 |
+
"loss": 0.4118,
|
| 1242 |
+
"step": 1750
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 1.0080183276059564,
|
| 1246 |
+
"grad_norm": 0.5480247806796007,
|
| 1247 |
+
"learning_rate": 5e-06,
|
| 1248 |
+
"loss": 0.3744,
|
| 1249 |
+
"step": 1760
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.013745704467354,
|
| 1253 |
+
"grad_norm": 0.5149756078279293,
|
| 1254 |
+
"learning_rate": 5e-06,
|
| 1255 |
+
"loss": 0.378,
|
| 1256 |
+
"step": 1770
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 1.0194730813287514,
|
| 1260 |
+
"grad_norm": 0.5171540890283103,
|
| 1261 |
+
"learning_rate": 5e-06,
|
| 1262 |
+
"loss": 0.3702,
|
| 1263 |
+
"step": 1780
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.0252004581901488,
|
| 1267 |
+
"grad_norm": 0.5125282145044338,
|
| 1268 |
+
"learning_rate": 5e-06,
|
| 1269 |
+
"loss": 0.3711,
|
| 1270 |
+
"step": 1790
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 1.0309278350515463,
|
| 1274 |
+
"grad_norm": 0.5207422414484574,
|
| 1275 |
+
"learning_rate": 5e-06,
|
| 1276 |
+
"loss": 0.3665,
|
| 1277 |
+
"step": 1800
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 1.036655211912944,
|
| 1281 |
+
"grad_norm": 0.5786010896530626,
|
| 1282 |
+
"learning_rate": 5e-06,
|
| 1283 |
+
"loss": 0.3787,
|
| 1284 |
+
"step": 1810
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 1.0423825887743414,
|
| 1288 |
+
"grad_norm": 0.5464322830590085,
|
| 1289 |
+
"learning_rate": 5e-06,
|
| 1290 |
+
"loss": 0.3701,
|
| 1291 |
+
"step": 1820
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 1.0481099656357389,
|
| 1295 |
+
"grad_norm": 0.5205995452066101,
|
| 1296 |
+
"learning_rate": 5e-06,
|
| 1297 |
+
"loss": 0.3775,
|
| 1298 |
+
"step": 1830
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 1.0538373424971363,
|
| 1302 |
+
"grad_norm": 0.519243616807458,
|
| 1303 |
+
"learning_rate": 5e-06,
|
| 1304 |
+
"loss": 0.3796,
|
| 1305 |
+
"step": 1840
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 1.0595647193585338,
|
| 1309 |
+
"grad_norm": 0.516879079058521,
|
| 1310 |
+
"learning_rate": 5e-06,
|
| 1311 |
+
"loss": 0.3682,
|
| 1312 |
+
"step": 1850
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 1.0652920962199313,
|
| 1316 |
+
"grad_norm": 0.5222769824122077,
|
| 1317 |
+
"learning_rate": 5e-06,
|
| 1318 |
+
"loss": 0.3727,
|
| 1319 |
+
"step": 1860
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 1.0710194730813287,
|
| 1323 |
+
"grad_norm": 0.5445361815315836,
|
| 1324 |
+
"learning_rate": 5e-06,
|
| 1325 |
+
"loss": 0.3808,
|
| 1326 |
+
"step": 1870
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 1.0767468499427262,
|
| 1330 |
+
"grad_norm": 0.5036623685893024,
|
| 1331 |
+
"learning_rate": 5e-06,
|
| 1332 |
+
"loss": 0.372,
|
| 1333 |
+
"step": 1880
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 1.0824742268041236,
|
| 1337 |
+
"grad_norm": 0.5348337621107334,
|
| 1338 |
+
"learning_rate": 5e-06,
|
| 1339 |
+
"loss": 0.3732,
|
| 1340 |
+
"step": 1890
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 1.088201603665521,
|
| 1344 |
+
"grad_norm": 0.5060065384189831,
|
| 1345 |
+
"learning_rate": 5e-06,
|
| 1346 |
+
"loss": 0.3846,
|
| 1347 |
+
"step": 1900
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 1.0939289805269188,
|
| 1351 |
+
"grad_norm": 0.5079630313899987,
|
| 1352 |
+
"learning_rate": 5e-06,
|
| 1353 |
+
"loss": 0.3734,
|
| 1354 |
+
"step": 1910
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 1.0996563573883162,
|
| 1358 |
+
"grad_norm": 0.5280819989664591,
|
| 1359 |
+
"learning_rate": 5e-06,
|
| 1360 |
+
"loss": 0.3795,
|
| 1361 |
+
"step": 1920
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 1.1053837342497137,
|
| 1365 |
+
"grad_norm": 0.5170704766163384,
|
| 1366 |
+
"learning_rate": 5e-06,
|
| 1367 |
+
"loss": 0.3813,
|
| 1368 |
+
"step": 1930
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 1.1111111111111112,
|
| 1372 |
+
"grad_norm": 0.490210122517552,
|
| 1373 |
+
"learning_rate": 5e-06,
|
| 1374 |
+
"loss": 0.3675,
|
| 1375 |
+
"step": 1940
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 1.1168384879725086,
|
| 1379 |
+
"grad_norm": 0.5132872988639963,
|
| 1380 |
+
"learning_rate": 5e-06,
|
| 1381 |
+
"loss": 0.3725,
|
| 1382 |
+
"step": 1950
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 1.122565864833906,
|
| 1386 |
+
"grad_norm": 0.5345634201153158,
|
| 1387 |
+
"learning_rate": 5e-06,
|
| 1388 |
+
"loss": 0.3717,
|
| 1389 |
+
"step": 1960
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 1.1282932416953035,
|
| 1393 |
+
"grad_norm": 0.509614500177815,
|
| 1394 |
+
"learning_rate": 5e-06,
|
| 1395 |
+
"loss": 0.376,
|
| 1396 |
+
"step": 1970
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 1.134020618556701,
|
| 1400 |
+
"grad_norm": 0.533789127642796,
|
| 1401 |
+
"learning_rate": 5e-06,
|
| 1402 |
+
"loss": 0.3813,
|
| 1403 |
+
"step": 1980
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 1.1397479954180985,
|
| 1407 |
+
"grad_norm": 0.5430074485285918,
|
| 1408 |
+
"learning_rate": 5e-06,
|
| 1409 |
+
"loss": 0.3836,
|
| 1410 |
+
"step": 1990
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 1.145475372279496,
|
| 1414 |
+
"grad_norm": 0.5313196763751518,
|
| 1415 |
+
"learning_rate": 5e-06,
|
| 1416 |
+
"loss": 0.3733,
|
| 1417 |
+
"step": 2000
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.1512027491408934,
|
| 1421 |
+
"grad_norm": 0.5044508200904723,
|
| 1422 |
+
"learning_rate": 5e-06,
|
| 1423 |
+
"loss": 0.3783,
|
| 1424 |
+
"step": 2010
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 1.1569301260022908,
|
| 1428 |
+
"grad_norm": 0.5686318202507047,
|
| 1429 |
+
"learning_rate": 5e-06,
|
| 1430 |
+
"loss": 0.3752,
|
| 1431 |
+
"step": 2020
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 1.1626575028636885,
|
| 1435 |
+
"grad_norm": 0.5467468838281937,
|
| 1436 |
+
"learning_rate": 5e-06,
|
| 1437 |
+
"loss": 0.3762,
|
| 1438 |
+
"step": 2030
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 1.168384879725086,
|
| 1442 |
+
"grad_norm": 0.5149539692191945,
|
| 1443 |
+
"learning_rate": 5e-06,
|
| 1444 |
+
"loss": 0.3743,
|
| 1445 |
+
"step": 2040
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 1.1741122565864834,
|
| 1449 |
+
"grad_norm": 0.5312244954262079,
|
| 1450 |
+
"learning_rate": 5e-06,
|
| 1451 |
+
"loss": 0.3746,
|
| 1452 |
+
"step": 2050
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 1.179839633447881,
|
| 1456 |
+
"grad_norm": 0.5219006331104148,
|
| 1457 |
+
"learning_rate": 5e-06,
|
| 1458 |
+
"loss": 0.3723,
|
| 1459 |
+
"step": 2060
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 1.1855670103092784,
|
| 1463 |
+
"grad_norm": 0.538164681785682,
|
| 1464 |
+
"learning_rate": 5e-06,
|
| 1465 |
+
"loss": 0.375,
|
| 1466 |
+
"step": 2070
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 1.1912943871706758,
|
| 1470 |
+
"grad_norm": 0.5113415155626444,
|
| 1471 |
+
"learning_rate": 5e-06,
|
| 1472 |
+
"loss": 0.3808,
|
| 1473 |
+
"step": 2080
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 1.1970217640320733,
|
| 1477 |
+
"grad_norm": 0.5321490376681038,
|
| 1478 |
+
"learning_rate": 5e-06,
|
| 1479 |
+
"loss": 0.3699,
|
| 1480 |
+
"step": 2090
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 1.2027491408934707,
|
| 1484 |
+
"grad_norm": 0.5024285063567219,
|
| 1485 |
+
"learning_rate": 5e-06,
|
| 1486 |
+
"loss": 0.37,
|
| 1487 |
+
"step": 2100
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 1.2084765177548682,
|
| 1491 |
+
"grad_norm": 0.500252002027109,
|
| 1492 |
+
"learning_rate": 5e-06,
|
| 1493 |
+
"loss": 0.3782,
|
| 1494 |
+
"step": 2110
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 1.2142038946162657,
|
| 1498 |
+
"grad_norm": 0.4924458421349769,
|
| 1499 |
+
"learning_rate": 5e-06,
|
| 1500 |
+
"loss": 0.3743,
|
| 1501 |
+
"step": 2120
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.2199312714776633,
|
| 1505 |
+
"grad_norm": 0.4994293087338354,
|
| 1506 |
+
"learning_rate": 5e-06,
|
| 1507 |
+
"loss": 0.3758,
|
| 1508 |
+
"step": 2130
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 1.2256586483390608,
|
| 1512 |
+
"grad_norm": 0.5610101096098137,
|
| 1513 |
+
"learning_rate": 5e-06,
|
| 1514 |
+
"loss": 0.3805,
|
| 1515 |
+
"step": 2140
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 1.2313860252004583,
|
| 1519 |
+
"grad_norm": 0.537084984018735,
|
| 1520 |
+
"learning_rate": 5e-06,
|
| 1521 |
+
"loss": 0.3803,
|
| 1522 |
+
"step": 2150
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 1.2371134020618557,
|
| 1526 |
+
"grad_norm": 0.5371742151025276,
|
| 1527 |
+
"learning_rate": 5e-06,
|
| 1528 |
+
"loss": 0.38,
|
| 1529 |
+
"step": 2160
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 1.2428407789232532,
|
| 1533 |
+
"grad_norm": 0.5349028198444167,
|
| 1534 |
+
"learning_rate": 5e-06,
|
| 1535 |
+
"loss": 0.3808,
|
| 1536 |
+
"step": 2170
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 1.2485681557846506,
|
| 1540 |
+
"grad_norm": 0.5168637085023853,
|
| 1541 |
+
"learning_rate": 5e-06,
|
| 1542 |
+
"loss": 0.3762,
|
| 1543 |
+
"step": 2180
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 1.254295532646048,
|
| 1547 |
+
"grad_norm": 0.521040039917101,
|
| 1548 |
+
"learning_rate": 5e-06,
|
| 1549 |
+
"loss": 0.3724,
|
| 1550 |
+
"step": 2190
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 1.2600229095074456,
|
| 1554 |
+
"grad_norm": 0.547586207928514,
|
| 1555 |
+
"learning_rate": 5e-06,
|
| 1556 |
+
"loss": 0.379,
|
| 1557 |
+
"step": 2200
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 1.265750286368843,
|
| 1561 |
+
"grad_norm": 0.5217616276100534,
|
| 1562 |
+
"learning_rate": 5e-06,
|
| 1563 |
+
"loss": 0.3801,
|
| 1564 |
+
"step": 2210
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 1.2714776632302405,
|
| 1568 |
+
"grad_norm": 0.5119693142628771,
|
| 1569 |
+
"learning_rate": 5e-06,
|
| 1570 |
+
"loss": 0.3717,
|
| 1571 |
+
"step": 2220
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.277205040091638,
|
| 1575 |
+
"grad_norm": 0.49171715528514415,
|
| 1576 |
+
"learning_rate": 5e-06,
|
| 1577 |
+
"loss": 0.3733,
|
| 1578 |
+
"step": 2230
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 1.2829324169530354,
|
| 1582 |
+
"grad_norm": 0.5583188472803077,
|
| 1583 |
+
"learning_rate": 5e-06,
|
| 1584 |
+
"loss": 0.3848,
|
| 1585 |
+
"step": 2240
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.2886597938144329,
|
| 1589 |
+
"grad_norm": 0.5384351613915103,
|
| 1590 |
+
"learning_rate": 5e-06,
|
| 1591 |
+
"loss": 0.3708,
|
| 1592 |
+
"step": 2250
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 1.2943871706758305,
|
| 1596 |
+
"grad_norm": 0.5283801778798911,
|
| 1597 |
+
"learning_rate": 5e-06,
|
| 1598 |
+
"loss": 0.3737,
|
| 1599 |
+
"step": 2260
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 1.300114547537228,
|
| 1603 |
+
"grad_norm": 0.516459149172914,
|
| 1604 |
+
"learning_rate": 5e-06,
|
| 1605 |
+
"loss": 0.3778,
|
| 1606 |
+
"step": 2270
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 1.3058419243986255,
|
| 1610 |
+
"grad_norm": 0.5064553031768843,
|
| 1611 |
+
"learning_rate": 5e-06,
|
| 1612 |
+
"loss": 0.3803,
|
| 1613 |
+
"step": 2280
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 1.311569301260023,
|
| 1617 |
+
"grad_norm": 0.4992554472760309,
|
| 1618 |
+
"learning_rate": 5e-06,
|
| 1619 |
+
"loss": 0.3644,
|
| 1620 |
+
"step": 2290
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 1.3172966781214204,
|
| 1624 |
+
"grad_norm": 0.5351102245997288,
|
| 1625 |
+
"learning_rate": 5e-06,
|
| 1626 |
+
"loss": 0.3754,
|
| 1627 |
+
"step": 2300
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 1.3230240549828178,
|
| 1631 |
+
"grad_norm": 0.5744380836159825,
|
| 1632 |
+
"learning_rate": 5e-06,
|
| 1633 |
+
"loss": 0.3773,
|
| 1634 |
+
"step": 2310
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 1.3287514318442153,
|
| 1638 |
+
"grad_norm": 0.5390997556454836,
|
| 1639 |
+
"learning_rate": 5e-06,
|
| 1640 |
+
"loss": 0.3702,
|
| 1641 |
+
"step": 2320
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 1.3344788087056128,
|
| 1645 |
+
"grad_norm": 0.5159422766312324,
|
| 1646 |
+
"learning_rate": 5e-06,
|
| 1647 |
+
"loss": 0.3662,
|
| 1648 |
+
"step": 2330
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 1.3402061855670104,
|
| 1652 |
+
"grad_norm": 0.5122893955215309,
|
| 1653 |
+
"learning_rate": 5e-06,
|
| 1654 |
+
"loss": 0.3662,
|
| 1655 |
+
"step": 2340
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 1.345933562428408,
|
| 1659 |
+
"grad_norm": 0.5435516572685336,
|
| 1660 |
+
"learning_rate": 5e-06,
|
| 1661 |
+
"loss": 0.3692,
|
| 1662 |
+
"step": 2350
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 1.3516609392898054,
|
| 1666 |
+
"grad_norm": 0.506519754552173,
|
| 1667 |
+
"learning_rate": 5e-06,
|
| 1668 |
+
"loss": 0.3775,
|
| 1669 |
+
"step": 2360
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 1.3573883161512028,
|
| 1673 |
+
"grad_norm": 0.5382970585180483,
|
| 1674 |
+
"learning_rate": 5e-06,
|
| 1675 |
+
"loss": 0.3744,
|
| 1676 |
+
"step": 2370
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 1.3631156930126003,
|
| 1680 |
+
"grad_norm": 0.5098385678364493,
|
| 1681 |
+
"learning_rate": 5e-06,
|
| 1682 |
+
"loss": 0.3817,
|
| 1683 |
+
"step": 2380
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 1.3688430698739977,
|
| 1687 |
+
"grad_norm": 0.554626835720726,
|
| 1688 |
+
"learning_rate": 5e-06,
|
| 1689 |
+
"loss": 0.3846,
|
| 1690 |
+
"step": 2390
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 1.3745704467353952,
|
| 1694 |
+
"grad_norm": 0.5148678581013812,
|
| 1695 |
+
"learning_rate": 5e-06,
|
| 1696 |
+
"loss": 0.3673,
|
| 1697 |
+
"step": 2400
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 1.3802978235967927,
|
| 1701 |
+
"grad_norm": 0.5160690729715556,
|
| 1702 |
+
"learning_rate": 5e-06,
|
| 1703 |
+
"loss": 0.3733,
|
| 1704 |
+
"step": 2410
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 1.38602520045819,
|
| 1708 |
+
"grad_norm": 0.49317947843655263,
|
| 1709 |
+
"learning_rate": 5e-06,
|
| 1710 |
+
"loss": 0.3676,
|
| 1711 |
+
"step": 2420
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 1.3917525773195876,
|
| 1715 |
+
"grad_norm": 0.491680558889149,
|
| 1716 |
+
"learning_rate": 5e-06,
|
| 1717 |
+
"loss": 0.3726,
|
| 1718 |
+
"step": 2430
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 1.397479954180985,
|
| 1722 |
+
"grad_norm": 0.5159563830253588,
|
| 1723 |
+
"learning_rate": 5e-06,
|
| 1724 |
+
"loss": 0.3722,
|
| 1725 |
+
"step": 2440
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 1.4032073310423825,
|
| 1729 |
+
"grad_norm": 0.5153913293053918,
|
| 1730 |
+
"learning_rate": 5e-06,
|
| 1731 |
+
"loss": 0.378,
|
| 1732 |
+
"step": 2450
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 1.40893470790378,
|
| 1736 |
+
"grad_norm": 0.5264135232150384,
|
| 1737 |
+
"learning_rate": 5e-06,
|
| 1738 |
+
"loss": 0.3788,
|
| 1739 |
+
"step": 2460
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 1.4146620847651776,
|
| 1743 |
+
"grad_norm": 0.5382150315757761,
|
| 1744 |
+
"learning_rate": 5e-06,
|
| 1745 |
+
"loss": 0.3728,
|
| 1746 |
+
"step": 2470
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 1.420389461626575,
|
| 1750 |
+
"grad_norm": 0.5310778600445981,
|
| 1751 |
+
"learning_rate": 5e-06,
|
| 1752 |
+
"loss": 0.3725,
|
| 1753 |
+
"step": 2480
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 1.4261168384879725,
|
| 1757 |
+
"grad_norm": 0.5190694903204963,
|
| 1758 |
+
"learning_rate": 5e-06,
|
| 1759 |
+
"loss": 0.3708,
|
| 1760 |
+
"step": 2490
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 1.43184421534937,
|
| 1764 |
+
"grad_norm": 0.5074266048391012,
|
| 1765 |
+
"learning_rate": 5e-06,
|
| 1766 |
+
"loss": 0.3816,
|
| 1767 |
+
"step": 2500
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 1.4375715922107675,
|
| 1771 |
+
"grad_norm": 0.527668984455016,
|
| 1772 |
+
"learning_rate": 5e-06,
|
| 1773 |
+
"loss": 0.3692,
|
| 1774 |
+
"step": 2510
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 1.443298969072165,
|
| 1778 |
+
"grad_norm": 0.5676081369199447,
|
| 1779 |
+
"learning_rate": 5e-06,
|
| 1780 |
+
"loss": 0.3726,
|
| 1781 |
+
"step": 2520
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 1.4490263459335624,
|
| 1785 |
+
"grad_norm": 0.734177855082109,
|
| 1786 |
+
"learning_rate": 5e-06,
|
| 1787 |
+
"loss": 0.3626,
|
| 1788 |
+
"step": 2530
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 1.4547537227949598,
|
| 1792 |
+
"grad_norm": 0.5303161196964384,
|
| 1793 |
+
"learning_rate": 5e-06,
|
| 1794 |
+
"loss": 0.3736,
|
| 1795 |
+
"step": 2540
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 1.4604810996563573,
|
| 1799 |
+
"grad_norm": 0.4844944790072633,
|
| 1800 |
+
"learning_rate": 5e-06,
|
| 1801 |
+
"loss": 0.3717,
|
| 1802 |
+
"step": 2550
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 1.466208476517755,
|
| 1806 |
+
"grad_norm": 0.5315356528221638,
|
| 1807 |
+
"learning_rate": 5e-06,
|
| 1808 |
+
"loss": 0.3739,
|
| 1809 |
+
"step": 2560
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 1.4719358533791524,
|
| 1813 |
+
"grad_norm": 0.4772067027942158,
|
| 1814 |
+
"learning_rate": 5e-06,
|
| 1815 |
+
"loss": 0.3812,
|
| 1816 |
+
"step": 2570
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 1.47766323024055,
|
| 1820 |
+
"grad_norm": 0.4939040642118836,
|
| 1821 |
+
"learning_rate": 5e-06,
|
| 1822 |
+
"loss": 0.3672,
|
| 1823 |
+
"step": 2580
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 1.4833906071019474,
|
| 1827 |
+
"grad_norm": 0.508946651609973,
|
| 1828 |
+
"learning_rate": 5e-06,
|
| 1829 |
+
"loss": 0.3746,
|
| 1830 |
+
"step": 2590
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 1.4891179839633448,
|
| 1834 |
+
"grad_norm": 0.5298078020553034,
|
| 1835 |
+
"learning_rate": 5e-06,
|
| 1836 |
+
"loss": 0.3733,
|
| 1837 |
+
"step": 2600
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 1.4948453608247423,
|
| 1841 |
+
"grad_norm": 0.5432646094312185,
|
| 1842 |
+
"learning_rate": 5e-06,
|
| 1843 |
+
"loss": 0.3758,
|
| 1844 |
+
"step": 2610
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 1.5005727376861397,
|
| 1848 |
+
"grad_norm": 0.5105782450661664,
|
| 1849 |
+
"learning_rate": 5e-06,
|
| 1850 |
+
"loss": 0.3805,
|
| 1851 |
+
"step": 2620
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 1.5063001145475372,
|
| 1855 |
+
"grad_norm": 0.5382093019088371,
|
| 1856 |
+
"learning_rate": 5e-06,
|
| 1857 |
+
"loss": 0.381,
|
| 1858 |
+
"step": 2630
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 1.5120274914089347,
|
| 1862 |
+
"grad_norm": 0.5204647981548546,
|
| 1863 |
+
"learning_rate": 5e-06,
|
| 1864 |
+
"loss": 0.3787,
|
| 1865 |
+
"step": 2640
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 1.5177548682703321,
|
| 1869 |
+
"grad_norm": 0.5407064147138322,
|
| 1870 |
+
"learning_rate": 5e-06,
|
| 1871 |
+
"loss": 0.379,
|
| 1872 |
+
"step": 2650
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 1.5234822451317296,
|
| 1876 |
+
"grad_norm": 0.5213399797673937,
|
| 1877 |
+
"learning_rate": 5e-06,
|
| 1878 |
+
"loss": 0.3815,
|
| 1879 |
+
"step": 2660
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 1.529209621993127,
|
| 1883 |
+
"grad_norm": 0.5213816003555798,
|
| 1884 |
+
"learning_rate": 5e-06,
|
| 1885 |
+
"loss": 0.3881,
|
| 1886 |
+
"step": 2670
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 1.5349369988545245,
|
| 1890 |
+
"grad_norm": 0.5142228705336312,
|
| 1891 |
+
"learning_rate": 5e-06,
|
| 1892 |
+
"loss": 0.3776,
|
| 1893 |
+
"step": 2680
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 1.540664375715922,
|
| 1897 |
+
"grad_norm": 0.520697436811985,
|
| 1898 |
+
"learning_rate": 5e-06,
|
| 1899 |
+
"loss": 0.3806,
|
| 1900 |
+
"step": 2690
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 1.5463917525773194,
|
| 1904 |
+
"grad_norm": 0.5637817682476384,
|
| 1905 |
+
"learning_rate": 5e-06,
|
| 1906 |
+
"loss": 0.3715,
|
| 1907 |
+
"step": 2700
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 1.552119129438717,
|
| 1911 |
+
"grad_norm": 0.4915517576152862,
|
| 1912 |
+
"learning_rate": 5e-06,
|
| 1913 |
+
"loss": 0.3727,
|
| 1914 |
+
"step": 2710
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 1.5578465063001146,
|
| 1918 |
+
"grad_norm": 0.5237728548563729,
|
| 1919 |
+
"learning_rate": 5e-06,
|
| 1920 |
+
"loss": 0.3709,
|
| 1921 |
+
"step": 2720
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 1.563573883161512,
|
| 1925 |
+
"grad_norm": 0.7190217505740812,
|
| 1926 |
+
"learning_rate": 5e-06,
|
| 1927 |
+
"loss": 0.3803,
|
| 1928 |
+
"step": 2730
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 1.5693012600229095,
|
| 1932 |
+
"grad_norm": 0.5209643317738942,
|
| 1933 |
+
"learning_rate": 5e-06,
|
| 1934 |
+
"loss": 0.371,
|
| 1935 |
+
"step": 2740
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 1.575028636884307,
|
| 1939 |
+
"grad_norm": 0.5247668495585847,
|
| 1940 |
+
"learning_rate": 5e-06,
|
| 1941 |
+
"loss": 0.3808,
|
| 1942 |
+
"step": 2750
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 1.5807560137457046,
|
| 1946 |
+
"grad_norm": 0.532703132334951,
|
| 1947 |
+
"learning_rate": 5e-06,
|
| 1948 |
+
"loss": 0.378,
|
| 1949 |
+
"step": 2760
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 1.586483390607102,
|
| 1953 |
+
"grad_norm": 0.5117006044420915,
|
| 1954 |
+
"learning_rate": 5e-06,
|
| 1955 |
+
"loss": 0.3695,
|
| 1956 |
+
"step": 2770
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 1.5922107674684995,
|
| 1960 |
+
"grad_norm": 0.5272216030753158,
|
| 1961 |
+
"learning_rate": 5e-06,
|
| 1962 |
+
"loss": 0.3666,
|
| 1963 |
+
"step": 2780
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 1.597938144329897,
|
| 1967 |
+
"grad_norm": 0.5392842250351713,
|
| 1968 |
+
"learning_rate": 5e-06,
|
| 1969 |
+
"loss": 0.3783,
|
| 1970 |
+
"step": 2790
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 1.6036655211912945,
|
| 1974 |
+
"grad_norm": 0.5211103103323186,
|
| 1975 |
+
"learning_rate": 5e-06,
|
| 1976 |
+
"loss": 0.3809,
|
| 1977 |
+
"step": 2800
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 1.609392898052692,
|
| 1981 |
+
"grad_norm": 0.5434144298333962,
|
| 1982 |
+
"learning_rate": 5e-06,
|
| 1983 |
+
"loss": 0.3839,
|
| 1984 |
+
"step": 2810
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 1.6151202749140894,
|
| 1988 |
+
"grad_norm": 0.6244778503333394,
|
| 1989 |
+
"learning_rate": 5e-06,
|
| 1990 |
+
"loss": 0.375,
|
| 1991 |
+
"step": 2820
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 1.6208476517754868,
|
| 1995 |
+
"grad_norm": 0.7929144005943337,
|
| 1996 |
+
"learning_rate": 5e-06,
|
| 1997 |
+
"loss": 0.3843,
|
| 1998 |
+
"step": 2830
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 1.6265750286368843,
|
| 2002 |
+
"grad_norm": 0.5076072639573785,
|
| 2003 |
+
"learning_rate": 5e-06,
|
| 2004 |
+
"loss": 0.3764,
|
| 2005 |
+
"step": 2840
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 1.6323024054982818,
|
| 2009 |
+
"grad_norm": 0.499119817733024,
|
| 2010 |
+
"learning_rate": 5e-06,
|
| 2011 |
+
"loss": 0.3814,
|
| 2012 |
+
"step": 2850
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 1.6380297823596792,
|
| 2016 |
+
"grad_norm": 0.5427237791683881,
|
| 2017 |
+
"learning_rate": 5e-06,
|
| 2018 |
+
"loss": 0.3732,
|
| 2019 |
+
"step": 2860
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 1.6437571592210767,
|
| 2023 |
+
"grad_norm": 0.5072517280886356,
|
| 2024 |
+
"learning_rate": 5e-06,
|
| 2025 |
+
"loss": 0.3716,
|
| 2026 |
+
"step": 2870
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 1.6494845360824741,
|
| 2030 |
+
"grad_norm": 0.5424242226752932,
|
| 2031 |
+
"learning_rate": 5e-06,
|
| 2032 |
+
"loss": 0.3725,
|
| 2033 |
+
"step": 2880
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 1.6552119129438716,
|
| 2037 |
+
"grad_norm": 0.5239482084133175,
|
| 2038 |
+
"learning_rate": 5e-06,
|
| 2039 |
+
"loss": 0.3797,
|
| 2040 |
+
"step": 2890
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 1.660939289805269,
|
| 2044 |
+
"grad_norm": 0.4877340871651499,
|
| 2045 |
+
"learning_rate": 5e-06,
|
| 2046 |
+
"loss": 0.3797,
|
| 2047 |
+
"step": 2900
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 1.6666666666666665,
|
| 2051 |
+
"grad_norm": 0.5187696419666292,
|
| 2052 |
+
"learning_rate": 5e-06,
|
| 2053 |
+
"loss": 0.3732,
|
| 2054 |
+
"step": 2910
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 1.6723940435280642,
|
| 2058 |
+
"grad_norm": 0.5165454581988194,
|
| 2059 |
+
"learning_rate": 5e-06,
|
| 2060 |
+
"loss": 0.3761,
|
| 2061 |
+
"step": 2920
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 1.6781214203894617,
|
| 2065 |
+
"grad_norm": 0.5295455488567519,
|
| 2066 |
+
"learning_rate": 5e-06,
|
| 2067 |
+
"loss": 0.3732,
|
| 2068 |
+
"step": 2930
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 1.6838487972508591,
|
| 2072 |
+
"grad_norm": 0.5356558582460169,
|
| 2073 |
+
"learning_rate": 5e-06,
|
| 2074 |
+
"loss": 0.3759,
|
| 2075 |
+
"step": 2940
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 1.6895761741122566,
|
| 2079 |
+
"grad_norm": 0.5382234203233621,
|
| 2080 |
+
"learning_rate": 5e-06,
|
| 2081 |
+
"loss": 0.3809,
|
| 2082 |
+
"step": 2950
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 1.695303550973654,
|
| 2086 |
+
"grad_norm": 0.5426839059978352,
|
| 2087 |
+
"learning_rate": 5e-06,
|
| 2088 |
+
"loss": 0.3818,
|
| 2089 |
+
"step": 2960
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"epoch": 1.7010309278350515,
|
| 2093 |
+
"grad_norm": 0.5589401799758412,
|
| 2094 |
+
"learning_rate": 5e-06,
|
| 2095 |
+
"loss": 0.3735,
|
| 2096 |
+
"step": 2970
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 1.7067583046964492,
|
| 2100 |
+
"grad_norm": 0.5400070080828028,
|
| 2101 |
+
"learning_rate": 5e-06,
|
| 2102 |
+
"loss": 0.3715,
|
| 2103 |
+
"step": 2980
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 1.7124856815578466,
|
| 2107 |
+
"grad_norm": 0.5125880859835712,
|
| 2108 |
+
"learning_rate": 5e-06,
|
| 2109 |
+
"loss": 0.3692,
|
| 2110 |
+
"step": 2990
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 1.718213058419244,
|
| 2114 |
+
"grad_norm": 0.521750795912995,
|
| 2115 |
+
"learning_rate": 5e-06,
|
| 2116 |
+
"loss": 0.3734,
|
| 2117 |
+
"step": 3000
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 1.7239404352806416,
|
| 2121 |
+
"grad_norm": 0.49977791549900996,
|
| 2122 |
+
"learning_rate": 5e-06,
|
| 2123 |
+
"loss": 0.3714,
|
| 2124 |
+
"step": 3010
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 1.729667812142039,
|
| 2128 |
+
"grad_norm": 0.517176383552139,
|
| 2129 |
+
"learning_rate": 5e-06,
|
| 2130 |
+
"loss": 0.3798,
|
| 2131 |
+
"step": 3020
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 1.7353951890034365,
|
| 2135 |
+
"grad_norm": 0.5569996765160103,
|
| 2136 |
+
"learning_rate": 5e-06,
|
| 2137 |
+
"loss": 0.3704,
|
| 2138 |
+
"step": 3030
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 1.741122565864834,
|
| 2142 |
+
"grad_norm": 0.5152384355166261,
|
| 2143 |
+
"learning_rate": 5e-06,
|
| 2144 |
+
"loss": 0.3687,
|
| 2145 |
+
"step": 3040
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 1.7468499427262314,
|
| 2149 |
+
"grad_norm": 0.5152215076607375,
|
| 2150 |
+
"learning_rate": 5e-06,
|
| 2151 |
+
"loss": 0.3711,
|
| 2152 |
+
"step": 3050
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 1.7525773195876289,
|
| 2156 |
+
"grad_norm": 0.5209367759111013,
|
| 2157 |
+
"learning_rate": 5e-06,
|
| 2158 |
+
"loss": 0.367,
|
| 2159 |
+
"step": 3060
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 1.7583046964490263,
|
| 2163 |
+
"grad_norm": 0.72478206720126,
|
| 2164 |
+
"learning_rate": 5e-06,
|
| 2165 |
+
"loss": 0.3753,
|
| 2166 |
+
"step": 3070
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 1.7640320733104238,
|
| 2170 |
+
"grad_norm": 0.5108998808880466,
|
| 2171 |
+
"learning_rate": 5e-06,
|
| 2172 |
+
"loss": 0.3776,
|
| 2173 |
+
"step": 3080
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 1.7697594501718212,
|
| 2177 |
+
"grad_norm": 0.49545669464589914,
|
| 2178 |
+
"learning_rate": 5e-06,
|
| 2179 |
+
"loss": 0.3691,
|
| 2180 |
+
"step": 3090
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 1.7754868270332187,
|
| 2184 |
+
"grad_norm": 0.5370814607474198,
|
| 2185 |
+
"learning_rate": 5e-06,
|
| 2186 |
+
"loss": 0.3746,
|
| 2187 |
+
"step": 3100
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 1.7812142038946162,
|
| 2191 |
+
"grad_norm": 0.6072170324437888,
|
| 2192 |
+
"learning_rate": 5e-06,
|
| 2193 |
+
"loss": 0.3729,
|
| 2194 |
+
"step": 3110
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 1.7869415807560136,
|
| 2198 |
+
"grad_norm": 0.5271754924806036,
|
| 2199 |
+
"learning_rate": 5e-06,
|
| 2200 |
+
"loss": 0.3588,
|
| 2201 |
+
"step": 3120
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 1.792668957617411,
|
| 2205 |
+
"grad_norm": 0.5754508589270141,
|
| 2206 |
+
"learning_rate": 5e-06,
|
| 2207 |
+
"loss": 0.3776,
|
| 2208 |
+
"step": 3130
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 1.7983963344788088,
|
| 2212 |
+
"grad_norm": 0.5403755251129506,
|
| 2213 |
+
"learning_rate": 5e-06,
|
| 2214 |
+
"loss": 0.3762,
|
| 2215 |
+
"step": 3140
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 1.8041237113402062,
|
| 2219 |
+
"grad_norm": 0.5138981496909732,
|
| 2220 |
+
"learning_rate": 5e-06,
|
| 2221 |
+
"loss": 0.3711,
|
| 2222 |
+
"step": 3150
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 1.8098510882016037,
|
| 2226 |
+
"grad_norm": 0.534919240445055,
|
| 2227 |
+
"learning_rate": 5e-06,
|
| 2228 |
+
"loss": 0.3763,
|
| 2229 |
+
"step": 3160
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 1.8155784650630011,
|
| 2233 |
+
"grad_norm": 0.5146135891570391,
|
| 2234 |
+
"learning_rate": 5e-06,
|
| 2235 |
+
"loss": 0.3806,
|
| 2236 |
+
"step": 3170
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 1.8213058419243986,
|
| 2240 |
+
"grad_norm": 0.505605657209453,
|
| 2241 |
+
"learning_rate": 5e-06,
|
| 2242 |
+
"loss": 0.3766,
|
| 2243 |
+
"step": 3180
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 1.827033218785796,
|
| 2247 |
+
"grad_norm": 0.5239153217495089,
|
| 2248 |
+
"learning_rate": 5e-06,
|
| 2249 |
+
"loss": 0.3659,
|
| 2250 |
+
"step": 3190
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 1.8327605956471937,
|
| 2254 |
+
"grad_norm": 0.9509826091058188,
|
| 2255 |
+
"learning_rate": 5e-06,
|
| 2256 |
+
"loss": 0.3792,
|
| 2257 |
+
"step": 3200
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 1.8384879725085912,
|
| 2261 |
+
"grad_norm": 0.5059729711091226,
|
| 2262 |
+
"learning_rate": 5e-06,
|
| 2263 |
+
"loss": 0.3721,
|
| 2264 |
+
"step": 3210
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 1.8442153493699887,
|
| 2268 |
+
"grad_norm": 0.5245359643059282,
|
| 2269 |
+
"learning_rate": 5e-06,
|
| 2270 |
+
"loss": 0.379,
|
| 2271 |
+
"step": 3220
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 1.8499427262313861,
|
| 2275 |
+
"grad_norm": 0.5050698399570109,
|
| 2276 |
+
"learning_rate": 5e-06,
|
| 2277 |
+
"loss": 0.3739,
|
| 2278 |
+
"step": 3230
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 1.8556701030927836,
|
| 2282 |
+
"grad_norm": 0.5065697206422057,
|
| 2283 |
+
"learning_rate": 5e-06,
|
| 2284 |
+
"loss": 0.3849,
|
| 2285 |
+
"step": 3240
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 1.861397479954181,
|
| 2289 |
+
"grad_norm": 0.5209454651612013,
|
| 2290 |
+
"learning_rate": 5e-06,
|
| 2291 |
+
"loss": 0.3707,
|
| 2292 |
+
"step": 3250
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 1.8671248568155785,
|
| 2296 |
+
"grad_norm": 0.5430271570033155,
|
| 2297 |
+
"learning_rate": 5e-06,
|
| 2298 |
+
"loss": 0.3696,
|
| 2299 |
+
"step": 3260
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 1.872852233676976,
|
| 2303 |
+
"grad_norm": 0.5281297299023378,
|
| 2304 |
+
"learning_rate": 5e-06,
|
| 2305 |
+
"loss": 0.3712,
|
| 2306 |
+
"step": 3270
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 1.8785796105383734,
|
| 2310 |
+
"grad_norm": 0.5292724874313788,
|
| 2311 |
+
"learning_rate": 5e-06,
|
| 2312 |
+
"loss": 0.3726,
|
| 2313 |
+
"step": 3280
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 1.8843069873997709,
|
| 2317 |
+
"grad_norm": 0.5350000049801282,
|
| 2318 |
+
"learning_rate": 5e-06,
|
| 2319 |
+
"loss": 0.3723,
|
| 2320 |
+
"step": 3290
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 1.8900343642611683,
|
| 2324 |
+
"grad_norm": 0.5208368353063128,
|
| 2325 |
+
"learning_rate": 5e-06,
|
| 2326 |
+
"loss": 0.3677,
|
| 2327 |
+
"step": 3300
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 1.8957617411225658,
|
| 2331 |
+
"grad_norm": 0.5279268625494672,
|
| 2332 |
+
"learning_rate": 5e-06,
|
| 2333 |
+
"loss": 0.3753,
|
| 2334 |
+
"step": 3310
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 1.9014891179839633,
|
| 2338 |
+
"grad_norm": 0.5144354886651064,
|
| 2339 |
+
"learning_rate": 5e-06,
|
| 2340 |
+
"loss": 0.3741,
|
| 2341 |
+
"step": 3320
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 1.9072164948453607,
|
| 2345 |
+
"grad_norm": 0.5082140336402234,
|
| 2346 |
+
"learning_rate": 5e-06,
|
| 2347 |
+
"loss": 0.3635,
|
| 2348 |
+
"step": 3330
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 1.9129438717067582,
|
| 2352 |
+
"grad_norm": 0.5252002760881638,
|
| 2353 |
+
"learning_rate": 5e-06,
|
| 2354 |
+
"loss": 0.3787,
|
| 2355 |
+
"step": 3340
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 1.9186712485681556,
|
| 2359 |
+
"grad_norm": 0.5070002443248173,
|
| 2360 |
+
"learning_rate": 5e-06,
|
| 2361 |
+
"loss": 0.3718,
|
| 2362 |
+
"step": 3350
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 1.9243986254295533,
|
| 2366 |
+
"grad_norm": 0.5316148384707632,
|
| 2367 |
+
"learning_rate": 5e-06,
|
| 2368 |
+
"loss": 0.3664,
|
| 2369 |
+
"step": 3360
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 1.9301260022909508,
|
| 2373 |
+
"grad_norm": 0.5193287345279981,
|
| 2374 |
+
"learning_rate": 5e-06,
|
| 2375 |
+
"loss": 0.3684,
|
| 2376 |
+
"step": 3370
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 1.9358533791523482,
|
| 2380 |
+
"grad_norm": 0.506552724781655,
|
| 2381 |
+
"learning_rate": 5e-06,
|
| 2382 |
+
"loss": 0.372,
|
| 2383 |
+
"step": 3380
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 1.9415807560137457,
|
| 2387 |
+
"grad_norm": 0.571442610460597,
|
| 2388 |
+
"learning_rate": 5e-06,
|
| 2389 |
+
"loss": 0.379,
|
| 2390 |
+
"step": 3390
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 1.9473081328751431,
|
| 2394 |
+
"grad_norm": 0.5688327247671221,
|
| 2395 |
+
"learning_rate": 5e-06,
|
| 2396 |
+
"loss": 0.3687,
|
| 2397 |
+
"step": 3400
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 1.9530355097365406,
|
| 2401 |
+
"grad_norm": 0.5042217231517675,
|
| 2402 |
+
"learning_rate": 5e-06,
|
| 2403 |
+
"loss": 0.3738,
|
| 2404 |
+
"step": 3410
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 1.9587628865979383,
|
| 2408 |
+
"grad_norm": 0.5321243972550566,
|
| 2409 |
+
"learning_rate": 5e-06,
|
| 2410 |
+
"loss": 0.37,
|
| 2411 |
+
"step": 3420
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 1.9644902634593358,
|
| 2415 |
+
"grad_norm": 0.5188108633782941,
|
| 2416 |
+
"learning_rate": 5e-06,
|
| 2417 |
+
"loss": 0.3766,
|
| 2418 |
+
"step": 3430
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 1.9702176403207332,
|
| 2422 |
+
"grad_norm": 0.49989426974744805,
|
| 2423 |
+
"learning_rate": 5e-06,
|
| 2424 |
+
"loss": 0.3703,
|
| 2425 |
+
"step": 3440
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 1.9759450171821307,
|
| 2429 |
+
"grad_norm": 0.5428089945263836,
|
| 2430 |
+
"learning_rate": 5e-06,
|
| 2431 |
+
"loss": 0.3689,
|
| 2432 |
+
"step": 3450
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 1.9816723940435281,
|
| 2436 |
+
"grad_norm": 0.5404439065398382,
|
| 2437 |
+
"learning_rate": 5e-06,
|
| 2438 |
+
"loss": 0.3729,
|
| 2439 |
+
"step": 3460
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 1.9873997709049256,
|
| 2443 |
+
"grad_norm": 0.5410379700662915,
|
| 2444 |
+
"learning_rate": 5e-06,
|
| 2445 |
+
"loss": 0.3726,
|
| 2446 |
+
"step": 3470
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 1.993127147766323,
|
| 2450 |
+
"grad_norm": 0.5225961306378605,
|
| 2451 |
+
"learning_rate": 5e-06,
|
| 2452 |
+
"loss": 0.3782,
|
| 2453 |
+
"step": 3480
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 1.9988545246277205,
|
| 2457 |
+
"grad_norm": 0.5535365053498117,
|
| 2458 |
+
"learning_rate": 5e-06,
|
| 2459 |
+
"loss": 0.3757,
|
| 2460 |
+
"step": 3490
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 2.0,
|
| 2464 |
+
"eval_loss": 0.42091062664985657,
|
| 2465 |
+
"eval_runtime": 444.2503,
|
| 2466 |
+
"eval_samples_per_second": 26.476,
|
| 2467 |
+
"eval_steps_per_second": 0.414,
|
| 2468 |
+
"step": 3492
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 2.004581901489118,
|
| 2472 |
+
"grad_norm": 0.5715227391793876,
|
| 2473 |
+
"learning_rate": 5e-06,
|
| 2474 |
+
"loss": 0.3266,
|
| 2475 |
+
"step": 3500
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 2.0103092783505154,
|
| 2479 |
+
"grad_norm": 0.5115288441622716,
|
| 2480 |
+
"learning_rate": 5e-06,
|
| 2481 |
+
"loss": 0.3123,
|
| 2482 |
+
"step": 3510
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 2.016036655211913,
|
| 2486 |
+
"grad_norm": 0.5114186476378918,
|
| 2487 |
+
"learning_rate": 5e-06,
|
| 2488 |
+
"loss": 0.3064,
|
| 2489 |
+
"step": 3520
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 2.0217640320733103,
|
| 2493 |
+
"grad_norm": 0.5174383277753903,
|
| 2494 |
+
"learning_rate": 5e-06,
|
| 2495 |
+
"loss": 0.3115,
|
| 2496 |
+
"step": 3530
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 2.027491408934708,
|
| 2500 |
+
"grad_norm": 0.5180676439991899,
|
| 2501 |
+
"learning_rate": 5e-06,
|
| 2502 |
+
"loss": 0.3092,
|
| 2503 |
+
"step": 3540
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 2.0332187857961053,
|
| 2507 |
+
"grad_norm": 0.5521743990569499,
|
| 2508 |
+
"learning_rate": 5e-06,
|
| 2509 |
+
"loss": 0.3154,
|
| 2510 |
+
"step": 3550
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 2.0389461626575027,
|
| 2514 |
+
"grad_norm": 0.5325564269068459,
|
| 2515 |
+
"learning_rate": 5e-06,
|
| 2516 |
+
"loss": 0.3139,
|
| 2517 |
+
"step": 3560
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 2.0446735395189,
|
| 2521 |
+
"grad_norm": 0.5307121051642008,
|
| 2522 |
+
"learning_rate": 5e-06,
|
| 2523 |
+
"loss": 0.3143,
|
| 2524 |
+
"step": 3570
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 2.0504009163802976,
|
| 2528 |
+
"grad_norm": 0.5039022580188123,
|
| 2529 |
+
"learning_rate": 5e-06,
|
| 2530 |
+
"loss": 0.3139,
|
| 2531 |
+
"step": 3580
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 2.056128293241695,
|
| 2535 |
+
"grad_norm": 0.5386599812235912,
|
| 2536 |
+
"learning_rate": 5e-06,
|
| 2537 |
+
"loss": 0.3274,
|
| 2538 |
+
"step": 3590
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 2.0618556701030926,
|
| 2542 |
+
"grad_norm": 0.5510122632798343,
|
| 2543 |
+
"learning_rate": 5e-06,
|
| 2544 |
+
"loss": 0.3218,
|
| 2545 |
+
"step": 3600
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 2.0675830469644905,
|
| 2549 |
+
"grad_norm": 0.5314262243630897,
|
| 2550 |
+
"learning_rate": 5e-06,
|
| 2551 |
+
"loss": 0.3126,
|
| 2552 |
+
"step": 3610
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 2.073310423825888,
|
| 2556 |
+
"grad_norm": 0.4939184680389998,
|
| 2557 |
+
"learning_rate": 5e-06,
|
| 2558 |
+
"loss": 0.313,
|
| 2559 |
+
"step": 3620
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 2.0790378006872854,
|
| 2563 |
+
"grad_norm": 0.5312396764234006,
|
| 2564 |
+
"learning_rate": 5e-06,
|
| 2565 |
+
"loss": 0.3105,
|
| 2566 |
+
"step": 3630
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 2.084765177548683,
|
| 2570 |
+
"grad_norm": 0.4876119734890598,
|
| 2571 |
+
"learning_rate": 5e-06,
|
| 2572 |
+
"loss": 0.3124,
|
| 2573 |
+
"step": 3640
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 2.0904925544100803,
|
| 2577 |
+
"grad_norm": 0.5103227941442228,
|
| 2578 |
+
"learning_rate": 5e-06,
|
| 2579 |
+
"loss": 0.3242,
|
| 2580 |
+
"step": 3650
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 2.0962199312714778,
|
| 2584 |
+
"grad_norm": 0.5042566969256604,
|
| 2585 |
+
"learning_rate": 5e-06,
|
| 2586 |
+
"loss": 0.3184,
|
| 2587 |
+
"step": 3660
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 2.1019473081328752,
|
| 2591 |
+
"grad_norm": 0.5141761625174609,
|
| 2592 |
+
"learning_rate": 5e-06,
|
| 2593 |
+
"loss": 0.3143,
|
| 2594 |
+
"step": 3670
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 2.1076746849942727,
|
| 2598 |
+
"grad_norm": 0.5382583222094659,
|
| 2599 |
+
"learning_rate": 5e-06,
|
| 2600 |
+
"loss": 0.3166,
|
| 2601 |
+
"step": 3680
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 2.11340206185567,
|
| 2605 |
+
"grad_norm": 0.4804864664629375,
|
| 2606 |
+
"learning_rate": 5e-06,
|
| 2607 |
+
"loss": 0.3122,
|
| 2608 |
+
"step": 3690
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 2.1191294387170676,
|
| 2612 |
+
"grad_norm": 0.5165262098916743,
|
| 2613 |
+
"learning_rate": 5e-06,
|
| 2614 |
+
"loss": 0.3135,
|
| 2615 |
+
"step": 3700
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 2.124856815578465,
|
| 2619 |
+
"grad_norm": 0.5253973489118585,
|
| 2620 |
+
"learning_rate": 5e-06,
|
| 2621 |
+
"loss": 0.3103,
|
| 2622 |
+
"step": 3710
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 2.1305841924398625,
|
| 2626 |
+
"grad_norm": 0.5153233857956726,
|
| 2627 |
+
"learning_rate": 5e-06,
|
| 2628 |
+
"loss": 0.319,
|
| 2629 |
+
"step": 3720
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 2.13631156930126,
|
| 2633 |
+
"grad_norm": 0.5024911676484453,
|
| 2634 |
+
"learning_rate": 5e-06,
|
| 2635 |
+
"loss": 0.3153,
|
| 2636 |
+
"step": 3730
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 2.1420389461626574,
|
| 2640 |
+
"grad_norm": 0.48649497978161954,
|
| 2641 |
+
"learning_rate": 5e-06,
|
| 2642 |
+
"loss": 0.3125,
|
| 2643 |
+
"step": 3740
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 2.147766323024055,
|
| 2647 |
+
"grad_norm": 0.5398906998916857,
|
| 2648 |
+
"learning_rate": 5e-06,
|
| 2649 |
+
"loss": 0.3207,
|
| 2650 |
+
"step": 3750
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 2.1534936998854524,
|
| 2654 |
+
"grad_norm": 0.5130668071668567,
|
| 2655 |
+
"learning_rate": 5e-06,
|
| 2656 |
+
"loss": 0.3141,
|
| 2657 |
+
"step": 3760
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 2.15922107674685,
|
| 2661 |
+
"grad_norm": 0.5518874676047263,
|
| 2662 |
+
"learning_rate": 5e-06,
|
| 2663 |
+
"loss": 0.3184,
|
| 2664 |
+
"step": 3770
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 2.1649484536082473,
|
| 2668 |
+
"grad_norm": 0.5160046944915151,
|
| 2669 |
+
"learning_rate": 5e-06,
|
| 2670 |
+
"loss": 0.3172,
|
| 2671 |
+
"step": 3780
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 2.1706758304696447,
|
| 2675 |
+
"grad_norm": 0.48834912906902556,
|
| 2676 |
+
"learning_rate": 5e-06,
|
| 2677 |
+
"loss": 0.3234,
|
| 2678 |
+
"step": 3790
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 2.176403207331042,
|
| 2682 |
+
"grad_norm": 0.5171110867595475,
|
| 2683 |
+
"learning_rate": 5e-06,
|
| 2684 |
+
"loss": 0.3151,
|
| 2685 |
+
"step": 3800
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 2.1821305841924397,
|
| 2689 |
+
"grad_norm": 0.5180778501865309,
|
| 2690 |
+
"learning_rate": 5e-06,
|
| 2691 |
+
"loss": 0.3157,
|
| 2692 |
+
"step": 3810
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 2.1878579610538376,
|
| 2696 |
+
"grad_norm": 0.5068334930400914,
|
| 2697 |
+
"learning_rate": 5e-06,
|
| 2698 |
+
"loss": 0.3126,
|
| 2699 |
+
"step": 3820
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 2.193585337915235,
|
| 2703 |
+
"grad_norm": 0.5383783523767623,
|
| 2704 |
+
"learning_rate": 5e-06,
|
| 2705 |
+
"loss": 0.313,
|
| 2706 |
+
"step": 3830
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 2.1993127147766325,
|
| 2710 |
+
"grad_norm": 0.5231488887255323,
|
| 2711 |
+
"learning_rate": 5e-06,
|
| 2712 |
+
"loss": 0.3199,
|
| 2713 |
+
"step": 3840
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 2.20504009163803,
|
| 2717 |
+
"grad_norm": 0.5130199004591844,
|
| 2718 |
+
"learning_rate": 5e-06,
|
| 2719 |
+
"loss": 0.3168,
|
| 2720 |
+
"step": 3850
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 2.2107674684994274,
|
| 2724 |
+
"grad_norm": 0.5482589908800253,
|
| 2725 |
+
"learning_rate": 5e-06,
|
| 2726 |
+
"loss": 0.3223,
|
| 2727 |
+
"step": 3860
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 2.216494845360825,
|
| 2731 |
+
"grad_norm": 0.53120848952022,
|
| 2732 |
+
"learning_rate": 5e-06,
|
| 2733 |
+
"loss": 0.321,
|
| 2734 |
+
"step": 3870
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 2.2222222222222223,
|
| 2738 |
+
"grad_norm": 0.5071013057492006,
|
| 2739 |
+
"learning_rate": 5e-06,
|
| 2740 |
+
"loss": 0.3203,
|
| 2741 |
+
"step": 3880
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 2.22794959908362,
|
| 2745 |
+
"grad_norm": 0.531266327465737,
|
| 2746 |
+
"learning_rate": 5e-06,
|
| 2747 |
+
"loss": 0.324,
|
| 2748 |
+
"step": 3890
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 2.2336769759450172,
|
| 2752 |
+
"grad_norm": 0.507052566002429,
|
| 2753 |
+
"learning_rate": 5e-06,
|
| 2754 |
+
"loss": 0.3137,
|
| 2755 |
+
"step": 3900
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 2.2394043528064147,
|
| 2759 |
+
"grad_norm": 0.5271639744898703,
|
| 2760 |
+
"learning_rate": 5e-06,
|
| 2761 |
+
"loss": 0.3159,
|
| 2762 |
+
"step": 3910
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 2.245131729667812,
|
| 2766 |
+
"grad_norm": 0.5388005656911042,
|
| 2767 |
+
"learning_rate": 5e-06,
|
| 2768 |
+
"loss": 0.3228,
|
| 2769 |
+
"step": 3920
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 2.2508591065292096,
|
| 2773 |
+
"grad_norm": 0.5060194168439476,
|
| 2774 |
+
"learning_rate": 5e-06,
|
| 2775 |
+
"loss": 0.3124,
|
| 2776 |
+
"step": 3930
|
| 2777 |
+
},
|
| 2778 |
+
{
|
| 2779 |
+
"epoch": 2.256586483390607,
|
| 2780 |
+
"grad_norm": 0.48546648398689385,
|
| 2781 |
+
"learning_rate": 5e-06,
|
| 2782 |
+
"loss": 0.3136,
|
| 2783 |
+
"step": 3940
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 2.2623138602520045,
|
| 2787 |
+
"grad_norm": 0.5076224304840579,
|
| 2788 |
+
"learning_rate": 5e-06,
|
| 2789 |
+
"loss": 0.3167,
|
| 2790 |
+
"step": 3950
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 2.268041237113402,
|
| 2794 |
+
"grad_norm": 0.5296050756375043,
|
| 2795 |
+
"learning_rate": 5e-06,
|
| 2796 |
+
"loss": 0.3204,
|
| 2797 |
+
"step": 3960
|
| 2798 |
+
},
|
| 2799 |
+
{
|
| 2800 |
+
"epoch": 2.2737686139747995,
|
| 2801 |
+
"grad_norm": 0.5177826901430715,
|
| 2802 |
+
"learning_rate": 5e-06,
|
| 2803 |
+
"loss": 0.3232,
|
| 2804 |
+
"step": 3970
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 2.279495990836197,
|
| 2808 |
+
"grad_norm": 0.5132840661554632,
|
| 2809 |
+
"learning_rate": 5e-06,
|
| 2810 |
+
"loss": 0.3158,
|
| 2811 |
+
"step": 3980
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 2.2852233676975944,
|
| 2815 |
+
"grad_norm": 0.543452264339846,
|
| 2816 |
+
"learning_rate": 5e-06,
|
| 2817 |
+
"loss": 0.3192,
|
| 2818 |
+
"step": 3990
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 2.290950744558992,
|
| 2822 |
+
"grad_norm": 0.5139552493291161,
|
| 2823 |
+
"learning_rate": 5e-06,
|
| 2824 |
+
"loss": 0.3209,
|
| 2825 |
+
"step": 4000
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 2.2966781214203893,
|
| 2829 |
+
"grad_norm": 0.4969492673393391,
|
| 2830 |
+
"learning_rate": 5e-06,
|
| 2831 |
+
"loss": 0.3153,
|
| 2832 |
+
"step": 4010
|
| 2833 |
+
},
|
| 2834 |
+
{
|
| 2835 |
+
"epoch": 2.3024054982817868,
|
| 2836 |
+
"grad_norm": 0.5143134020150514,
|
| 2837 |
+
"learning_rate": 5e-06,
|
| 2838 |
+
"loss": 0.3201,
|
| 2839 |
+
"step": 4020
|
| 2840 |
+
},
|
| 2841 |
+
{
|
| 2842 |
+
"epoch": 2.308132875143184,
|
| 2843 |
+
"grad_norm": 0.4939050512552889,
|
| 2844 |
+
"learning_rate": 5e-06,
|
| 2845 |
+
"loss": 0.3106,
|
| 2846 |
+
"step": 4030
|
| 2847 |
+
},
|
| 2848 |
+
{
|
| 2849 |
+
"epoch": 2.3138602520045817,
|
| 2850 |
+
"grad_norm": 0.4886273394938644,
|
| 2851 |
+
"learning_rate": 5e-06,
|
| 2852 |
+
"loss": 0.3225,
|
| 2853 |
+
"step": 4040
|
| 2854 |
+
},
|
| 2855 |
+
{
|
| 2856 |
+
"epoch": 2.319587628865979,
|
| 2857 |
+
"grad_norm": 0.5380016012501289,
|
| 2858 |
+
"learning_rate": 5e-06,
|
| 2859 |
+
"loss": 0.3183,
|
| 2860 |
+
"step": 4050
|
| 2861 |
+
},
|
| 2862 |
+
{
|
| 2863 |
+
"epoch": 2.325315005727377,
|
| 2864 |
+
"grad_norm": 0.5160375481982886,
|
| 2865 |
+
"learning_rate": 5e-06,
|
| 2866 |
+
"loss": 0.3127,
|
| 2867 |
+
"step": 4060
|
| 2868 |
+
},
|
| 2869 |
+
{
|
| 2870 |
+
"epoch": 2.3310423825887745,
|
| 2871 |
+
"grad_norm": 0.5217824879654166,
|
| 2872 |
+
"learning_rate": 5e-06,
|
| 2873 |
+
"loss": 0.3168,
|
| 2874 |
+
"step": 4070
|
| 2875 |
+
},
|
| 2876 |
+
{
|
| 2877 |
+
"epoch": 2.336769759450172,
|
| 2878 |
+
"grad_norm": 0.4987920921308887,
|
| 2879 |
+
"learning_rate": 5e-06,
|
| 2880 |
+
"loss": 0.3243,
|
| 2881 |
+
"step": 4080
|
| 2882 |
+
},
|
| 2883 |
+
{
|
| 2884 |
+
"epoch": 2.3424971363115694,
|
| 2885 |
+
"grad_norm": 0.540718540329614,
|
| 2886 |
+
"learning_rate": 5e-06,
|
| 2887 |
+
"loss": 0.3267,
|
| 2888 |
+
"step": 4090
|
| 2889 |
+
},
|
| 2890 |
+
{
|
| 2891 |
+
"epoch": 2.348224513172967,
|
| 2892 |
+
"grad_norm": 0.4928104897116783,
|
| 2893 |
+
"learning_rate": 5e-06,
|
| 2894 |
+
"loss": 0.3154,
|
| 2895 |
+
"step": 4100
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 2.3539518900343643,
|
| 2899 |
+
"grad_norm": 0.5332905023045372,
|
| 2900 |
+
"learning_rate": 5e-06,
|
| 2901 |
+
"loss": 0.3294,
|
| 2902 |
+
"step": 4110
|
| 2903 |
+
},
|
| 2904 |
+
{
|
| 2905 |
+
"epoch": 2.359679266895762,
|
| 2906 |
+
"grad_norm": 0.5190878954413758,
|
| 2907 |
+
"learning_rate": 5e-06,
|
| 2908 |
+
"loss": 0.3135,
|
| 2909 |
+
"step": 4120
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 2.3654066437571593,
|
| 2913 |
+
"grad_norm": 0.49770844545756987,
|
| 2914 |
+
"learning_rate": 5e-06,
|
| 2915 |
+
"loss": 0.3217,
|
| 2916 |
+
"step": 4130
|
| 2917 |
+
},
|
| 2918 |
+
{
|
| 2919 |
+
"epoch": 2.3711340206185567,
|
| 2920 |
+
"grad_norm": 0.5158902735624868,
|
| 2921 |
+
"learning_rate": 5e-06,
|
| 2922 |
+
"loss": 0.3182,
|
| 2923 |
+
"step": 4140
|
| 2924 |
+
},
|
| 2925 |
+
{
|
| 2926 |
+
"epoch": 2.376861397479954,
|
| 2927 |
+
"grad_norm": 0.5258004488461884,
|
| 2928 |
+
"learning_rate": 5e-06,
|
| 2929 |
+
"loss": 0.3211,
|
| 2930 |
+
"step": 4150
|
| 2931 |
+
},
|
| 2932 |
+
{
|
| 2933 |
+
"epoch": 2.3825887743413516,
|
| 2934 |
+
"grad_norm": 0.5006246348277473,
|
| 2935 |
+
"learning_rate": 5e-06,
|
| 2936 |
+
"loss": 0.311,
|
| 2937 |
+
"step": 4160
|
| 2938 |
+
},
|
| 2939 |
+
{
|
| 2940 |
+
"epoch": 2.388316151202749,
|
| 2941 |
+
"grad_norm": 0.5152750751947476,
|
| 2942 |
+
"learning_rate": 5e-06,
|
| 2943 |
+
"loss": 0.3219,
|
| 2944 |
+
"step": 4170
|
| 2945 |
+
},
|
| 2946 |
+
{
|
| 2947 |
+
"epoch": 2.3940435280641466,
|
| 2948 |
+
"grad_norm": 0.52068469477404,
|
| 2949 |
+
"learning_rate": 5e-06,
|
| 2950 |
+
"loss": 0.3164,
|
| 2951 |
+
"step": 4180
|
| 2952 |
+
},
|
| 2953 |
+
{
|
| 2954 |
+
"epoch": 2.399770904925544,
|
| 2955 |
+
"grad_norm": 0.5323767239444738,
|
| 2956 |
+
"learning_rate": 5e-06,
|
| 2957 |
+
"loss": 0.3208,
|
| 2958 |
+
"step": 4190
|
| 2959 |
+
},
|
| 2960 |
+
{
|
| 2961 |
+
"epoch": 2.4054982817869415,
|
| 2962 |
+
"grad_norm": 0.5003101909687979,
|
| 2963 |
+
"learning_rate": 5e-06,
|
| 2964 |
+
"loss": 0.3228,
|
| 2965 |
+
"step": 4200
|
| 2966 |
+
},
|
| 2967 |
+
{
|
| 2968 |
+
"epoch": 2.411225658648339,
|
| 2969 |
+
"grad_norm": 0.5111739731033206,
|
| 2970 |
+
"learning_rate": 5e-06,
|
| 2971 |
+
"loss": 0.3271,
|
| 2972 |
+
"step": 4210
|
| 2973 |
+
},
|
| 2974 |
+
{
|
| 2975 |
+
"epoch": 2.4169530355097364,
|
| 2976 |
+
"grad_norm": 0.5131151305511839,
|
| 2977 |
+
"learning_rate": 5e-06,
|
| 2978 |
+
"loss": 0.3147,
|
| 2979 |
+
"step": 4220
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 2.422680412371134,
|
| 2983 |
+
"grad_norm": 0.5163340601407542,
|
| 2984 |
+
"learning_rate": 5e-06,
|
| 2985 |
+
"loss": 0.3194,
|
| 2986 |
+
"step": 4230
|
| 2987 |
+
},
|
| 2988 |
+
{
|
| 2989 |
+
"epoch": 2.4284077892325313,
|
| 2990 |
+
"grad_norm": 0.4958432705018228,
|
| 2991 |
+
"learning_rate": 5e-06,
|
| 2992 |
+
"loss": 0.3188,
|
| 2993 |
+
"step": 4240
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 2.434135166093929,
|
| 2997 |
+
"grad_norm": 0.5468740104263736,
|
| 2998 |
+
"learning_rate": 5e-06,
|
| 2999 |
+
"loss": 0.3249,
|
| 3000 |
+
"step": 4250
|
| 3001 |
+
},
|
| 3002 |
+
{
|
| 3003 |
+
"epoch": 2.4398625429553267,
|
| 3004 |
+
"grad_norm": 0.5410705485869296,
|
| 3005 |
+
"learning_rate": 5e-06,
|
| 3006 |
+
"loss": 0.3278,
|
| 3007 |
+
"step": 4260
|
| 3008 |
+
},
|
| 3009 |
+
{
|
| 3010 |
+
"epoch": 2.445589919816724,
|
| 3011 |
+
"grad_norm": 0.5368673433466151,
|
| 3012 |
+
"learning_rate": 5e-06,
|
| 3013 |
+
"loss": 0.3076,
|
| 3014 |
+
"step": 4270
|
| 3015 |
+
},
|
| 3016 |
+
{
|
| 3017 |
+
"epoch": 2.4513172966781216,
|
| 3018 |
+
"grad_norm": 0.5128839768898594,
|
| 3019 |
+
"learning_rate": 5e-06,
|
| 3020 |
+
"loss": 0.3184,
|
| 3021 |
+
"step": 4280
|
| 3022 |
+
},
|
| 3023 |
+
{
|
| 3024 |
+
"epoch": 2.457044673539519,
|
| 3025 |
+
"grad_norm": 0.5068335080164907,
|
| 3026 |
+
"learning_rate": 5e-06,
|
| 3027 |
+
"loss": 0.3169,
|
| 3028 |
+
"step": 4290
|
| 3029 |
+
},
|
| 3030 |
+
{
|
| 3031 |
+
"epoch": 2.4627720504009165,
|
| 3032 |
+
"grad_norm": 0.4997641306590389,
|
| 3033 |
+
"learning_rate": 5e-06,
|
| 3034 |
+
"loss": 0.3182,
|
| 3035 |
+
"step": 4300
|
| 3036 |
+
},
|
| 3037 |
+
{
|
| 3038 |
+
"epoch": 2.468499427262314,
|
| 3039 |
+
"grad_norm": 0.5191935789555608,
|
| 3040 |
+
"learning_rate": 5e-06,
|
| 3041 |
+
"loss": 0.3216,
|
| 3042 |
+
"step": 4310
|
| 3043 |
+
},
|
| 3044 |
+
{
|
| 3045 |
+
"epoch": 2.4742268041237114,
|
| 3046 |
+
"grad_norm": 0.5256779474029062,
|
| 3047 |
+
"learning_rate": 5e-06,
|
| 3048 |
+
"loss": 0.3233,
|
| 3049 |
+
"step": 4320
|
| 3050 |
+
},
|
| 3051 |
+
{
|
| 3052 |
+
"epoch": 2.479954180985109,
|
| 3053 |
+
"grad_norm": 0.4865715340962184,
|
| 3054 |
+
"learning_rate": 5e-06,
|
| 3055 |
+
"loss": 0.3204,
|
| 3056 |
+
"step": 4330
|
| 3057 |
+
},
|
| 3058 |
+
{
|
| 3059 |
+
"epoch": 2.4856815578465064,
|
| 3060 |
+
"grad_norm": 0.526248237049682,
|
| 3061 |
+
"learning_rate": 5e-06,
|
| 3062 |
+
"loss": 0.3205,
|
| 3063 |
+
"step": 4340
|
| 3064 |
+
},
|
| 3065 |
+
{
|
| 3066 |
+
"epoch": 2.491408934707904,
|
| 3067 |
+
"grad_norm": 0.5130011897003942,
|
| 3068 |
+
"learning_rate": 5e-06,
|
| 3069 |
+
"loss": 0.3259,
|
| 3070 |
+
"step": 4350
|
| 3071 |
+
},
|
| 3072 |
+
{
|
| 3073 |
+
"epoch": 2.4971363115693013,
|
| 3074 |
+
"grad_norm": 0.5162374829270555,
|
| 3075 |
+
"learning_rate": 5e-06,
|
| 3076 |
+
"loss": 0.3195,
|
| 3077 |
+
"step": 4360
|
| 3078 |
+
},
|
| 3079 |
+
{
|
| 3080 |
+
"epoch": 2.5028636884306987,
|
| 3081 |
+
"grad_norm": 0.5014077765674124,
|
| 3082 |
+
"learning_rate": 5e-06,
|
| 3083 |
+
"loss": 0.3228,
|
| 3084 |
+
"step": 4370
|
| 3085 |
+
},
|
| 3086 |
+
{
|
| 3087 |
+
"epoch": 2.508591065292096,
|
| 3088 |
+
"grad_norm": 0.5026646925273252,
|
| 3089 |
+
"learning_rate": 5e-06,
|
| 3090 |
+
"loss": 0.3207,
|
| 3091 |
+
"step": 4380
|
| 3092 |
+
},
|
| 3093 |
+
{
|
| 3094 |
+
"epoch": 2.5143184421534936,
|
| 3095 |
+
"grad_norm": 0.489097601061641,
|
| 3096 |
+
"learning_rate": 5e-06,
|
| 3097 |
+
"loss": 0.3176,
|
| 3098 |
+
"step": 4390
|
| 3099 |
+
},
|
| 3100 |
+
{
|
| 3101 |
+
"epoch": 2.520045819014891,
|
| 3102 |
+
"grad_norm": 0.5078735499696359,
|
| 3103 |
+
"learning_rate": 5e-06,
|
| 3104 |
+
"loss": 0.3224,
|
| 3105 |
+
"step": 4400
|
| 3106 |
+
},
|
| 3107 |
+
{
|
| 3108 |
+
"epoch": 2.5257731958762886,
|
| 3109 |
+
"grad_norm": 0.5271652940063883,
|
| 3110 |
+
"learning_rate": 5e-06,
|
| 3111 |
+
"loss": 0.323,
|
| 3112 |
+
"step": 4410
|
| 3113 |
+
},
|
| 3114 |
+
{
|
| 3115 |
+
"epoch": 2.531500572737686,
|
| 3116 |
+
"grad_norm": 0.507078592732912,
|
| 3117 |
+
"learning_rate": 5e-06,
|
| 3118 |
+
"loss": 0.3203,
|
| 3119 |
+
"step": 4420
|
| 3120 |
+
},
|
| 3121 |
+
{
|
| 3122 |
+
"epoch": 2.5372279495990835,
|
| 3123 |
+
"grad_norm": 0.5337080175384938,
|
| 3124 |
+
"learning_rate": 5e-06,
|
| 3125 |
+
"loss": 0.3228,
|
| 3126 |
+
"step": 4430
|
| 3127 |
+
},
|
| 3128 |
+
{
|
| 3129 |
+
"epoch": 2.542955326460481,
|
| 3130 |
+
"grad_norm": 0.50176908036031,
|
| 3131 |
+
"learning_rate": 5e-06,
|
| 3132 |
+
"loss": 0.3202,
|
| 3133 |
+
"step": 4440
|
| 3134 |
+
},
|
| 3135 |
+
{
|
| 3136 |
+
"epoch": 2.5486827033218784,
|
| 3137 |
+
"grad_norm": 0.5120277803126527,
|
| 3138 |
+
"learning_rate": 5e-06,
|
| 3139 |
+
"loss": 0.3197,
|
| 3140 |
+
"step": 4450
|
| 3141 |
+
},
|
| 3142 |
+
{
|
| 3143 |
+
"epoch": 2.554410080183276,
|
| 3144 |
+
"grad_norm": 0.5366578152814068,
|
| 3145 |
+
"learning_rate": 5e-06,
|
| 3146 |
+
"loss": 0.3244,
|
| 3147 |
+
"step": 4460
|
| 3148 |
+
},
|
| 3149 |
+
{
|
| 3150 |
+
"epoch": 2.5601374570446733,
|
| 3151 |
+
"grad_norm": 0.522159840101665,
|
| 3152 |
+
"learning_rate": 5e-06,
|
| 3153 |
+
"loss": 0.3185,
|
| 3154 |
+
"step": 4470
|
| 3155 |
+
},
|
| 3156 |
+
{
|
| 3157 |
+
"epoch": 2.565864833906071,
|
| 3158 |
+
"grad_norm": 0.48196371586373893,
|
| 3159 |
+
"learning_rate": 5e-06,
|
| 3160 |
+
"loss": 0.3173,
|
| 3161 |
+
"step": 4480
|
| 3162 |
+
},
|
| 3163 |
+
{
|
| 3164 |
+
"epoch": 2.5715922107674682,
|
| 3165 |
+
"grad_norm": 0.5255302877210084,
|
| 3166 |
+
"learning_rate": 5e-06,
|
| 3167 |
+
"loss": 0.3278,
|
| 3168 |
+
"step": 4490
|
| 3169 |
+
},
|
| 3170 |
+
{
|
| 3171 |
+
"epoch": 2.5773195876288657,
|
| 3172 |
+
"grad_norm": 0.5089869190508646,
|
| 3173 |
+
"learning_rate": 5e-06,
|
| 3174 |
+
"loss": 0.3198,
|
| 3175 |
+
"step": 4500
|
| 3176 |
+
},
|
| 3177 |
+
{
|
| 3178 |
+
"epoch": 2.5830469644902636,
|
| 3179 |
+
"grad_norm": 0.530959409102577,
|
| 3180 |
+
"learning_rate": 5e-06,
|
| 3181 |
+
"loss": 0.3304,
|
| 3182 |
+
"step": 4510
|
| 3183 |
+
},
|
| 3184 |
+
{
|
| 3185 |
+
"epoch": 2.588774341351661,
|
| 3186 |
+
"grad_norm": 0.5531110752439229,
|
| 3187 |
+
"learning_rate": 5e-06,
|
| 3188 |
+
"loss": 0.3215,
|
| 3189 |
+
"step": 4520
|
| 3190 |
+
},
|
| 3191 |
+
{
|
| 3192 |
+
"epoch": 2.5945017182130585,
|
| 3193 |
+
"grad_norm": 0.5260126942389958,
|
| 3194 |
+
"learning_rate": 5e-06,
|
| 3195 |
+
"loss": 0.316,
|
| 3196 |
+
"step": 4530
|
| 3197 |
+
},
|
| 3198 |
+
{
|
| 3199 |
+
"epoch": 2.600229095074456,
|
| 3200 |
+
"grad_norm": 0.5160707506662007,
|
| 3201 |
+
"learning_rate": 5e-06,
|
| 3202 |
+
"loss": 0.32,
|
| 3203 |
+
"step": 4540
|
| 3204 |
+
},
|
| 3205 |
+
{
|
| 3206 |
+
"epoch": 2.6059564719358534,
|
| 3207 |
+
"grad_norm": 0.5410876892071711,
|
| 3208 |
+
"learning_rate": 5e-06,
|
| 3209 |
+
"loss": 0.324,
|
| 3210 |
+
"step": 4550
|
| 3211 |
+
},
|
| 3212 |
+
{
|
| 3213 |
+
"epoch": 2.611683848797251,
|
| 3214 |
+
"grad_norm": 0.5149983976318435,
|
| 3215 |
+
"learning_rate": 5e-06,
|
| 3216 |
+
"loss": 0.3168,
|
| 3217 |
+
"step": 4560
|
| 3218 |
+
},
|
| 3219 |
+
{
|
| 3220 |
+
"epoch": 2.6174112256586484,
|
| 3221 |
+
"grad_norm": 0.5066882681673974,
|
| 3222 |
+
"learning_rate": 5e-06,
|
| 3223 |
+
"loss": 0.3164,
|
| 3224 |
+
"step": 4570
|
| 3225 |
+
},
|
| 3226 |
+
{
|
| 3227 |
+
"epoch": 2.623138602520046,
|
| 3228 |
+
"grad_norm": 0.5033788230089447,
|
| 3229 |
+
"learning_rate": 5e-06,
|
| 3230 |
+
"loss": 0.3163,
|
| 3231 |
+
"step": 4580
|
| 3232 |
+
},
|
| 3233 |
+
{
|
| 3234 |
+
"epoch": 2.6288659793814433,
|
| 3235 |
+
"grad_norm": 0.5247195850361597,
|
| 3236 |
+
"learning_rate": 5e-06,
|
| 3237 |
+
"loss": 0.3175,
|
| 3238 |
+
"step": 4590
|
| 3239 |
+
},
|
| 3240 |
+
{
|
| 3241 |
+
"epoch": 2.6345933562428407,
|
| 3242 |
+
"grad_norm": 0.5282100918029693,
|
| 3243 |
+
"learning_rate": 5e-06,
|
| 3244 |
+
"loss": 0.3272,
|
| 3245 |
+
"step": 4600
|
| 3246 |
+
},
|
| 3247 |
+
{
|
| 3248 |
+
"epoch": 2.640320733104238,
|
| 3249 |
+
"grad_norm": 0.52717520655256,
|
| 3250 |
+
"learning_rate": 5e-06,
|
| 3251 |
+
"loss": 0.3203,
|
| 3252 |
+
"step": 4610
|
| 3253 |
+
},
|
| 3254 |
+
{
|
| 3255 |
+
"epoch": 2.6460481099656357,
|
| 3256 |
+
"grad_norm": 0.5022071775606154,
|
| 3257 |
+
"learning_rate": 5e-06,
|
| 3258 |
+
"loss": 0.3244,
|
| 3259 |
+
"step": 4620
|
| 3260 |
+
},
|
| 3261 |
+
{
|
| 3262 |
+
"epoch": 2.651775486827033,
|
| 3263 |
+
"grad_norm": 0.5211994501987708,
|
| 3264 |
+
"learning_rate": 5e-06,
|
| 3265 |
+
"loss": 0.3131,
|
| 3266 |
+
"step": 4630
|
| 3267 |
+
},
|
| 3268 |
+
{
|
| 3269 |
+
"epoch": 2.6575028636884306,
|
| 3270 |
+
"grad_norm": 0.5124524380744641,
|
| 3271 |
+
"learning_rate": 5e-06,
|
| 3272 |
+
"loss": 0.3253,
|
| 3273 |
+
"step": 4640
|
| 3274 |
+
},
|
| 3275 |
+
{
|
| 3276 |
+
"epoch": 2.663230240549828,
|
| 3277 |
+
"grad_norm": 0.5060206463796821,
|
| 3278 |
+
"learning_rate": 5e-06,
|
| 3279 |
+
"loss": 0.3173,
|
| 3280 |
+
"step": 4650
|
| 3281 |
+
},
|
| 3282 |
+
{
|
| 3283 |
+
"epoch": 2.6689576174112255,
|
| 3284 |
+
"grad_norm": 0.5162650933726732,
|
| 3285 |
+
"learning_rate": 5e-06,
|
| 3286 |
+
"loss": 0.3198,
|
| 3287 |
+
"step": 4660
|
| 3288 |
+
},
|
| 3289 |
+
{
|
| 3290 |
+
"epoch": 2.6746849942726234,
|
| 3291 |
+
"grad_norm": 0.524408740926564,
|
| 3292 |
+
"learning_rate": 5e-06,
|
| 3293 |
+
"loss": 0.3302,
|
| 3294 |
+
"step": 4670
|
| 3295 |
+
},
|
| 3296 |
+
{
|
| 3297 |
+
"epoch": 2.680412371134021,
|
| 3298 |
+
"grad_norm": 0.5146510600522526,
|
| 3299 |
+
"learning_rate": 5e-06,
|
| 3300 |
+
"loss": 0.3234,
|
| 3301 |
+
"step": 4680
|
| 3302 |
+
},
|
| 3303 |
+
{
|
| 3304 |
+
"epoch": 2.6861397479954183,
|
| 3305 |
+
"grad_norm": 0.5092928068009863,
|
| 3306 |
+
"learning_rate": 5e-06,
|
| 3307 |
+
"loss": 0.322,
|
| 3308 |
+
"step": 4690
|
| 3309 |
+
},
|
| 3310 |
+
{
|
| 3311 |
+
"epoch": 2.691867124856816,
|
| 3312 |
+
"grad_norm": 0.5757942628503555,
|
| 3313 |
+
"learning_rate": 5e-06,
|
| 3314 |
+
"loss": 0.3117,
|
| 3315 |
+
"step": 4700
|
| 3316 |
+
},
|
| 3317 |
+
{
|
| 3318 |
+
"epoch": 2.6975945017182132,
|
| 3319 |
+
"grad_norm": 0.4889008213010037,
|
| 3320 |
+
"learning_rate": 5e-06,
|
| 3321 |
+
"loss": 0.3231,
|
| 3322 |
+
"step": 4710
|
| 3323 |
+
},
|
| 3324 |
+
{
|
| 3325 |
+
"epoch": 2.7033218785796107,
|
| 3326 |
+
"grad_norm": 0.5302974244163768,
|
| 3327 |
+
"learning_rate": 5e-06,
|
| 3328 |
+
"loss": 0.3178,
|
| 3329 |
+
"step": 4720
|
| 3330 |
+
},
|
| 3331 |
+
{
|
| 3332 |
+
"epoch": 2.709049255441008,
|
| 3333 |
+
"grad_norm": 0.52479839402123,
|
| 3334 |
+
"learning_rate": 5e-06,
|
| 3335 |
+
"loss": 0.3217,
|
| 3336 |
+
"step": 4730
|
| 3337 |
+
},
|
| 3338 |
+
{
|
| 3339 |
+
"epoch": 2.7147766323024056,
|
| 3340 |
+
"grad_norm": 0.5039321357679906,
|
| 3341 |
+
"learning_rate": 5e-06,
|
| 3342 |
+
"loss": 0.3257,
|
| 3343 |
+
"step": 4740
|
| 3344 |
+
},
|
| 3345 |
+
{
|
| 3346 |
+
"epoch": 2.720504009163803,
|
| 3347 |
+
"grad_norm": 0.5101102471452367,
|
| 3348 |
+
"learning_rate": 5e-06,
|
| 3349 |
+
"loss": 0.3222,
|
| 3350 |
+
"step": 4750
|
| 3351 |
+
},
|
| 3352 |
+
{
|
| 3353 |
+
"epoch": 2.7262313860252005,
|
| 3354 |
+
"grad_norm": 0.5430506058097,
|
| 3355 |
+
"learning_rate": 5e-06,
|
| 3356 |
+
"loss": 0.319,
|
| 3357 |
+
"step": 4760
|
| 3358 |
+
},
|
| 3359 |
+
{
|
| 3360 |
+
"epoch": 2.731958762886598,
|
| 3361 |
+
"grad_norm": 0.5355617323855306,
|
| 3362 |
+
"learning_rate": 5e-06,
|
| 3363 |
+
"loss": 0.3269,
|
| 3364 |
+
"step": 4770
|
| 3365 |
+
},
|
| 3366 |
+
{
|
| 3367 |
+
"epoch": 2.7376861397479955,
|
| 3368 |
+
"grad_norm": 0.4986561926249133,
|
| 3369 |
+
"learning_rate": 5e-06,
|
| 3370 |
+
"loss": 0.3149,
|
| 3371 |
+
"step": 4780
|
| 3372 |
+
},
|
| 3373 |
+
{
|
| 3374 |
+
"epoch": 2.743413516609393,
|
| 3375 |
+
"grad_norm": 0.5053547200537913,
|
| 3376 |
+
"learning_rate": 5e-06,
|
| 3377 |
+
"loss": 0.3308,
|
| 3378 |
+
"step": 4790
|
| 3379 |
+
},
|
| 3380 |
+
{
|
| 3381 |
+
"epoch": 2.7491408934707904,
|
| 3382 |
+
"grad_norm": 0.48632254517347523,
|
| 3383 |
+
"learning_rate": 5e-06,
|
| 3384 |
+
"loss": 0.3185,
|
| 3385 |
+
"step": 4800
|
| 3386 |
+
},
|
| 3387 |
+
{
|
| 3388 |
+
"epoch": 2.754868270332188,
|
| 3389 |
+
"grad_norm": 0.486358386605019,
|
| 3390 |
+
"learning_rate": 5e-06,
|
| 3391 |
+
"loss": 0.3103,
|
| 3392 |
+
"step": 4810
|
| 3393 |
+
},
|
| 3394 |
+
{
|
| 3395 |
+
"epoch": 2.7605956471935853,
|
| 3396 |
+
"grad_norm": 0.5374470855558577,
|
| 3397 |
+
"learning_rate": 5e-06,
|
| 3398 |
+
"loss": 0.3298,
|
| 3399 |
+
"step": 4820
|
| 3400 |
+
},
|
| 3401 |
+
{
|
| 3402 |
+
"epoch": 2.7663230240549828,
|
| 3403 |
+
"grad_norm": 0.5153628917876052,
|
| 3404 |
+
"learning_rate": 5e-06,
|
| 3405 |
+
"loss": 0.3233,
|
| 3406 |
+
"step": 4830
|
| 3407 |
+
},
|
| 3408 |
+
{
|
| 3409 |
+
"epoch": 2.77205040091638,
|
| 3410 |
+
"grad_norm": 0.5054154722646462,
|
| 3411 |
+
"learning_rate": 5e-06,
|
| 3412 |
+
"loss": 0.3275,
|
| 3413 |
+
"step": 4840
|
| 3414 |
+
},
|
| 3415 |
+
{
|
| 3416 |
+
"epoch": 2.7777777777777777,
|
| 3417 |
+
"grad_norm": 0.5149761064711591,
|
| 3418 |
+
"learning_rate": 5e-06,
|
| 3419 |
+
"loss": 0.317,
|
| 3420 |
+
"step": 4850
|
| 3421 |
+
},
|
| 3422 |
+
{
|
| 3423 |
+
"epoch": 2.783505154639175,
|
| 3424 |
+
"grad_norm": 0.5148931201825926,
|
| 3425 |
+
"learning_rate": 5e-06,
|
| 3426 |
+
"loss": 0.3233,
|
| 3427 |
+
"step": 4860
|
| 3428 |
+
},
|
| 3429 |
+
{
|
| 3430 |
+
"epoch": 2.7892325315005726,
|
| 3431 |
+
"grad_norm": 0.5191662793295072,
|
| 3432 |
+
"learning_rate": 5e-06,
|
| 3433 |
+
"loss": 0.3184,
|
| 3434 |
+
"step": 4870
|
| 3435 |
+
},
|
| 3436 |
+
{
|
| 3437 |
+
"epoch": 2.79495990836197,
|
| 3438 |
+
"grad_norm": 0.5675073234527983,
|
| 3439 |
+
"learning_rate": 5e-06,
|
| 3440 |
+
"loss": 0.3251,
|
| 3441 |
+
"step": 4880
|
| 3442 |
+
},
|
| 3443 |
+
{
|
| 3444 |
+
"epoch": 2.8006872852233675,
|
| 3445 |
+
"grad_norm": 0.5038525244128136,
|
| 3446 |
+
"learning_rate": 5e-06,
|
| 3447 |
+
"loss": 0.3232,
|
| 3448 |
+
"step": 4890
|
| 3449 |
+
},
|
| 3450 |
+
{
|
| 3451 |
+
"epoch": 2.806414662084765,
|
| 3452 |
+
"grad_norm": 0.4918634907852925,
|
| 3453 |
+
"learning_rate": 5e-06,
|
| 3454 |
+
"loss": 0.3175,
|
| 3455 |
+
"step": 4900
|
| 3456 |
+
},
|
| 3457 |
+
{
|
| 3458 |
+
"epoch": 2.8121420389461624,
|
| 3459 |
+
"grad_norm": 0.5284779685936459,
|
| 3460 |
+
"learning_rate": 5e-06,
|
| 3461 |
+
"loss": 0.3208,
|
| 3462 |
+
"step": 4910
|
| 3463 |
+
},
|
| 3464 |
+
{
|
| 3465 |
+
"epoch": 2.81786941580756,
|
| 3466 |
+
"grad_norm": 0.5108078870269828,
|
| 3467 |
+
"learning_rate": 5e-06,
|
| 3468 |
+
"loss": 0.3272,
|
| 3469 |
+
"step": 4920
|
| 3470 |
+
},
|
| 3471 |
+
{
|
| 3472 |
+
"epoch": 2.8235967926689574,
|
| 3473 |
+
"grad_norm": 0.4902747564813571,
|
| 3474 |
+
"learning_rate": 5e-06,
|
| 3475 |
+
"loss": 0.3257,
|
| 3476 |
+
"step": 4930
|
| 3477 |
+
},
|
| 3478 |
+
{
|
| 3479 |
+
"epoch": 2.8293241695303553,
|
| 3480 |
+
"grad_norm": 0.5030822928949704,
|
| 3481 |
+
"learning_rate": 5e-06,
|
| 3482 |
+
"loss": 0.32,
|
| 3483 |
+
"step": 4940
|
| 3484 |
+
},
|
| 3485 |
+
{
|
| 3486 |
+
"epoch": 2.8350515463917527,
|
| 3487 |
+
"grad_norm": 0.5272339203419648,
|
| 3488 |
+
"learning_rate": 5e-06,
|
| 3489 |
+
"loss": 0.3288,
|
| 3490 |
+
"step": 4950
|
| 3491 |
+
},
|
| 3492 |
+
{
|
| 3493 |
+
"epoch": 2.84077892325315,
|
| 3494 |
+
"grad_norm": 0.5340619443879843,
|
| 3495 |
+
"learning_rate": 5e-06,
|
| 3496 |
+
"loss": 0.3277,
|
| 3497 |
+
"step": 4960
|
| 3498 |
+
},
|
| 3499 |
+
{
|
| 3500 |
+
"epoch": 2.8465063001145476,
|
| 3501 |
+
"grad_norm": 0.5139875993534403,
|
| 3502 |
+
"learning_rate": 5e-06,
|
| 3503 |
+
"loss": 0.3167,
|
| 3504 |
+
"step": 4970
|
| 3505 |
+
},
|
| 3506 |
+
{
|
| 3507 |
+
"epoch": 2.852233676975945,
|
| 3508 |
+
"grad_norm": 0.5197113622112696,
|
| 3509 |
+
"learning_rate": 5e-06,
|
| 3510 |
+
"loss": 0.3221,
|
| 3511 |
+
"step": 4980
|
| 3512 |
+
},
|
| 3513 |
+
{
|
| 3514 |
+
"epoch": 2.8579610538373426,
|
| 3515 |
+
"grad_norm": 0.4932910708419312,
|
| 3516 |
+
"learning_rate": 5e-06,
|
| 3517 |
+
"loss": 0.319,
|
| 3518 |
+
"step": 4990
|
| 3519 |
+
},
|
| 3520 |
+
{
|
| 3521 |
+
"epoch": 2.86368843069874,
|
| 3522 |
+
"grad_norm": 0.48961874681440903,
|
| 3523 |
+
"learning_rate": 5e-06,
|
| 3524 |
+
"loss": 0.3216,
|
| 3525 |
+
"step": 5000
|
| 3526 |
+
},
|
| 3527 |
+
{
|
| 3528 |
+
"epoch": 2.8694158075601375,
|
| 3529 |
+
"grad_norm": 0.5306801859906654,
|
| 3530 |
+
"learning_rate": 5e-06,
|
| 3531 |
+
"loss": 0.3294,
|
| 3532 |
+
"step": 5010
|
| 3533 |
+
},
|
| 3534 |
+
{
|
| 3535 |
+
"epoch": 2.875143184421535,
|
| 3536 |
+
"grad_norm": 0.5245022306778138,
|
| 3537 |
+
"learning_rate": 5e-06,
|
| 3538 |
+
"loss": 0.3238,
|
| 3539 |
+
"step": 5020
|
| 3540 |
+
},
|
| 3541 |
+
{
|
| 3542 |
+
"epoch": 2.8808705612829324,
|
| 3543 |
+
"grad_norm": 0.5294236720116192,
|
| 3544 |
+
"learning_rate": 5e-06,
|
| 3545 |
+
"loss": 0.3172,
|
| 3546 |
+
"step": 5030
|
| 3547 |
+
},
|
| 3548 |
+
{
|
| 3549 |
+
"epoch": 2.88659793814433,
|
| 3550 |
+
"grad_norm": 0.49805002487820915,
|
| 3551 |
+
"learning_rate": 5e-06,
|
| 3552 |
+
"loss": 0.3154,
|
| 3553 |
+
"step": 5040
|
| 3554 |
+
},
|
| 3555 |
+
{
|
| 3556 |
+
"epoch": 2.8923253150057273,
|
| 3557 |
+
"grad_norm": 0.5118585995041246,
|
| 3558 |
+
"learning_rate": 5e-06,
|
| 3559 |
+
"loss": 0.3264,
|
| 3560 |
+
"step": 5050
|
| 3561 |
+
},
|
| 3562 |
+
{
|
| 3563 |
+
"epoch": 2.8980526918671248,
|
| 3564 |
+
"grad_norm": 0.4778628715181792,
|
| 3565 |
+
"learning_rate": 5e-06,
|
| 3566 |
+
"loss": 0.3215,
|
| 3567 |
+
"step": 5060
|
| 3568 |
+
},
|
| 3569 |
+
{
|
| 3570 |
+
"epoch": 2.9037800687285222,
|
| 3571 |
+
"grad_norm": 0.502141421148047,
|
| 3572 |
+
"learning_rate": 5e-06,
|
| 3573 |
+
"loss": 0.3195,
|
| 3574 |
+
"step": 5070
|
| 3575 |
+
},
|
| 3576 |
+
{
|
| 3577 |
+
"epoch": 2.9095074455899197,
|
| 3578 |
+
"grad_norm": 0.5216144225169003,
|
| 3579 |
+
"learning_rate": 5e-06,
|
| 3580 |
+
"loss": 0.322,
|
| 3581 |
+
"step": 5080
|
| 3582 |
+
},
|
| 3583 |
+
{
|
| 3584 |
+
"epoch": 2.915234822451317,
|
| 3585 |
+
"grad_norm": 0.5061317551186302,
|
| 3586 |
+
"learning_rate": 5e-06,
|
| 3587 |
+
"loss": 0.3181,
|
| 3588 |
+
"step": 5090
|
| 3589 |
+
},
|
| 3590 |
+
{
|
| 3591 |
+
"epoch": 2.9209621993127146,
|
| 3592 |
+
"grad_norm": 0.5258296449599809,
|
| 3593 |
+
"learning_rate": 5e-06,
|
| 3594 |
+
"loss": 0.3196,
|
| 3595 |
+
"step": 5100
|
| 3596 |
+
},
|
| 3597 |
+
{
|
| 3598 |
+
"epoch": 2.9266895761741125,
|
| 3599 |
+
"grad_norm": 0.5205129759914494,
|
| 3600 |
+
"learning_rate": 5e-06,
|
| 3601 |
+
"loss": 0.3298,
|
| 3602 |
+
"step": 5110
|
| 3603 |
+
},
|
| 3604 |
+
{
|
| 3605 |
+
"epoch": 2.93241695303551,
|
| 3606 |
+
"grad_norm": 0.5441889203090563,
|
| 3607 |
+
"learning_rate": 5e-06,
|
| 3608 |
+
"loss": 0.3247,
|
| 3609 |
+
"step": 5120
|
| 3610 |
+
},
|
| 3611 |
+
{
|
| 3612 |
+
"epoch": 2.9381443298969074,
|
| 3613 |
+
"grad_norm": 0.5039522474134192,
|
| 3614 |
+
"learning_rate": 5e-06,
|
| 3615 |
+
"loss": 0.3181,
|
| 3616 |
+
"step": 5130
|
| 3617 |
+
},
|
| 3618 |
+
{
|
| 3619 |
+
"epoch": 2.943871706758305,
|
| 3620 |
+
"grad_norm": 0.5232711939948509,
|
| 3621 |
+
"learning_rate": 5e-06,
|
| 3622 |
+
"loss": 0.3196,
|
| 3623 |
+
"step": 5140
|
| 3624 |
+
},
|
| 3625 |
+
{
|
| 3626 |
+
"epoch": 2.9495990836197024,
|
| 3627 |
+
"grad_norm": 0.5119250727874893,
|
| 3628 |
+
"learning_rate": 5e-06,
|
| 3629 |
+
"loss": 0.3131,
|
| 3630 |
+
"step": 5150
|
| 3631 |
+
},
|
| 3632 |
+
{
|
| 3633 |
+
"epoch": 2.9553264604811,
|
| 3634 |
+
"grad_norm": 0.5050585849003654,
|
| 3635 |
+
"learning_rate": 5e-06,
|
| 3636 |
+
"loss": 0.3248,
|
| 3637 |
+
"step": 5160
|
| 3638 |
+
},
|
| 3639 |
+
{
|
| 3640 |
+
"epoch": 2.9610538373424973,
|
| 3641 |
+
"grad_norm": 0.5164002352592711,
|
| 3642 |
+
"learning_rate": 5e-06,
|
| 3643 |
+
"loss": 0.32,
|
| 3644 |
+
"step": 5170
|
| 3645 |
+
},
|
| 3646 |
+
{
|
| 3647 |
+
"epoch": 2.9667812142038947,
|
| 3648 |
+
"grad_norm": 0.5133798037929324,
|
| 3649 |
+
"learning_rate": 5e-06,
|
| 3650 |
+
"loss": 0.3185,
|
| 3651 |
+
"step": 5180
|
| 3652 |
+
},
|
| 3653 |
+
{
|
| 3654 |
+
"epoch": 2.972508591065292,
|
| 3655 |
+
"grad_norm": 0.5131629955273521,
|
| 3656 |
+
"learning_rate": 5e-06,
|
| 3657 |
+
"loss": 0.3169,
|
| 3658 |
+
"step": 5190
|
| 3659 |
+
},
|
| 3660 |
+
{
|
| 3661 |
+
"epoch": 2.9782359679266897,
|
| 3662 |
+
"grad_norm": 0.5354127604897547,
|
| 3663 |
+
"learning_rate": 5e-06,
|
| 3664 |
+
"loss": 0.3265,
|
| 3665 |
+
"step": 5200
|
| 3666 |
+
},
|
| 3667 |
+
{
|
| 3668 |
+
"epoch": 2.983963344788087,
|
| 3669 |
+
"grad_norm": 0.486875785468725,
|
| 3670 |
+
"learning_rate": 5e-06,
|
| 3671 |
+
"loss": 0.3185,
|
| 3672 |
+
"step": 5210
|
| 3673 |
+
},
|
| 3674 |
+
{
|
| 3675 |
+
"epoch": 2.9896907216494846,
|
| 3676 |
+
"grad_norm": 0.524969500670149,
|
| 3677 |
+
"learning_rate": 5e-06,
|
| 3678 |
+
"loss": 0.3161,
|
| 3679 |
+
"step": 5220
|
| 3680 |
+
},
|
| 3681 |
+
{
|
| 3682 |
+
"epoch": 2.995418098510882,
|
| 3683 |
+
"grad_norm": 0.5109832301608017,
|
| 3684 |
+
"learning_rate": 5e-06,
|
| 3685 |
+
"loss": 0.3268,
|
| 3686 |
+
"step": 5230
|
| 3687 |
+
},
|
| 3688 |
+
{
|
| 3689 |
+
"epoch": 3.0,
|
| 3690 |
+
"eval_loss": 0.42792943120002747,
|
| 3691 |
+
"eval_runtime": 443.2838,
|
| 3692 |
+
"eval_samples_per_second": 26.534,
|
| 3693 |
+
"eval_steps_per_second": 0.415,
|
| 3694 |
+
"step": 5238
|
| 3695 |
+
},
|
| 3696 |
+
{
|
| 3697 |
+
"epoch": 3.0,
|
| 3698 |
+
"step": 5238,
|
| 3699 |
+
"total_flos": 2745782658662400.0,
|
| 3700 |
+
"train_loss": 0.3825414228548572,
|
| 3701 |
+
"train_runtime": 71090.721,
|
| 3702 |
+
"train_samples_per_second": 9.431,
|
| 3703 |
+
"train_steps_per_second": 0.074
|
| 3704 |
+
}
|
| 3705 |
+
],
|
| 3706 |
+
"logging_steps": 10,
|
| 3707 |
+
"max_steps": 5238,
|
| 3708 |
+
"num_input_tokens_seen": 0,
|
| 3709 |
+
"num_train_epochs": 3,
|
| 3710 |
+
"save_steps": 500,
|
| 3711 |
+
"stateful_callbacks": {
|
| 3712 |
+
"TrainerControl": {
|
| 3713 |
+
"args": {
|
| 3714 |
+
"should_epoch_stop": false,
|
| 3715 |
+
"should_evaluate": false,
|
| 3716 |
+
"should_log": false,
|
| 3717 |
+
"should_save": true,
|
| 3718 |
+
"should_training_stop": true
|
| 3719 |
+
},
|
| 3720 |
+
"attributes": {}
|
| 3721 |
+
}
|
| 3722 |
+
},
|
| 3723 |
+
"total_flos": 2745782658662400.0,
|
| 3724 |
+
"train_batch_size": 8,
|
| 3725 |
+
"trial_name": null,
|
| 3726 |
+
"trial_params": null
|
| 3727 |
+
}
|
training_eval_loss.png
ADDED
|
training_loss.png
ADDED
|