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 +962 -0
- training_eval_loss.png +0 -0
- training_loss.png +0 -0
README.md
CHANGED
|
@@ -4,6 +4,7 @@ license: llama3.1
|
|
| 4 |
base_model: meta-llama/Meta-Llama-3.1-8B
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
|
|
|
| 7 |
- generated_from_trainer
|
| 8 |
model-index:
|
| 9 |
- name: llama3-1_8b_math_500000_samples
|
|
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
# llama3-1_8b_math_500000_samples
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B) on
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
- Loss: 0.5069
|
| 21 |
|
|
|
|
| 4 |
base_model: meta-llama/Meta-Llama-3.1-8B
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
| 7 |
+
- full
|
| 8 |
- generated_from_trainer
|
| 9 |
model-index:
|
| 10 |
- name: llama3-1_8b_math_500000_samples
|
|
|
|
| 16 |
|
| 17 |
# llama3-1_8b_math_500000_samples
|
| 18 |
|
| 19 |
+
This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B) on the mlfoundations-dev/math_500000_samples dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
- Loss: 0.5069
|
| 22 |
|
all_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_loss": 0.5069288611412048,
|
| 4 |
+
"eval_runtime": 41.6306,
|
| 5 |
+
"eval_samples_per_second": 276.311,
|
| 6 |
+
"eval_steps_per_second": 1.081,
|
| 7 |
+
"total_flos": 2145722711408640.0,
|
| 8 |
+
"train_loss": 0.4994005665799587,
|
| 9 |
+
"train_runtime": 8118.1452,
|
| 10 |
+
"train_samples_per_second": 80.761,
|
| 11 |
+
"train_steps_per_second": 0.158
|
| 12 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_loss": 0.5069288611412048,
|
| 4 |
+
"eval_runtime": 41.6306,
|
| 5 |
+
"eval_samples_per_second": 276.311,
|
| 6 |
+
"eval_steps_per_second": 1.081
|
| 7 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 2145722711408640.0,
|
| 4 |
+
"train_loss": 0.4994005665799587,
|
| 5 |
+
"train_runtime": 8118.1452,
|
| 6 |
+
"train_samples_per_second": 80.761,
|
| 7 |
+
"train_steps_per_second": 0.158
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,962 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1281,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0234192037470726,
|
| 13 |
+
"grad_norm": 2.481647447902812,
|
| 14 |
+
"learning_rate": 5e-06,
|
| 15 |
+
"loss": 0.7124,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0468384074941452,
|
| 20 |
+
"grad_norm": 0.7262736099420435,
|
| 21 |
+
"learning_rate": 5e-06,
|
| 22 |
+
"loss": 0.6153,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.0702576112412178,
|
| 27 |
+
"grad_norm": 0.7790978795240339,
|
| 28 |
+
"learning_rate": 5e-06,
|
| 29 |
+
"loss": 0.5859,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.0936768149882904,
|
| 34 |
+
"grad_norm": 0.6467962423003429,
|
| 35 |
+
"learning_rate": 5e-06,
|
| 36 |
+
"loss": 0.5711,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.117096018735363,
|
| 41 |
+
"grad_norm": 0.5143977117058031,
|
| 42 |
+
"learning_rate": 5e-06,
|
| 43 |
+
"loss": 0.5668,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.1405152224824356,
|
| 48 |
+
"grad_norm": 0.600469324649081,
|
| 49 |
+
"learning_rate": 5e-06,
|
| 50 |
+
"loss": 0.559,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.16393442622950818,
|
| 55 |
+
"grad_norm": 0.5054683279603859,
|
| 56 |
+
"learning_rate": 5e-06,
|
| 57 |
+
"loss": 0.555,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.1873536299765808,
|
| 62 |
+
"grad_norm": 0.520518786315232,
|
| 63 |
+
"learning_rate": 5e-06,
|
| 64 |
+
"loss": 0.548,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.2107728337236534,
|
| 69 |
+
"grad_norm": 1.008487324396366,
|
| 70 |
+
"learning_rate": 5e-06,
|
| 71 |
+
"loss": 0.5481,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.234192037470726,
|
| 76 |
+
"grad_norm": 0.6357300755902857,
|
| 77 |
+
"learning_rate": 5e-06,
|
| 78 |
+
"loss": 0.5449,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.2576112412177986,
|
| 83 |
+
"grad_norm": 0.4471586394958172,
|
| 84 |
+
"learning_rate": 5e-06,
|
| 85 |
+
"loss": 0.541,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.2810304449648712,
|
| 90 |
+
"grad_norm": 0.42874334693852684,
|
| 91 |
+
"learning_rate": 5e-06,
|
| 92 |
+
"loss": 0.5397,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.3044496487119438,
|
| 97 |
+
"grad_norm": 0.5493546613118618,
|
| 98 |
+
"learning_rate": 5e-06,
|
| 99 |
+
"loss": 0.5357,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.32786885245901637,
|
| 104 |
+
"grad_norm": 0.7309908680131434,
|
| 105 |
+
"learning_rate": 5e-06,
|
| 106 |
+
"loss": 0.535,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.351288056206089,
|
| 111 |
+
"grad_norm": 0.5231239547569523,
|
| 112 |
+
"learning_rate": 5e-06,
|
| 113 |
+
"loss": 0.5373,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.3747072599531616,
|
| 118 |
+
"grad_norm": 0.47995952741525777,
|
| 119 |
+
"learning_rate": 5e-06,
|
| 120 |
+
"loss": 0.5346,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.3981264637002342,
|
| 125 |
+
"grad_norm": 0.562861554643795,
|
| 126 |
+
"learning_rate": 5e-06,
|
| 127 |
+
"loss": 0.5312,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.4215456674473068,
|
| 132 |
+
"grad_norm": 0.6323157343840412,
|
| 133 |
+
"learning_rate": 5e-06,
|
| 134 |
+
"loss": 0.5332,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.4449648711943794,
|
| 139 |
+
"grad_norm": 0.6022835533604722,
|
| 140 |
+
"learning_rate": 5e-06,
|
| 141 |
+
"loss": 0.5368,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.468384074941452,
|
| 146 |
+
"grad_norm": 0.5241376216453505,
|
| 147 |
+
"learning_rate": 5e-06,
|
| 148 |
+
"loss": 0.5348,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.4918032786885246,
|
| 153 |
+
"grad_norm": 0.4560855929707371,
|
| 154 |
+
"learning_rate": 5e-06,
|
| 155 |
+
"loss": 0.5298,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.5152224824355972,
|
| 160 |
+
"grad_norm": 0.4775160386802804,
|
| 161 |
+
"learning_rate": 5e-06,
|
| 162 |
+
"loss": 0.527,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.5386416861826698,
|
| 167 |
+
"grad_norm": 0.5062169069603658,
|
| 168 |
+
"learning_rate": 5e-06,
|
| 169 |
+
"loss": 0.5295,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.5620608899297423,
|
| 174 |
+
"grad_norm": 0.43311731742710935,
|
| 175 |
+
"learning_rate": 5e-06,
|
| 176 |
+
"loss": 0.5264,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.585480093676815,
|
| 181 |
+
"grad_norm": 0.4911544553586758,
|
| 182 |
+
"learning_rate": 5e-06,
|
| 183 |
+
"loss": 0.5292,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.6088992974238876,
|
| 188 |
+
"grad_norm": 0.5749030738074241,
|
| 189 |
+
"learning_rate": 5e-06,
|
| 190 |
+
"loss": 0.5253,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.6323185011709602,
|
| 195 |
+
"grad_norm": 0.5580576917700846,
|
| 196 |
+
"learning_rate": 5e-06,
|
| 197 |
+
"loss": 0.5238,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.6557377049180327,
|
| 202 |
+
"grad_norm": 0.4760038513178861,
|
| 203 |
+
"learning_rate": 5e-06,
|
| 204 |
+
"loss": 0.522,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.6791569086651054,
|
| 209 |
+
"grad_norm": 0.5216361788356643,
|
| 210 |
+
"learning_rate": 5e-06,
|
| 211 |
+
"loss": 0.5243,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.702576112412178,
|
| 216 |
+
"grad_norm": 0.46276681576162904,
|
| 217 |
+
"learning_rate": 5e-06,
|
| 218 |
+
"loss": 0.5217,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.7259953161592506,
|
| 223 |
+
"grad_norm": 0.4747845254798478,
|
| 224 |
+
"learning_rate": 5e-06,
|
| 225 |
+
"loss": 0.5226,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.7494145199063232,
|
| 230 |
+
"grad_norm": 0.5960080896982851,
|
| 231 |
+
"learning_rate": 5e-06,
|
| 232 |
+
"loss": 0.521,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.7728337236533958,
|
| 237 |
+
"grad_norm": 0.5212901742846952,
|
| 238 |
+
"learning_rate": 5e-06,
|
| 239 |
+
"loss": 0.5201,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.7962529274004684,
|
| 244 |
+
"grad_norm": 0.5133380400247927,
|
| 245 |
+
"learning_rate": 5e-06,
|
| 246 |
+
"loss": 0.5199,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.819672131147541,
|
| 251 |
+
"grad_norm": 0.4368489322167074,
|
| 252 |
+
"learning_rate": 5e-06,
|
| 253 |
+
"loss": 0.5204,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.8430913348946136,
|
| 258 |
+
"grad_norm": 0.5101656467162784,
|
| 259 |
+
"learning_rate": 5e-06,
|
| 260 |
+
"loss": 0.5221,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.8665105386416861,
|
| 265 |
+
"grad_norm": 0.5226647593757339,
|
| 266 |
+
"learning_rate": 5e-06,
|
| 267 |
+
"loss": 0.519,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.8899297423887588,
|
| 272 |
+
"grad_norm": 0.4265472443674912,
|
| 273 |
+
"learning_rate": 5e-06,
|
| 274 |
+
"loss": 0.5193,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.9133489461358314,
|
| 279 |
+
"grad_norm": 0.42184387318773725,
|
| 280 |
+
"learning_rate": 5e-06,
|
| 281 |
+
"loss": 0.5178,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.936768149882904,
|
| 286 |
+
"grad_norm": 0.4773759814767807,
|
| 287 |
+
"learning_rate": 5e-06,
|
| 288 |
+
"loss": 0.5201,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.9601873536299765,
|
| 293 |
+
"grad_norm": 0.5049194555570107,
|
| 294 |
+
"learning_rate": 5e-06,
|
| 295 |
+
"loss": 0.5146,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.9836065573770492,
|
| 300 |
+
"grad_norm": 0.4836846129062704,
|
| 301 |
+
"learning_rate": 5e-06,
|
| 302 |
+
"loss": 0.514,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 1.0,
|
| 307 |
+
"eval_loss": 0.5167434215545654,
|
| 308 |
+
"eval_runtime": 41.6336,
|
| 309 |
+
"eval_samples_per_second": 276.291,
|
| 310 |
+
"eval_steps_per_second": 1.081,
|
| 311 |
+
"step": 427
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 1.0070257611241218,
|
| 315 |
+
"grad_norm": 0.4520838332937754,
|
| 316 |
+
"learning_rate": 5e-06,
|
| 317 |
+
"loss": 0.51,
|
| 318 |
+
"step": 430
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 1.0304449648711944,
|
| 322 |
+
"grad_norm": 0.4866921792731558,
|
| 323 |
+
"learning_rate": 5e-06,
|
| 324 |
+
"loss": 0.4977,
|
| 325 |
+
"step": 440
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 1.053864168618267,
|
| 329 |
+
"grad_norm": 0.46612714801971344,
|
| 330 |
+
"learning_rate": 5e-06,
|
| 331 |
+
"loss": 0.4974,
|
| 332 |
+
"step": 450
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.0772833723653397,
|
| 336 |
+
"grad_norm": 0.47487180548015817,
|
| 337 |
+
"learning_rate": 5e-06,
|
| 338 |
+
"loss": 0.4947,
|
| 339 |
+
"step": 460
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 1.100702576112412,
|
| 343 |
+
"grad_norm": 0.5023694335040125,
|
| 344 |
+
"learning_rate": 5e-06,
|
| 345 |
+
"loss": 0.4909,
|
| 346 |
+
"step": 470
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 1.1241217798594847,
|
| 350 |
+
"grad_norm": 0.5257140618717152,
|
| 351 |
+
"learning_rate": 5e-06,
|
| 352 |
+
"loss": 0.4932,
|
| 353 |
+
"step": 480
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.1475409836065573,
|
| 357 |
+
"grad_norm": 0.39736116687044365,
|
| 358 |
+
"learning_rate": 5e-06,
|
| 359 |
+
"loss": 0.4894,
|
| 360 |
+
"step": 490
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 1.17096018735363,
|
| 364 |
+
"grad_norm": 0.42753669693798274,
|
| 365 |
+
"learning_rate": 5e-06,
|
| 366 |
+
"loss": 0.4943,
|
| 367 |
+
"step": 500
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 1.1943793911007026,
|
| 371 |
+
"grad_norm": 0.4539073963232022,
|
| 372 |
+
"learning_rate": 5e-06,
|
| 373 |
+
"loss": 0.497,
|
| 374 |
+
"step": 510
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.2177985948477752,
|
| 378 |
+
"grad_norm": 0.40221054966287245,
|
| 379 |
+
"learning_rate": 5e-06,
|
| 380 |
+
"loss": 0.4916,
|
| 381 |
+
"step": 520
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 1.2412177985948478,
|
| 385 |
+
"grad_norm": 0.40071702689512967,
|
| 386 |
+
"learning_rate": 5e-06,
|
| 387 |
+
"loss": 0.4904,
|
| 388 |
+
"step": 530
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 1.2646370023419204,
|
| 392 |
+
"grad_norm": 0.4914284242936407,
|
| 393 |
+
"learning_rate": 5e-06,
|
| 394 |
+
"loss": 0.4941,
|
| 395 |
+
"step": 540
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.288056206088993,
|
| 399 |
+
"grad_norm": 0.4409757840452587,
|
| 400 |
+
"learning_rate": 5e-06,
|
| 401 |
+
"loss": 0.4924,
|
| 402 |
+
"step": 550
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 1.3114754098360657,
|
| 406 |
+
"grad_norm": 0.4137810147419586,
|
| 407 |
+
"learning_rate": 5e-06,
|
| 408 |
+
"loss": 0.4925,
|
| 409 |
+
"step": 560
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 1.334894613583138,
|
| 413 |
+
"grad_norm": 0.48726970931068914,
|
| 414 |
+
"learning_rate": 5e-06,
|
| 415 |
+
"loss": 0.4943,
|
| 416 |
+
"step": 570
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 1.3583138173302107,
|
| 420 |
+
"grad_norm": 0.5494057601615131,
|
| 421 |
+
"learning_rate": 5e-06,
|
| 422 |
+
"loss": 0.4935,
|
| 423 |
+
"step": 580
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 1.3817330210772834,
|
| 427 |
+
"grad_norm": 0.41251177208283474,
|
| 428 |
+
"learning_rate": 5e-06,
|
| 429 |
+
"loss": 0.4956,
|
| 430 |
+
"step": 590
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 1.405152224824356,
|
| 434 |
+
"grad_norm": 0.39118790995394576,
|
| 435 |
+
"learning_rate": 5e-06,
|
| 436 |
+
"loss": 0.4883,
|
| 437 |
+
"step": 600
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.4285714285714286,
|
| 441 |
+
"grad_norm": 0.49321602369191925,
|
| 442 |
+
"learning_rate": 5e-06,
|
| 443 |
+
"loss": 0.496,
|
| 444 |
+
"step": 610
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 1.4519906323185012,
|
| 448 |
+
"grad_norm": 0.5208838513759695,
|
| 449 |
+
"learning_rate": 5e-06,
|
| 450 |
+
"loss": 0.4953,
|
| 451 |
+
"step": 620
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 1.4754098360655736,
|
| 455 |
+
"grad_norm": 0.43794255642976027,
|
| 456 |
+
"learning_rate": 5e-06,
|
| 457 |
+
"loss": 0.4951,
|
| 458 |
+
"step": 630
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 1.4988290398126463,
|
| 462 |
+
"grad_norm": 0.5557497245301519,
|
| 463 |
+
"learning_rate": 5e-06,
|
| 464 |
+
"loss": 0.4929,
|
| 465 |
+
"step": 640
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 1.5222482435597189,
|
| 469 |
+
"grad_norm": 0.5087889678152296,
|
| 470 |
+
"learning_rate": 5e-06,
|
| 471 |
+
"loss": 0.4896,
|
| 472 |
+
"step": 650
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 1.5456674473067915,
|
| 476 |
+
"grad_norm": 0.38311143724497726,
|
| 477 |
+
"learning_rate": 5e-06,
|
| 478 |
+
"loss": 0.4928,
|
| 479 |
+
"step": 660
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.5690866510538641,
|
| 483 |
+
"grad_norm": 0.48635422306380777,
|
| 484 |
+
"learning_rate": 5e-06,
|
| 485 |
+
"loss": 0.488,
|
| 486 |
+
"step": 670
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 1.5925058548009368,
|
| 490 |
+
"grad_norm": 0.6136317892186699,
|
| 491 |
+
"learning_rate": 5e-06,
|
| 492 |
+
"loss": 0.4949,
|
| 493 |
+
"step": 680
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 1.6159250585480094,
|
| 497 |
+
"grad_norm": 0.47270773640049263,
|
| 498 |
+
"learning_rate": 5e-06,
|
| 499 |
+
"loss": 0.4883,
|
| 500 |
+
"step": 690
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 1.639344262295082,
|
| 504 |
+
"grad_norm": 0.5783159568227515,
|
| 505 |
+
"learning_rate": 5e-06,
|
| 506 |
+
"loss": 0.4936,
|
| 507 |
+
"step": 700
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 1.6627634660421546,
|
| 511 |
+
"grad_norm": 0.43336426848256004,
|
| 512 |
+
"learning_rate": 5e-06,
|
| 513 |
+
"loss": 0.4938,
|
| 514 |
+
"step": 710
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 1.6861826697892273,
|
| 518 |
+
"grad_norm": 0.4314026408601509,
|
| 519 |
+
"learning_rate": 5e-06,
|
| 520 |
+
"loss": 0.4967,
|
| 521 |
+
"step": 720
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 1.7096018735362999,
|
| 525 |
+
"grad_norm": 0.38409075682155797,
|
| 526 |
+
"learning_rate": 5e-06,
|
| 527 |
+
"loss": 0.4891,
|
| 528 |
+
"step": 730
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 1.7330210772833725,
|
| 532 |
+
"grad_norm": 0.5211936246593988,
|
| 533 |
+
"learning_rate": 5e-06,
|
| 534 |
+
"loss": 0.491,
|
| 535 |
+
"step": 740
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 1.756440281030445,
|
| 539 |
+
"grad_norm": 0.46808950576809366,
|
| 540 |
+
"learning_rate": 5e-06,
|
| 541 |
+
"loss": 0.4949,
|
| 542 |
+
"step": 750
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 1.7798594847775175,
|
| 546 |
+
"grad_norm": 0.3970884421781967,
|
| 547 |
+
"learning_rate": 5e-06,
|
| 548 |
+
"loss": 0.4864,
|
| 549 |
+
"step": 760
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 1.8032786885245902,
|
| 553 |
+
"grad_norm": 0.4338133573209166,
|
| 554 |
+
"learning_rate": 5e-06,
|
| 555 |
+
"loss": 0.4877,
|
| 556 |
+
"step": 770
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 1.8266978922716628,
|
| 560 |
+
"grad_norm": 0.5302929743875919,
|
| 561 |
+
"learning_rate": 5e-06,
|
| 562 |
+
"loss": 0.491,
|
| 563 |
+
"step": 780
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 1.8501170960187352,
|
| 567 |
+
"grad_norm": 0.48309786628518675,
|
| 568 |
+
"learning_rate": 5e-06,
|
| 569 |
+
"loss": 0.4919,
|
| 570 |
+
"step": 790
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 1.8735362997658078,
|
| 574 |
+
"grad_norm": 0.44165745399512696,
|
| 575 |
+
"learning_rate": 5e-06,
|
| 576 |
+
"loss": 0.4875,
|
| 577 |
+
"step": 800
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 1.8969555035128804,
|
| 581 |
+
"grad_norm": 0.46223180789120477,
|
| 582 |
+
"learning_rate": 5e-06,
|
| 583 |
+
"loss": 0.4936,
|
| 584 |
+
"step": 810
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 1.920374707259953,
|
| 588 |
+
"grad_norm": 0.48226738955985293,
|
| 589 |
+
"learning_rate": 5e-06,
|
| 590 |
+
"loss": 0.4888,
|
| 591 |
+
"step": 820
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 1.9437939110070257,
|
| 595 |
+
"grad_norm": 0.39089405316919673,
|
| 596 |
+
"learning_rate": 5e-06,
|
| 597 |
+
"loss": 0.4893,
|
| 598 |
+
"step": 830
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 1.9672131147540983,
|
| 602 |
+
"grad_norm": 0.3967255297864678,
|
| 603 |
+
"learning_rate": 5e-06,
|
| 604 |
+
"loss": 0.4915,
|
| 605 |
+
"step": 840
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 1.990632318501171,
|
| 609 |
+
"grad_norm": 0.4053292002238054,
|
| 610 |
+
"learning_rate": 5e-06,
|
| 611 |
+
"loss": 0.4902,
|
| 612 |
+
"step": 850
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 2.0,
|
| 616 |
+
"eval_loss": 0.5077147483825684,
|
| 617 |
+
"eval_runtime": 41.7501,
|
| 618 |
+
"eval_samples_per_second": 275.521,
|
| 619 |
+
"eval_steps_per_second": 1.078,
|
| 620 |
+
"step": 854
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 2.0140515222482436,
|
| 624 |
+
"grad_norm": 0.5268697371178085,
|
| 625 |
+
"learning_rate": 5e-06,
|
| 626 |
+
"loss": 0.4752,
|
| 627 |
+
"step": 860
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 2.037470725995316,
|
| 631 |
+
"grad_norm": 0.5447609017298904,
|
| 632 |
+
"learning_rate": 5e-06,
|
| 633 |
+
"loss": 0.4669,
|
| 634 |
+
"step": 870
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 2.060889929742389,
|
| 638 |
+
"grad_norm": 0.5637527273857,
|
| 639 |
+
"learning_rate": 5e-06,
|
| 640 |
+
"loss": 0.4614,
|
| 641 |
+
"step": 880
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 2.0843091334894615,
|
| 645 |
+
"grad_norm": 0.3841529769069848,
|
| 646 |
+
"learning_rate": 5e-06,
|
| 647 |
+
"loss": 0.4659,
|
| 648 |
+
"step": 890
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 2.107728337236534,
|
| 652 |
+
"grad_norm": 0.42444670935905104,
|
| 653 |
+
"learning_rate": 5e-06,
|
| 654 |
+
"loss": 0.4593,
|
| 655 |
+
"step": 900
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 2.1311475409836067,
|
| 659 |
+
"grad_norm": 0.4111116772372611,
|
| 660 |
+
"learning_rate": 5e-06,
|
| 661 |
+
"loss": 0.4677,
|
| 662 |
+
"step": 910
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 2.1545667447306793,
|
| 666 |
+
"grad_norm": 0.6373094219643812,
|
| 667 |
+
"learning_rate": 5e-06,
|
| 668 |
+
"loss": 0.4652,
|
| 669 |
+
"step": 920
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 2.177985948477752,
|
| 673 |
+
"grad_norm": 0.6160202621538906,
|
| 674 |
+
"learning_rate": 5e-06,
|
| 675 |
+
"loss": 0.4681,
|
| 676 |
+
"step": 930
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 2.201405152224824,
|
| 680 |
+
"grad_norm": 0.37701921381345616,
|
| 681 |
+
"learning_rate": 5e-06,
|
| 682 |
+
"loss": 0.4586,
|
| 683 |
+
"step": 940
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 2.2248243559718968,
|
| 687 |
+
"grad_norm": 0.4220733688149805,
|
| 688 |
+
"learning_rate": 5e-06,
|
| 689 |
+
"loss": 0.4669,
|
| 690 |
+
"step": 950
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 2.2482435597189694,
|
| 694 |
+
"grad_norm": 0.5051333977952606,
|
| 695 |
+
"learning_rate": 5e-06,
|
| 696 |
+
"loss": 0.4665,
|
| 697 |
+
"step": 960
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 2.271662763466042,
|
| 701 |
+
"grad_norm": 0.45928248167044633,
|
| 702 |
+
"learning_rate": 5e-06,
|
| 703 |
+
"loss": 0.4665,
|
| 704 |
+
"step": 970
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 2.2950819672131146,
|
| 708 |
+
"grad_norm": 0.4495617426425146,
|
| 709 |
+
"learning_rate": 5e-06,
|
| 710 |
+
"loss": 0.4662,
|
| 711 |
+
"step": 980
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 2.3185011709601873,
|
| 715 |
+
"grad_norm": 0.4517571833465374,
|
| 716 |
+
"learning_rate": 5e-06,
|
| 717 |
+
"loss": 0.4646,
|
| 718 |
+
"step": 990
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 2.34192037470726,
|
| 722 |
+
"grad_norm": 0.4138564151461986,
|
| 723 |
+
"learning_rate": 5e-06,
|
| 724 |
+
"loss": 0.4647,
|
| 725 |
+
"step": 1000
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 2.3653395784543325,
|
| 729 |
+
"grad_norm": 0.5738034318601656,
|
| 730 |
+
"learning_rate": 5e-06,
|
| 731 |
+
"loss": 0.4659,
|
| 732 |
+
"step": 1010
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 2.388758782201405,
|
| 736 |
+
"grad_norm": 0.631953187344588,
|
| 737 |
+
"learning_rate": 5e-06,
|
| 738 |
+
"loss": 0.4663,
|
| 739 |
+
"step": 1020
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 2.4121779859484778,
|
| 743 |
+
"grad_norm": 0.46549566538492926,
|
| 744 |
+
"learning_rate": 5e-06,
|
| 745 |
+
"loss": 0.4676,
|
| 746 |
+
"step": 1030
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 2.4355971896955504,
|
| 750 |
+
"grad_norm": 0.43668516137249047,
|
| 751 |
+
"learning_rate": 5e-06,
|
| 752 |
+
"loss": 0.4673,
|
| 753 |
+
"step": 1040
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 2.459016393442623,
|
| 757 |
+
"grad_norm": 0.42948375636759983,
|
| 758 |
+
"learning_rate": 5e-06,
|
| 759 |
+
"loss": 0.4654,
|
| 760 |
+
"step": 1050
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 2.4824355971896956,
|
| 764 |
+
"grad_norm": 0.4391102312189122,
|
| 765 |
+
"learning_rate": 5e-06,
|
| 766 |
+
"loss": 0.469,
|
| 767 |
+
"step": 1060
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 2.5058548009367683,
|
| 771 |
+
"grad_norm": 0.49842008382902114,
|
| 772 |
+
"learning_rate": 5e-06,
|
| 773 |
+
"loss": 0.47,
|
| 774 |
+
"step": 1070
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 2.529274004683841,
|
| 778 |
+
"grad_norm": 0.5588818151397613,
|
| 779 |
+
"learning_rate": 5e-06,
|
| 780 |
+
"loss": 0.4685,
|
| 781 |
+
"step": 1080
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 2.552693208430913,
|
| 785 |
+
"grad_norm": 0.4037726802957124,
|
| 786 |
+
"learning_rate": 5e-06,
|
| 787 |
+
"loss": 0.4677,
|
| 788 |
+
"step": 1090
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 2.576112412177986,
|
| 792 |
+
"grad_norm": 0.48821405764171133,
|
| 793 |
+
"learning_rate": 5e-06,
|
| 794 |
+
"loss": 0.4695,
|
| 795 |
+
"step": 1100
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 2.5995316159250583,
|
| 799 |
+
"grad_norm": 0.3858402261553971,
|
| 800 |
+
"learning_rate": 5e-06,
|
| 801 |
+
"loss": 0.4677,
|
| 802 |
+
"step": 1110
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 2.6229508196721314,
|
| 806 |
+
"grad_norm": 0.4203466943287186,
|
| 807 |
+
"learning_rate": 5e-06,
|
| 808 |
+
"loss": 0.4696,
|
| 809 |
+
"step": 1120
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 2.6463700234192036,
|
| 813 |
+
"grad_norm": 0.48915413911853883,
|
| 814 |
+
"learning_rate": 5e-06,
|
| 815 |
+
"loss": 0.4664,
|
| 816 |
+
"step": 1130
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 2.669789227166276,
|
| 820 |
+
"grad_norm": 0.43449855854006547,
|
| 821 |
+
"learning_rate": 5e-06,
|
| 822 |
+
"loss": 0.4693,
|
| 823 |
+
"step": 1140
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 2.693208430913349,
|
| 827 |
+
"grad_norm": 0.5093894446854779,
|
| 828 |
+
"learning_rate": 5e-06,
|
| 829 |
+
"loss": 0.4692,
|
| 830 |
+
"step": 1150
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 2.7166276346604215,
|
| 834 |
+
"grad_norm": 0.44982206423576104,
|
| 835 |
+
"learning_rate": 5e-06,
|
| 836 |
+
"loss": 0.4704,
|
| 837 |
+
"step": 1160
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 2.740046838407494,
|
| 841 |
+
"grad_norm": 0.40753147448920485,
|
| 842 |
+
"learning_rate": 5e-06,
|
| 843 |
+
"loss": 0.4648,
|
| 844 |
+
"step": 1170
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 2.7634660421545667,
|
| 848 |
+
"grad_norm": 0.36811758162537594,
|
| 849 |
+
"learning_rate": 5e-06,
|
| 850 |
+
"loss": 0.4629,
|
| 851 |
+
"step": 1180
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 2.7868852459016393,
|
| 855 |
+
"grad_norm": 0.3949349460619723,
|
| 856 |
+
"learning_rate": 5e-06,
|
| 857 |
+
"loss": 0.4704,
|
| 858 |
+
"step": 1190
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 2.810304449648712,
|
| 862 |
+
"grad_norm": 0.41402013254949793,
|
| 863 |
+
"learning_rate": 5e-06,
|
| 864 |
+
"loss": 0.4707,
|
| 865 |
+
"step": 1200
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 2.8337236533957846,
|
| 869 |
+
"grad_norm": 0.5432806369874894,
|
| 870 |
+
"learning_rate": 5e-06,
|
| 871 |
+
"loss": 0.4679,
|
| 872 |
+
"step": 1210
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 2.857142857142857,
|
| 876 |
+
"grad_norm": 0.3788514557857583,
|
| 877 |
+
"learning_rate": 5e-06,
|
| 878 |
+
"loss": 0.4711,
|
| 879 |
+
"step": 1220
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 2.88056206088993,
|
| 883 |
+
"grad_norm": 0.39636258806401947,
|
| 884 |
+
"learning_rate": 5e-06,
|
| 885 |
+
"loss": 0.4653,
|
| 886 |
+
"step": 1230
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 2.9039812646370025,
|
| 890 |
+
"grad_norm": 0.3936100271064975,
|
| 891 |
+
"learning_rate": 5e-06,
|
| 892 |
+
"loss": 0.4667,
|
| 893 |
+
"step": 1240
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 2.927400468384075,
|
| 897 |
+
"grad_norm": 0.4271813759224886,
|
| 898 |
+
"learning_rate": 5e-06,
|
| 899 |
+
"loss": 0.4664,
|
| 900 |
+
"step": 1250
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 2.9508196721311473,
|
| 904 |
+
"grad_norm": 0.4720023108766376,
|
| 905 |
+
"learning_rate": 5e-06,
|
| 906 |
+
"loss": 0.468,
|
| 907 |
+
"step": 1260
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 2.9742388758782203,
|
| 911 |
+
"grad_norm": 0.42625818049428776,
|
| 912 |
+
"learning_rate": 5e-06,
|
| 913 |
+
"loss": 0.4654,
|
| 914 |
+
"step": 1270
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 2.9976580796252925,
|
| 918 |
+
"grad_norm": 0.4282654656901964,
|
| 919 |
+
"learning_rate": 5e-06,
|
| 920 |
+
"loss": 0.47,
|
| 921 |
+
"step": 1280
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 3.0,
|
| 925 |
+
"eval_loss": 0.5069288611412048,
|
| 926 |
+
"eval_runtime": 41.6671,
|
| 927 |
+
"eval_samples_per_second": 276.069,
|
| 928 |
+
"eval_steps_per_second": 1.08,
|
| 929 |
+
"step": 1281
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 3.0,
|
| 933 |
+
"step": 1281,
|
| 934 |
+
"total_flos": 2145722711408640.0,
|
| 935 |
+
"train_loss": 0.4994005665799587,
|
| 936 |
+
"train_runtime": 8118.1452,
|
| 937 |
+
"train_samples_per_second": 80.761,
|
| 938 |
+
"train_steps_per_second": 0.158
|
| 939 |
+
}
|
| 940 |
+
],
|
| 941 |
+
"logging_steps": 10,
|
| 942 |
+
"max_steps": 1281,
|
| 943 |
+
"num_input_tokens_seen": 0,
|
| 944 |
+
"num_train_epochs": 3,
|
| 945 |
+
"save_steps": 500,
|
| 946 |
+
"stateful_callbacks": {
|
| 947 |
+
"TrainerControl": {
|
| 948 |
+
"args": {
|
| 949 |
+
"should_epoch_stop": false,
|
| 950 |
+
"should_evaluate": false,
|
| 951 |
+
"should_log": false,
|
| 952 |
+
"should_save": true,
|
| 953 |
+
"should_training_stop": true
|
| 954 |
+
},
|
| 955 |
+
"attributes": {}
|
| 956 |
+
}
|
| 957 |
+
},
|
| 958 |
+
"total_flos": 2145722711408640.0,
|
| 959 |
+
"train_batch_size": 16,
|
| 960 |
+
"trial_name": null,
|
| 961 |
+
"trial_params": null
|
| 962 |
+
}
|
training_eval_loss.png
ADDED
|
training_loss.png
ADDED
|