End of training
Browse files- README.md +2 -1
- all_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +2205 -0
- training_loss.png +0 -0
README.md
CHANGED
|
@@ -4,6 +4,7 @@ license: mit
|
|
| 4 |
base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-7B
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
|
|
|
| 7 |
- generated_from_trainer
|
| 8 |
model-index:
|
| 9 |
- name: VD-DS-DSFormat-Clean-8k_DeepSeek-R1-Distill-Qwen-7B_full_sft_1e-5
|
|
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
# VD-DS-DSFormat-Clean-8k_DeepSeek-R1-Distill-Qwen-7B_full_sft_1e-5
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [deepseek-ai/DeepSeek-R1-Distill-Qwen-7B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B) on
|
| 19 |
|
| 20 |
## Model description
|
| 21 |
|
|
|
|
| 4 |
base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-7B
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
| 7 |
+
- full
|
| 8 |
- generated_from_trainer
|
| 9 |
model-index:
|
| 10 |
- name: VD-DS-DSFormat-Clean-8k_DeepSeek-R1-Distill-Qwen-7B_full_sft_1e-5
|
|
|
|
| 16 |
|
| 17 |
# VD-DS-DSFormat-Clean-8k_DeepSeek-R1-Distill-Qwen-7B_full_sft_1e-5
|
| 18 |
|
| 19 |
+
This model is a fine-tuned version of [deepseek-ai/DeepSeek-R1-Distill-Qwen-7B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B) on the VD-DS-DSFormat-Clean-8k dataset.
|
| 20 |
|
| 21 |
## Model description
|
| 22 |
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.9734939759036143,
|
| 3 |
+
"total_flos": 1.0191470808845517e+17,
|
| 4 |
+
"train_loss": 0.6618778399857889,
|
| 5 |
+
"train_runtime": 3030.6803,
|
| 6 |
+
"train_samples_per_second": 4.928,
|
| 7 |
+
"train_steps_per_second": 0.102
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.9734939759036143,
|
| 3 |
+
"total_flos": 1.0191470808845517e+17,
|
| 4 |
+
"train_loss": 0.6618778399857889,
|
| 5 |
+
"train_runtime": 3030.6803,
|
| 6 |
+
"train_samples_per_second": 4.928,
|
| 7 |
+
"train_steps_per_second": 0.102
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2205 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.9734939759036143,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 309,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.00963855421686747,
|
| 13 |
+
"grad_norm": 0.7321725047335572,
|
| 14 |
+
"learning_rate": 3.2258064516129035e-07,
|
| 15 |
+
"loss": 1.2207,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.01927710843373494,
|
| 20 |
+
"grad_norm": 0.6665264239249077,
|
| 21 |
+
"learning_rate": 6.451612903225807e-07,
|
| 22 |
+
"loss": 1.1371,
|
| 23 |
+
"step": 2
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.02891566265060241,
|
| 27 |
+
"grad_norm": 0.7664904235774216,
|
| 28 |
+
"learning_rate": 9.67741935483871e-07,
|
| 29 |
+
"loss": 1.1605,
|
| 30 |
+
"step": 3
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.03855421686746988,
|
| 34 |
+
"grad_norm": 0.8973293033776428,
|
| 35 |
+
"learning_rate": 1.2903225806451614e-06,
|
| 36 |
+
"loss": 1.297,
|
| 37 |
+
"step": 4
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.04819277108433735,
|
| 41 |
+
"grad_norm": 0.7512289668671644,
|
| 42 |
+
"learning_rate": 1.6129032258064516e-06,
|
| 43 |
+
"loss": 1.2186,
|
| 44 |
+
"step": 5
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.05783132530120482,
|
| 48 |
+
"grad_norm": 0.5784006362716563,
|
| 49 |
+
"learning_rate": 1.935483870967742e-06,
|
| 50 |
+
"loss": 1.1326,
|
| 51 |
+
"step": 6
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.06746987951807229,
|
| 55 |
+
"grad_norm": 0.6067121698335344,
|
| 56 |
+
"learning_rate": 2.2580645161290324e-06,
|
| 57 |
+
"loss": 1.1416,
|
| 58 |
+
"step": 7
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.07710843373493977,
|
| 62 |
+
"grad_norm": 0.5822439970894898,
|
| 63 |
+
"learning_rate": 2.580645161290323e-06,
|
| 64 |
+
"loss": 1.2283,
|
| 65 |
+
"step": 8
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.08674698795180723,
|
| 69 |
+
"grad_norm": 0.43410811408302385,
|
| 70 |
+
"learning_rate": 2.903225806451613e-06,
|
| 71 |
+
"loss": 1.1301,
|
| 72 |
+
"step": 9
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.0963855421686747,
|
| 76 |
+
"grad_norm": 0.44969321494204945,
|
| 77 |
+
"learning_rate": 3.225806451612903e-06,
|
| 78 |
+
"loss": 1.1998,
|
| 79 |
+
"step": 10
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.10602409638554217,
|
| 83 |
+
"grad_norm": 0.40286147754479423,
|
| 84 |
+
"learning_rate": 3.548387096774194e-06,
|
| 85 |
+
"loss": 1.1185,
|
| 86 |
+
"step": 11
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.11566265060240964,
|
| 90 |
+
"grad_norm": 0.5535764409411175,
|
| 91 |
+
"learning_rate": 3.870967741935484e-06,
|
| 92 |
+
"loss": 1.0715,
|
| 93 |
+
"step": 12
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.12530120481927712,
|
| 97 |
+
"grad_norm": 0.48960428118771093,
|
| 98 |
+
"learning_rate": 4.193548387096774e-06,
|
| 99 |
+
"loss": 1.0387,
|
| 100 |
+
"step": 13
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.13493975903614458,
|
| 104 |
+
"grad_norm": 0.38578430703859284,
|
| 105 |
+
"learning_rate": 4.516129032258065e-06,
|
| 106 |
+
"loss": 1.0138,
|
| 107 |
+
"step": 14
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.14457831325301204,
|
| 111 |
+
"grad_norm": 0.36944902077760366,
|
| 112 |
+
"learning_rate": 4.838709677419355e-06,
|
| 113 |
+
"loss": 1.0815,
|
| 114 |
+
"step": 15
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.15421686746987953,
|
| 118 |
+
"grad_norm": 0.22615568618733844,
|
| 119 |
+
"learning_rate": 5.161290322580646e-06,
|
| 120 |
+
"loss": 0.9954,
|
| 121 |
+
"step": 16
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.163855421686747,
|
| 125 |
+
"grad_norm": 0.16924466530397725,
|
| 126 |
+
"learning_rate": 5.483870967741935e-06,
|
| 127 |
+
"loss": 0.9345,
|
| 128 |
+
"step": 17
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.17349397590361446,
|
| 132 |
+
"grad_norm": 0.17506858257513364,
|
| 133 |
+
"learning_rate": 5.806451612903226e-06,
|
| 134 |
+
"loss": 1.0057,
|
| 135 |
+
"step": 18
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.18313253012048192,
|
| 139 |
+
"grad_norm": 0.19256260025141983,
|
| 140 |
+
"learning_rate": 6.129032258064517e-06,
|
| 141 |
+
"loss": 0.961,
|
| 142 |
+
"step": 19
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.1927710843373494,
|
| 146 |
+
"grad_norm": 0.20045422860718864,
|
| 147 |
+
"learning_rate": 6.451612903225806e-06,
|
| 148 |
+
"loss": 0.9732,
|
| 149 |
+
"step": 20
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.20240963855421687,
|
| 153 |
+
"grad_norm": 0.20006737983005313,
|
| 154 |
+
"learning_rate": 6.774193548387097e-06,
|
| 155 |
+
"loss": 0.9114,
|
| 156 |
+
"step": 21
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.21204819277108433,
|
| 160 |
+
"grad_norm": 0.1804762390245476,
|
| 161 |
+
"learning_rate": 7.096774193548388e-06,
|
| 162 |
+
"loss": 0.9238,
|
| 163 |
+
"step": 22
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.2216867469879518,
|
| 167 |
+
"grad_norm": 0.19132885097999205,
|
| 168 |
+
"learning_rate": 7.4193548387096784e-06,
|
| 169 |
+
"loss": 0.9423,
|
| 170 |
+
"step": 23
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.23132530120481928,
|
| 174 |
+
"grad_norm": 0.1946117127819535,
|
| 175 |
+
"learning_rate": 7.741935483870968e-06,
|
| 176 |
+
"loss": 0.8987,
|
| 177 |
+
"step": 24
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.24096385542168675,
|
| 181 |
+
"grad_norm": 0.21044619139543286,
|
| 182 |
+
"learning_rate": 8.064516129032258e-06,
|
| 183 |
+
"loss": 0.923,
|
| 184 |
+
"step": 25
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.25060240963855424,
|
| 188 |
+
"grad_norm": 0.1894635936827489,
|
| 189 |
+
"learning_rate": 8.387096774193549e-06,
|
| 190 |
+
"loss": 0.8432,
|
| 191 |
+
"step": 26
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.26024096385542167,
|
| 195 |
+
"grad_norm": 0.2213813253619953,
|
| 196 |
+
"learning_rate": 8.70967741935484e-06,
|
| 197 |
+
"loss": 0.8921,
|
| 198 |
+
"step": 27
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.26987951807228916,
|
| 202 |
+
"grad_norm": 0.16421014736345188,
|
| 203 |
+
"learning_rate": 9.03225806451613e-06,
|
| 204 |
+
"loss": 0.891,
|
| 205 |
+
"step": 28
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.27951807228915665,
|
| 209 |
+
"grad_norm": 0.15888367691754784,
|
| 210 |
+
"learning_rate": 9.35483870967742e-06,
|
| 211 |
+
"loss": 0.8328,
|
| 212 |
+
"step": 29
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.2891566265060241,
|
| 216 |
+
"grad_norm": 0.1218262296356629,
|
| 217 |
+
"learning_rate": 9.67741935483871e-06,
|
| 218 |
+
"loss": 0.8109,
|
| 219 |
+
"step": 30
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.2987951807228916,
|
| 223 |
+
"grad_norm": 0.12772522304462855,
|
| 224 |
+
"learning_rate": 1e-05,
|
| 225 |
+
"loss": 0.8327,
|
| 226 |
+
"step": 31
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.30843373493975906,
|
| 230 |
+
"grad_norm": 0.11662156818152736,
|
| 231 |
+
"learning_rate": 9.999680739242022e-06,
|
| 232 |
+
"loss": 0.8228,
|
| 233 |
+
"step": 32
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.3180722891566265,
|
| 237 |
+
"grad_norm": 0.11423729898798743,
|
| 238 |
+
"learning_rate": 9.99872299773906e-06,
|
| 239 |
+
"loss": 0.8184,
|
| 240 |
+
"step": 33
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.327710843373494,
|
| 244 |
+
"grad_norm": 0.10732392773580121,
|
| 245 |
+
"learning_rate": 9.997126897798826e-06,
|
| 246 |
+
"loss": 0.8431,
|
| 247 |
+
"step": 34
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.3373493975903614,
|
| 251 |
+
"grad_norm": 0.10533931104652113,
|
| 252 |
+
"learning_rate": 9.994892643250147e-06,
|
| 253 |
+
"loss": 0.8436,
|
| 254 |
+
"step": 35
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.3469879518072289,
|
| 258 |
+
"grad_norm": 0.10349233667045869,
|
| 259 |
+
"learning_rate": 9.99202051941695e-06,
|
| 260 |
+
"loss": 0.8124,
|
| 261 |
+
"step": 36
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.3566265060240964,
|
| 265 |
+
"grad_norm": 0.12234917986226827,
|
| 266 |
+
"learning_rate": 9.9885108930818e-06,
|
| 267 |
+
"loss": 0.8402,
|
| 268 |
+
"step": 37
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.36626506024096384,
|
| 272 |
+
"grad_norm": 0.1084904559104385,
|
| 273 |
+
"learning_rate": 9.984364212439089e-06,
|
| 274 |
+
"loss": 0.7863,
|
| 275 |
+
"step": 38
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.3759036144578313,
|
| 279 |
+
"grad_norm": 0.09315599863192536,
|
| 280 |
+
"learning_rate": 9.979581007037776e-06,
|
| 281 |
+
"loss": 0.7494,
|
| 282 |
+
"step": 39
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.3855421686746988,
|
| 286 |
+
"grad_norm": 0.10455770281109837,
|
| 287 |
+
"learning_rate": 9.974161887713775e-06,
|
| 288 |
+
"loss": 0.7733,
|
| 289 |
+
"step": 40
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.39518072289156625,
|
| 293 |
+
"grad_norm": 0.09713116343076955,
|
| 294 |
+
"learning_rate": 9.968107546511942e-06,
|
| 295 |
+
"loss": 0.8069,
|
| 296 |
+
"step": 41
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.40481927710843374,
|
| 300 |
+
"grad_norm": 0.08994648197042518,
|
| 301 |
+
"learning_rate": 9.961418756597703e-06,
|
| 302 |
+
"loss": 0.7544,
|
| 303 |
+
"step": 42
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.41445783132530123,
|
| 307 |
+
"grad_norm": 0.08619129061295391,
|
| 308 |
+
"learning_rate": 9.95409637215831e-06,
|
| 309 |
+
"loss": 0.8046,
|
| 310 |
+
"step": 43
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.42409638554216866,
|
| 314 |
+
"grad_norm": 0.07881839593291304,
|
| 315 |
+
"learning_rate": 9.94614132829377e-06,
|
| 316 |
+
"loss": 0.7887,
|
| 317 |
+
"step": 44
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.43373493975903615,
|
| 321 |
+
"grad_norm": 0.08828454670099575,
|
| 322 |
+
"learning_rate": 9.937554640897414e-06,
|
| 323 |
+
"loss": 0.806,
|
| 324 |
+
"step": 45
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.4433734939759036,
|
| 328 |
+
"grad_norm": 0.08434767386597816,
|
| 329 |
+
"learning_rate": 9.928337406526172e-06,
|
| 330 |
+
"loss": 0.8187,
|
| 331 |
+
"step": 46
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.4530120481927711,
|
| 335 |
+
"grad_norm": 0.07330130217165154,
|
| 336 |
+
"learning_rate": 9.918490802260538e-06,
|
| 337 |
+
"loss": 0.7442,
|
| 338 |
+
"step": 47
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.46265060240963857,
|
| 342 |
+
"grad_norm": 0.08355675961610799,
|
| 343 |
+
"learning_rate": 9.90801608555425e-06,
|
| 344 |
+
"loss": 0.7288,
|
| 345 |
+
"step": 48
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.472289156626506,
|
| 349 |
+
"grad_norm": 0.08550275063984396,
|
| 350 |
+
"learning_rate": 9.896914594073703e-06,
|
| 351 |
+
"loss": 0.7876,
|
| 352 |
+
"step": 49
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.4819277108433735,
|
| 356 |
+
"grad_norm": 0.08574284988583232,
|
| 357 |
+
"learning_rate": 9.885187745527132e-06,
|
| 358 |
+
"loss": 0.7958,
|
| 359 |
+
"step": 50
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.491566265060241,
|
| 363 |
+
"grad_norm": 0.07092583421564527,
|
| 364 |
+
"learning_rate": 9.87283703748356e-06,
|
| 365 |
+
"loss": 0.7526,
|
| 366 |
+
"step": 51
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.5012048192771085,
|
| 370 |
+
"grad_norm": 0.08160685477674649,
|
| 371 |
+
"learning_rate": 9.859864047181551e-06,
|
| 372 |
+
"loss": 0.771,
|
| 373 |
+
"step": 52
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.5108433734939759,
|
| 377 |
+
"grad_norm": 0.08388500089586182,
|
| 378 |
+
"learning_rate": 9.846270431327793e-06,
|
| 379 |
+
"loss": 0.7916,
|
| 380 |
+
"step": 53
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.5204819277108433,
|
| 384 |
+
"grad_norm": 0.07274717865251835,
|
| 385 |
+
"learning_rate": 9.832057925885526e-06,
|
| 386 |
+
"loss": 0.7394,
|
| 387 |
+
"step": 54
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.5301204819277109,
|
| 391 |
+
"grad_norm": 0.07939094338552345,
|
| 392 |
+
"learning_rate": 9.817228345852853e-06,
|
| 393 |
+
"loss": 0.7336,
|
| 394 |
+
"step": 55
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.5397590361445783,
|
| 398 |
+
"grad_norm": 0.07845140622971908,
|
| 399 |
+
"learning_rate": 9.801783585030959e-06,
|
| 400 |
+
"loss": 0.7373,
|
| 401 |
+
"step": 56
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.5493975903614458,
|
| 405 |
+
"grad_norm": 0.07825843281252011,
|
| 406 |
+
"learning_rate": 9.785725615782262e-06,
|
| 407 |
+
"loss": 0.7453,
|
| 408 |
+
"step": 57
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.5590361445783133,
|
| 412 |
+
"grad_norm": 0.072056346411283,
|
| 413 |
+
"learning_rate": 9.769056488778538e-06,
|
| 414 |
+
"loss": 0.7468,
|
| 415 |
+
"step": 58
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.5686746987951807,
|
| 419 |
+
"grad_norm": 0.06493580070744535,
|
| 420 |
+
"learning_rate": 9.751778332739033e-06,
|
| 421 |
+
"loss": 0.6937,
|
| 422 |
+
"step": 59
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.5783132530120482,
|
| 426 |
+
"grad_norm": 0.06883589219210758,
|
| 427 |
+
"learning_rate": 9.733893354158628e-06,
|
| 428 |
+
"loss": 0.724,
|
| 429 |
+
"step": 60
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.5879518072289157,
|
| 433 |
+
"grad_norm": 0.08127915681080482,
|
| 434 |
+
"learning_rate": 9.715403837026046e-06,
|
| 435 |
+
"loss": 0.7357,
|
| 436 |
+
"step": 61
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.5975903614457831,
|
| 440 |
+
"grad_norm": 0.07532962426042335,
|
| 441 |
+
"learning_rate": 9.69631214253219e-06,
|
| 442 |
+
"loss": 0.7634,
|
| 443 |
+
"step": 62
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.6072289156626506,
|
| 447 |
+
"grad_norm": 0.07103096817411828,
|
| 448 |
+
"learning_rate": 9.676620708768608e-06,
|
| 449 |
+
"loss": 0.7507,
|
| 450 |
+
"step": 63
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.6168674698795181,
|
| 454 |
+
"grad_norm": 0.06730296630661546,
|
| 455 |
+
"learning_rate": 9.656332050416118e-06,
|
| 456 |
+
"loss": 0.7995,
|
| 457 |
+
"step": 64
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.6265060240963856,
|
| 461 |
+
"grad_norm": 0.07088532082738806,
|
| 462 |
+
"learning_rate": 9.635448758423703e-06,
|
| 463 |
+
"loss": 0.7284,
|
| 464 |
+
"step": 65
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.636144578313253,
|
| 468 |
+
"grad_norm": 0.07333112024028063,
|
| 469 |
+
"learning_rate": 9.613973499677613e-06,
|
| 470 |
+
"loss": 0.7757,
|
| 471 |
+
"step": 66
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.6457831325301204,
|
| 475 |
+
"grad_norm": 0.08025501213483908,
|
| 476 |
+
"learning_rate": 9.591909016660806e-06,
|
| 477 |
+
"loss": 0.7109,
|
| 478 |
+
"step": 67
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.655421686746988,
|
| 482 |
+
"grad_norm": 0.08343912298543543,
|
| 483 |
+
"learning_rate": 9.569258127102708e-06,
|
| 484 |
+
"loss": 0.6956,
|
| 485 |
+
"step": 68
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.6650602409638554,
|
| 489 |
+
"grad_norm": 0.0702913605748641,
|
| 490 |
+
"learning_rate": 9.546023723619387e-06,
|
| 491 |
+
"loss": 0.6946,
|
| 492 |
+
"step": 69
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.6746987951807228,
|
| 496 |
+
"grad_norm": 0.07487955447921284,
|
| 497 |
+
"learning_rate": 9.522208773344147e-06,
|
| 498 |
+
"loss": 0.6818,
|
| 499 |
+
"step": 70
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.6843373493975904,
|
| 503 |
+
"grad_norm": 0.07183727481397417,
|
| 504 |
+
"learning_rate": 9.497816317548625e-06,
|
| 505 |
+
"loss": 0.7371,
|
| 506 |
+
"step": 71
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.6939759036144578,
|
| 510 |
+
"grad_norm": 0.07423016586226384,
|
| 511 |
+
"learning_rate": 9.472849471254386e-06,
|
| 512 |
+
"loss": 0.7471,
|
| 513 |
+
"step": 72
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.7036144578313253,
|
| 517 |
+
"grad_norm": 0.07862160976107511,
|
| 518 |
+
"learning_rate": 9.447311422835141e-06,
|
| 519 |
+
"loss": 0.7603,
|
| 520 |
+
"step": 73
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.7132530120481928,
|
| 524 |
+
"grad_norm": 0.06406902535918113,
|
| 525 |
+
"learning_rate": 9.421205433609568e-06,
|
| 526 |
+
"loss": 0.6906,
|
| 527 |
+
"step": 74
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.7228915662650602,
|
| 531 |
+
"grad_norm": 0.06529603874190637,
|
| 532 |
+
"learning_rate": 9.39453483742483e-06,
|
| 533 |
+
"loss": 0.688,
|
| 534 |
+
"step": 75
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.7325301204819277,
|
| 538 |
+
"grad_norm": 0.0715253101992791,
|
| 539 |
+
"learning_rate": 9.367303040230828e-06,
|
| 540 |
+
"loss": 0.6933,
|
| 541 |
+
"step": 76
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.7421686746987952,
|
| 545 |
+
"grad_norm": 0.07138315068754539,
|
| 546 |
+
"learning_rate": 9.33951351964525e-06,
|
| 547 |
+
"loss": 0.7064,
|
| 548 |
+
"step": 77
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.7518072289156627,
|
| 552 |
+
"grad_norm": 0.06646488529034193,
|
| 553 |
+
"learning_rate": 9.311169824509454e-06,
|
| 554 |
+
"loss": 0.7174,
|
| 555 |
+
"step": 78
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.7614457831325301,
|
| 559 |
+
"grad_norm": 0.07024176035919157,
|
| 560 |
+
"learning_rate": 9.28227557443528e-06,
|
| 561 |
+
"loss": 0.7226,
|
| 562 |
+
"step": 79
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.7710843373493976,
|
| 566 |
+
"grad_norm": 0.062310962213733945,
|
| 567 |
+
"learning_rate": 9.252834459342801e-06,
|
| 568 |
+
"loss": 0.7238,
|
| 569 |
+
"step": 80
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.7807228915662651,
|
| 573 |
+
"grad_norm": 0.07322177936729748,
|
| 574 |
+
"learning_rate": 9.222850238989104e-06,
|
| 575 |
+
"loss": 0.706,
|
| 576 |
+
"step": 81
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.7903614457831325,
|
| 580 |
+
"grad_norm": 0.06195276994822025,
|
| 581 |
+
"learning_rate": 9.192326742488153e-06,
|
| 582 |
+
"loss": 0.7107,
|
| 583 |
+
"step": 82
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.8,
|
| 587 |
+
"grad_norm": 0.07022979646591547,
|
| 588 |
+
"learning_rate": 9.161267867821802e-06,
|
| 589 |
+
"loss": 0.678,
|
| 590 |
+
"step": 83
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.8096385542168675,
|
| 594 |
+
"grad_norm": 0.0757230181696156,
|
| 595 |
+
"learning_rate": 9.129677581342e-06,
|
| 596 |
+
"loss": 0.7487,
|
| 597 |
+
"step": 84
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.8192771084337349,
|
| 601 |
+
"grad_norm": 0.06342455930199235,
|
| 602 |
+
"learning_rate": 9.097559917264268e-06,
|
| 603 |
+
"loss": 0.6984,
|
| 604 |
+
"step": 85
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.8289156626506025,
|
| 608 |
+
"grad_norm": 0.06113100771712842,
|
| 609 |
+
"learning_rate": 9.064918977152517e-06,
|
| 610 |
+
"loss": 0.72,
|
| 611 |
+
"step": 86
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.8385542168674699,
|
| 615 |
+
"grad_norm": 0.06646354277708373,
|
| 616 |
+
"learning_rate": 9.031758929395259e-06,
|
| 617 |
+
"loss": 0.6858,
|
| 618 |
+
"step": 87
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.8481927710843373,
|
| 622 |
+
"grad_norm": 0.06414930650101153,
|
| 623 |
+
"learning_rate": 8.998084008673284e-06,
|
| 624 |
+
"loss": 0.7008,
|
| 625 |
+
"step": 88
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.8578313253012049,
|
| 629 |
+
"grad_norm": 0.06669153411991571,
|
| 630 |
+
"learning_rate": 8.963898515418885e-06,
|
| 631 |
+
"loss": 0.7093,
|
| 632 |
+
"step": 89
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.8674698795180723,
|
| 636 |
+
"grad_norm": 0.07412062045012623,
|
| 637 |
+
"learning_rate": 8.929206815266653e-06,
|
| 638 |
+
"loss": 0.6888,
|
| 639 |
+
"step": 90
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.8771084337349397,
|
| 643 |
+
"grad_norm": 0.07202834958462946,
|
| 644 |
+
"learning_rate": 8.89401333849598e-06,
|
| 645 |
+
"loss": 0.6948,
|
| 646 |
+
"step": 91
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.8867469879518072,
|
| 650 |
+
"grad_norm": 0.06873056807112735,
|
| 651 |
+
"learning_rate": 8.8583225794653e-06,
|
| 652 |
+
"loss": 0.7512,
|
| 653 |
+
"step": 92
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.8963855421686747,
|
| 657 |
+
"grad_norm": 0.06679929866753545,
|
| 658 |
+
"learning_rate": 8.82213909603812e-06,
|
| 659 |
+
"loss": 0.6964,
|
| 660 |
+
"step": 93
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.9060240963855422,
|
| 664 |
+
"grad_norm": 0.07323069668965577,
|
| 665 |
+
"learning_rate": 8.78546750900098e-06,
|
| 666 |
+
"loss": 0.7281,
|
| 667 |
+
"step": 94
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.9156626506024096,
|
| 671 |
+
"grad_norm": 0.07083868471644728,
|
| 672 |
+
"learning_rate": 8.748312501473351e-06,
|
| 673 |
+
"loss": 0.6991,
|
| 674 |
+
"step": 95
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.9253012048192771,
|
| 678 |
+
"grad_norm": 0.05937570692801768,
|
| 679 |
+
"learning_rate": 8.710678818309576e-06,
|
| 680 |
+
"loss": 0.6241,
|
| 681 |
+
"step": 96
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.9349397590361446,
|
| 685 |
+
"grad_norm": 0.061050430884171045,
|
| 686 |
+
"learning_rate": 8.672571265492944e-06,
|
| 687 |
+
"loss": 0.6176,
|
| 688 |
+
"step": 97
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.944578313253012,
|
| 692 |
+
"grad_norm": 0.0658465310603122,
|
| 693 |
+
"learning_rate": 8.63399470952193e-06,
|
| 694 |
+
"loss": 0.672,
|
| 695 |
+
"step": 98
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.9542168674698795,
|
| 699 |
+
"grad_norm": 0.06755179095814319,
|
| 700 |
+
"learning_rate": 8.594954076788736e-06,
|
| 701 |
+
"loss": 0.7228,
|
| 702 |
+
"step": 99
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.963855421686747,
|
| 706 |
+
"grad_norm": 0.07262602835248963,
|
| 707 |
+
"learning_rate": 8.555454352950161e-06,
|
| 708 |
+
"loss": 0.7101,
|
| 709 |
+
"step": 100
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.9734939759036144,
|
| 713 |
+
"grad_norm": 0.07303010678022298,
|
| 714 |
+
"learning_rate": 8.515500582290914e-06,
|
| 715 |
+
"loss": 0.6865,
|
| 716 |
+
"step": 101
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.983132530120482,
|
| 720 |
+
"grad_norm": 0.08187664550907275,
|
| 721 |
+
"learning_rate": 8.475097867079437e-06,
|
| 722 |
+
"loss": 0.6994,
|
| 723 |
+
"step": 102
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.9927710843373494,
|
| 727 |
+
"grad_norm": 0.06244683599564814,
|
| 728 |
+
"learning_rate": 8.434251366916323e-06,
|
| 729 |
+
"loss": 0.7047,
|
| 730 |
+
"step": 103
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 1.0,
|
| 734 |
+
"grad_norm": 0.06244683599564814,
|
| 735 |
+
"learning_rate": 8.392966298075413e-06,
|
| 736 |
+
"loss": 0.6859,
|
| 737 |
+
"step": 104
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 1.0096385542168675,
|
| 741 |
+
"grad_norm": 0.07725262140434404,
|
| 742 |
+
"learning_rate": 8.351247932837655e-06,
|
| 743 |
+
"loss": 0.6524,
|
| 744 |
+
"step": 105
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 1.0192771084337349,
|
| 748 |
+
"grad_norm": 0.05782618669079852,
|
| 749 |
+
"learning_rate": 8.309101598817812e-06,
|
| 750 |
+
"loss": 0.6007,
|
| 751 |
+
"step": 106
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 1.0289156626506024,
|
| 755 |
+
"grad_norm": 0.05932365316594588,
|
| 756 |
+
"learning_rate": 8.266532678284103e-06,
|
| 757 |
+
"loss": 0.6718,
|
| 758 |
+
"step": 107
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 1.03855421686747,
|
| 762 |
+
"grad_norm": 0.06611899088620404,
|
| 763 |
+
"learning_rate": 8.223546607470863e-06,
|
| 764 |
+
"loss": 0.62,
|
| 765 |
+
"step": 108
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 1.0481927710843373,
|
| 769 |
+
"grad_norm": 0.07180268403968028,
|
| 770 |
+
"learning_rate": 8.18014887588431e-06,
|
| 771 |
+
"loss": 0.5977,
|
| 772 |
+
"step": 109
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 1.0578313253012048,
|
| 776 |
+
"grad_norm": 0.061881962840029656,
|
| 777 |
+
"learning_rate": 8.13634502560152e-06,
|
| 778 |
+
"loss": 0.624,
|
| 779 |
+
"step": 110
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 1.0674698795180724,
|
| 783 |
+
"grad_norm": 0.06247008666424044,
|
| 784 |
+
"learning_rate": 8.092140650562665e-06,
|
| 785 |
+
"loss": 0.6297,
|
| 786 |
+
"step": 111
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 1.0771084337349397,
|
| 790 |
+
"grad_norm": 0.06286105854890581,
|
| 791 |
+
"learning_rate": 8.047541395856661e-06,
|
| 792 |
+
"loss": 0.628,
|
| 793 |
+
"step": 112
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 1.0867469879518072,
|
| 797 |
+
"grad_norm": 0.06730512890818845,
|
| 798 |
+
"learning_rate": 8.002552957000254e-06,
|
| 799 |
+
"loss": 0.6064,
|
| 800 |
+
"step": 113
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 1.0963855421686748,
|
| 804 |
+
"grad_norm": 0.06403895444236096,
|
| 805 |
+
"learning_rate": 7.957181079210676e-06,
|
| 806 |
+
"loss": 0.6034,
|
| 807 |
+
"step": 114
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 1.106024096385542,
|
| 811 |
+
"grad_norm": 0.05995505967096142,
|
| 812 |
+
"learning_rate": 7.911431556671967e-06,
|
| 813 |
+
"loss": 0.5827,
|
| 814 |
+
"step": 115
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 1.1156626506024097,
|
| 818 |
+
"grad_norm": 0.06106916100219134,
|
| 819 |
+
"learning_rate": 7.865310231795026e-06,
|
| 820 |
+
"loss": 0.6235,
|
| 821 |
+
"step": 116
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 1.1253012048192772,
|
| 825 |
+
"grad_norm": 0.0727899167728142,
|
| 826 |
+
"learning_rate": 7.818822994471504e-06,
|
| 827 |
+
"loss": 0.6082,
|
| 828 |
+
"step": 117
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 1.1349397590361445,
|
| 832 |
+
"grad_norm": 0.06595393108950595,
|
| 833 |
+
"learning_rate": 7.771975781321655e-06,
|
| 834 |
+
"loss": 0.6335,
|
| 835 |
+
"step": 118
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 1.144578313253012,
|
| 839 |
+
"grad_norm": 0.07220179641493146,
|
| 840 |
+
"learning_rate": 7.72477457493619e-06,
|
| 841 |
+
"loss": 0.6208,
|
| 842 |
+
"step": 119
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 1.1542168674698796,
|
| 846 |
+
"grad_norm": 0.053152296862338196,
|
| 847 |
+
"learning_rate": 7.677225403112277e-06,
|
| 848 |
+
"loss": 0.6314,
|
| 849 |
+
"step": 120
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 1.163855421686747,
|
| 853 |
+
"grad_norm": 0.06908539751424997,
|
| 854 |
+
"learning_rate": 7.629334338083774e-06,
|
| 855 |
+
"loss": 0.618,
|
| 856 |
+
"step": 121
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 1.1734939759036145,
|
| 860 |
+
"grad_norm": 0.0607666575336561,
|
| 861 |
+
"learning_rate": 7.58110749574577e-06,
|
| 862 |
+
"loss": 0.6127,
|
| 863 |
+
"step": 122
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 1.1831325301204818,
|
| 867 |
+
"grad_norm": 0.05725193856979287,
|
| 868 |
+
"learning_rate": 7.532551034873558e-06,
|
| 869 |
+
"loss": 0.6414,
|
| 870 |
+
"step": 123
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 1.1927710843373494,
|
| 874 |
+
"grad_norm": 0.06466648989573874,
|
| 875 |
+
"learning_rate": 7.483671156336142e-06,
|
| 876 |
+
"loss": 0.5967,
|
| 877 |
+
"step": 124
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 1.202409638554217,
|
| 881 |
+
"grad_norm": 0.0653325240595629,
|
| 882 |
+
"learning_rate": 7.43447410230435e-06,
|
| 883 |
+
"loss": 0.6185,
|
| 884 |
+
"step": 125
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 1.2120481927710842,
|
| 888 |
+
"grad_norm": 0.07538141090987788,
|
| 889 |
+
"learning_rate": 7.384966155453686e-06,
|
| 890 |
+
"loss": 0.6412,
|
| 891 |
+
"step": 126
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 1.2216867469879518,
|
| 895 |
+
"grad_norm": 0.07040259282115024,
|
| 896 |
+
"learning_rate": 7.335153638162005e-06,
|
| 897 |
+
"loss": 0.6173,
|
| 898 |
+
"step": 127
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 1.2313253012048193,
|
| 902 |
+
"grad_norm": 0.060551841344496767,
|
| 903 |
+
"learning_rate": 7.285042911702116e-06,
|
| 904 |
+
"loss": 0.5867,
|
| 905 |
+
"step": 128
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 1.2409638554216866,
|
| 909 |
+
"grad_norm": 0.060059786172390106,
|
| 910 |
+
"learning_rate": 7.234640375429427e-06,
|
| 911 |
+
"loss": 0.617,
|
| 912 |
+
"step": 129
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 1.2506024096385542,
|
| 916 |
+
"grad_norm": 0.06747743957174769,
|
| 917 |
+
"learning_rate": 7.183952465964711e-06,
|
| 918 |
+
"loss": 0.6193,
|
| 919 |
+
"step": 130
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 1.2602409638554217,
|
| 923 |
+
"grad_norm": 0.06828017138762647,
|
| 924 |
+
"learning_rate": 7.132985656372126e-06,
|
| 925 |
+
"loss": 0.6423,
|
| 926 |
+
"step": 131
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 1.269879518072289,
|
| 930 |
+
"grad_norm": 0.05392382216193354,
|
| 931 |
+
"learning_rate": 7.0817464553325764e-06,
|
| 932 |
+
"loss": 0.5984,
|
| 933 |
+
"step": 132
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 1.2795180722891566,
|
| 937 |
+
"grad_norm": 0.06241321312010483,
|
| 938 |
+
"learning_rate": 7.030241406312528e-06,
|
| 939 |
+
"loss": 0.6179,
|
| 940 |
+
"step": 133
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 1.2891566265060241,
|
| 944 |
+
"grad_norm": 0.06613062850109816,
|
| 945 |
+
"learning_rate": 6.978477086728375e-06,
|
| 946 |
+
"loss": 0.6758,
|
| 947 |
+
"step": 134
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 1.2987951807228915,
|
| 951 |
+
"grad_norm": 0.06410748091385764,
|
| 952 |
+
"learning_rate": 6.926460107106483e-06,
|
| 953 |
+
"loss": 0.6083,
|
| 954 |
+
"step": 135
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 1.308433734939759,
|
| 958 |
+
"grad_norm": 0.0608689353711739,
|
| 959 |
+
"learning_rate": 6.874197110238986e-06,
|
| 960 |
+
"loss": 0.621,
|
| 961 |
+
"step": 136
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 1.3180722891566266,
|
| 965 |
+
"grad_norm": 0.06714353872124008,
|
| 966 |
+
"learning_rate": 6.8216947703354815e-06,
|
| 967 |
+
"loss": 0.602,
|
| 968 |
+
"step": 137
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 1.3277108433734939,
|
| 972 |
+
"grad_norm": 0.0762090276876796,
|
| 973 |
+
"learning_rate": 6.7689597921707065e-06,
|
| 974 |
+
"loss": 0.6078,
|
| 975 |
+
"step": 138
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 1.3373493975903614,
|
| 979 |
+
"grad_norm": 0.06482193036022396,
|
| 980 |
+
"learning_rate": 6.715998910228296e-06,
|
| 981 |
+
"loss": 0.6012,
|
| 982 |
+
"step": 139
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 1.346987951807229,
|
| 986 |
+
"grad_norm": 0.06505573004876232,
|
| 987 |
+
"learning_rate": 6.6628188878407806e-06,
|
| 988 |
+
"loss": 0.5436,
|
| 989 |
+
"step": 140
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 1.3566265060240963,
|
| 993 |
+
"grad_norm": 0.062063475413644965,
|
| 994 |
+
"learning_rate": 6.609426516325859e-06,
|
| 995 |
+
"loss": 0.6328,
|
| 996 |
+
"step": 141
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 1.3662650602409638,
|
| 1000 |
+
"grad_norm": 0.05934060841850848,
|
| 1001 |
+
"learning_rate": 6.555828614119132e-06,
|
| 1002 |
+
"loss": 0.6157,
|
| 1003 |
+
"step": 142
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 1.3759036144578314,
|
| 1007 |
+
"grad_norm": 0.06467965488879493,
|
| 1008 |
+
"learning_rate": 6.502032025903356e-06,
|
| 1009 |
+
"loss": 0.668,
|
| 1010 |
+
"step": 143
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 1.3855421686746987,
|
| 1014 |
+
"grad_norm": 0.06412287381702947,
|
| 1015 |
+
"learning_rate": 6.4480436217343366e-06,
|
| 1016 |
+
"loss": 0.6411,
|
| 1017 |
+
"step": 144
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 1.3951807228915662,
|
| 1021 |
+
"grad_norm": 0.06321881583648611,
|
| 1022 |
+
"learning_rate": 6.393870296163616e-06,
|
| 1023 |
+
"loss": 0.6235,
|
| 1024 |
+
"step": 145
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 1.4048192771084338,
|
| 1028 |
+
"grad_norm": 0.06334369747563459,
|
| 1029 |
+
"learning_rate": 6.339518967357985e-06,
|
| 1030 |
+
"loss": 0.6281,
|
| 1031 |
+
"step": 146
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 1.4144578313253011,
|
| 1035 |
+
"grad_norm": 0.06504837084952969,
|
| 1036 |
+
"learning_rate": 6.284996576216014e-06,
|
| 1037 |
+
"loss": 0.5896,
|
| 1038 |
+
"step": 147
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 1.4240963855421687,
|
| 1042 |
+
"grad_norm": 0.05937400854730589,
|
| 1043 |
+
"learning_rate": 6.230310085481677e-06,
|
| 1044 |
+
"loss": 0.6225,
|
| 1045 |
+
"step": 148
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 1.4337349397590362,
|
| 1049 |
+
"grad_norm": 0.07525329007434042,
|
| 1050 |
+
"learning_rate": 6.175466478855161e-06,
|
| 1051 |
+
"loss": 0.6683,
|
| 1052 |
+
"step": 149
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 1.4433734939759035,
|
| 1056 |
+
"grad_norm": 0.0692386238529192,
|
| 1057 |
+
"learning_rate": 6.1204727601010396e-06,
|
| 1058 |
+
"loss": 0.665,
|
| 1059 |
+
"step": 150
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 1.453012048192771,
|
| 1063 |
+
"grad_norm": 0.06422226013162163,
|
| 1064 |
+
"learning_rate": 6.065335952153846e-06,
|
| 1065 |
+
"loss": 0.6313,
|
| 1066 |
+
"step": 151
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 1.4626506024096386,
|
| 1070 |
+
"grad_norm": 0.06068901796091081,
|
| 1071 |
+
"learning_rate": 6.010063096221215e-06,
|
| 1072 |
+
"loss": 0.6139,
|
| 1073 |
+
"step": 152
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 1.472289156626506,
|
| 1077 |
+
"grad_norm": 0.06840718236821038,
|
| 1078 |
+
"learning_rate": 5.954661250884704e-06,
|
| 1079 |
+
"loss": 0.6263,
|
| 1080 |
+
"step": 153
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 1.4819277108433735,
|
| 1084 |
+
"grad_norm": 0.06212232268715233,
|
| 1085 |
+
"learning_rate": 5.899137491198364e-06,
|
| 1086 |
+
"loss": 0.6635,
|
| 1087 |
+
"step": 154
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 1.491566265060241,
|
| 1091 |
+
"grad_norm": 0.07029083873722003,
|
| 1092 |
+
"learning_rate": 5.843498907785236e-06,
|
| 1093 |
+
"loss": 0.5506,
|
| 1094 |
+
"step": 155
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 1.5012048192771084,
|
| 1098 |
+
"grad_norm": 0.06346720559167707,
|
| 1099 |
+
"learning_rate": 5.78775260593185e-06,
|
| 1100 |
+
"loss": 0.6193,
|
| 1101 |
+
"step": 156
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 1.510843373493976,
|
| 1105 |
+
"grad_norm": 0.07558970902676904,
|
| 1106 |
+
"learning_rate": 5.731905704680834e-06,
|
| 1107 |
+
"loss": 0.6251,
|
| 1108 |
+
"step": 157
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 1.5204819277108435,
|
| 1112 |
+
"grad_norm": 0.05842912970122868,
|
| 1113 |
+
"learning_rate": 5.6759653359218e-06,
|
| 1114 |
+
"loss": 0.6761,
|
| 1115 |
+
"step": 158
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 1.5301204819277108,
|
| 1119 |
+
"grad_norm": 0.05813865281679669,
|
| 1120 |
+
"learning_rate": 5.6199386434805615e-06,
|
| 1121 |
+
"loss": 0.5842,
|
| 1122 |
+
"step": 159
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 1.5397590361445783,
|
| 1126 |
+
"grad_norm": 0.05744456895698129,
|
| 1127 |
+
"learning_rate": 5.563832782206835e-06,
|
| 1128 |
+
"loss": 0.6087,
|
| 1129 |
+
"step": 160
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 1.5493975903614459,
|
| 1133 |
+
"grad_norm": 0.06419357719628659,
|
| 1134 |
+
"learning_rate": 5.507654917060541e-06,
|
| 1135 |
+
"loss": 0.6059,
|
| 1136 |
+
"step": 161
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 1.5590361445783132,
|
| 1140 |
+
"grad_norm": 0.06081199348431484,
|
| 1141 |
+
"learning_rate": 5.451412222196801e-06,
|
| 1142 |
+
"loss": 0.5965,
|
| 1143 |
+
"step": 162
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 1.5686746987951807,
|
| 1147 |
+
"grad_norm": 0.06588220582304925,
|
| 1148 |
+
"learning_rate": 5.395111880049775e-06,
|
| 1149 |
+
"loss": 0.5697,
|
| 1150 |
+
"step": 163
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 1.5783132530120483,
|
| 1154 |
+
"grad_norm": 0.06367402188772205,
|
| 1155 |
+
"learning_rate": 5.338761080415425e-06,
|
| 1156 |
+
"loss": 0.6026,
|
| 1157 |
+
"step": 164
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 1.5879518072289156,
|
| 1161 |
+
"grad_norm": 0.06057405284771154,
|
| 1162 |
+
"learning_rate": 5.28236701953335e-06,
|
| 1163 |
+
"loss": 0.5812,
|
| 1164 |
+
"step": 165
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 1.5975903614457831,
|
| 1168 |
+
"grad_norm": 0.053482867334479156,
|
| 1169 |
+
"learning_rate": 5.225936899167803e-06,
|
| 1170 |
+
"loss": 0.6347,
|
| 1171 |
+
"step": 166
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 1.6072289156626507,
|
| 1175 |
+
"grad_norm": 0.06071788010807735,
|
| 1176 |
+
"learning_rate": 5.169477925687981e-06,
|
| 1177 |
+
"loss": 0.61,
|
| 1178 |
+
"step": 167
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 1.616867469879518,
|
| 1182 |
+
"grad_norm": 0.04803665860432399,
|
| 1183 |
+
"learning_rate": 5.112997309147753e-06,
|
| 1184 |
+
"loss": 0.6102,
|
| 1185 |
+
"step": 168
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 1.6265060240963856,
|
| 1189 |
+
"grad_norm": 0.07383190175754212,
|
| 1190 |
+
"learning_rate": 5.0565022623649e-06,
|
| 1191 |
+
"loss": 0.6634,
|
| 1192 |
+
"step": 169
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 1.636144578313253,
|
| 1196 |
+
"grad_norm": 0.06343533190073457,
|
| 1197 |
+
"learning_rate": 5e-06,
|
| 1198 |
+
"loss": 0.6221,
|
| 1199 |
+
"step": 170
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 1.6457831325301204,
|
| 1203 |
+
"grad_norm": 0.05067311084373564,
|
| 1204 |
+
"learning_rate": 4.943497737635103e-06,
|
| 1205 |
+
"loss": 0.576,
|
| 1206 |
+
"step": 171
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 1.655421686746988,
|
| 1210 |
+
"grad_norm": 0.07430397087873777,
|
| 1211 |
+
"learning_rate": 4.887002690852249e-06,
|
| 1212 |
+
"loss": 0.5677,
|
| 1213 |
+
"step": 172
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 1.6650602409638555,
|
| 1217 |
+
"grad_norm": 0.0573447314824425,
|
| 1218 |
+
"learning_rate": 4.830522074312019e-06,
|
| 1219 |
+
"loss": 0.6066,
|
| 1220 |
+
"step": 173
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 1.6746987951807228,
|
| 1224 |
+
"grad_norm": 0.0663766323629017,
|
| 1225 |
+
"learning_rate": 4.774063100832199e-06,
|
| 1226 |
+
"loss": 0.6569,
|
| 1227 |
+
"step": 174
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 1.6843373493975904,
|
| 1231 |
+
"grad_norm": 0.06715405212610676,
|
| 1232 |
+
"learning_rate": 4.717632980466652e-06,
|
| 1233 |
+
"loss": 0.6171,
|
| 1234 |
+
"step": 175
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 1.693975903614458,
|
| 1238 |
+
"grad_norm": 0.06792888500682888,
|
| 1239 |
+
"learning_rate": 4.661238919584578e-06,
|
| 1240 |
+
"loss": 0.6257,
|
| 1241 |
+
"step": 176
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 1.7036144578313253,
|
| 1245 |
+
"grad_norm": 0.06352940843145351,
|
| 1246 |
+
"learning_rate": 4.6048881199502265e-06,
|
| 1247 |
+
"loss": 0.5744,
|
| 1248 |
+
"step": 177
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 1.7132530120481928,
|
| 1252 |
+
"grad_norm": 0.06122504041319759,
|
| 1253 |
+
"learning_rate": 4.548587777803198e-06,
|
| 1254 |
+
"loss": 0.6479,
|
| 1255 |
+
"step": 178
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 1.7228915662650603,
|
| 1259 |
+
"grad_norm": 0.06714121511192614,
|
| 1260 |
+
"learning_rate": 4.49234508293946e-06,
|
| 1261 |
+
"loss": 0.5905,
|
| 1262 |
+
"step": 179
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 1.7325301204819277,
|
| 1266 |
+
"grad_norm": 0.06504871151522112,
|
| 1267 |
+
"learning_rate": 4.436167217793167e-06,
|
| 1268 |
+
"loss": 0.6377,
|
| 1269 |
+
"step": 180
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 1.7421686746987952,
|
| 1273 |
+
"grad_norm": 0.05746737774784524,
|
| 1274 |
+
"learning_rate": 4.38006135651944e-06,
|
| 1275 |
+
"loss": 0.6343,
|
| 1276 |
+
"step": 181
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 1.7518072289156628,
|
| 1280 |
+
"grad_norm": 0.0742332284296601,
|
| 1281 |
+
"learning_rate": 4.3240346640782014e-06,
|
| 1282 |
+
"loss": 0.604,
|
| 1283 |
+
"step": 182
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 1.76144578313253,
|
| 1287 |
+
"grad_norm": 0.07199906822994642,
|
| 1288 |
+
"learning_rate": 4.268094295319167e-06,
|
| 1289 |
+
"loss": 0.5884,
|
| 1290 |
+
"step": 183
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 1.7710843373493976,
|
| 1294 |
+
"grad_norm": 0.06146250524213322,
|
| 1295 |
+
"learning_rate": 4.212247394068151e-06,
|
| 1296 |
+
"loss": 0.5786,
|
| 1297 |
+
"step": 184
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 1.7807228915662652,
|
| 1301 |
+
"grad_norm": 0.06678830744721113,
|
| 1302 |
+
"learning_rate": 4.1565010922147644e-06,
|
| 1303 |
+
"loss": 0.568,
|
| 1304 |
+
"step": 185
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 1.7903614457831325,
|
| 1308 |
+
"grad_norm": 0.06225018767217524,
|
| 1309 |
+
"learning_rate": 4.100862508801639e-06,
|
| 1310 |
+
"loss": 0.6216,
|
| 1311 |
+
"step": 186
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 1.8,
|
| 1315 |
+
"grad_norm": 0.06390858568136344,
|
| 1316 |
+
"learning_rate": 4.045338749115299e-06,
|
| 1317 |
+
"loss": 0.58,
|
| 1318 |
+
"step": 187
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 1.8096385542168676,
|
| 1322 |
+
"grad_norm": 0.06358790048246866,
|
| 1323 |
+
"learning_rate": 3.989936903778785e-06,
|
| 1324 |
+
"loss": 0.6225,
|
| 1325 |
+
"step": 188
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 1.819277108433735,
|
| 1329 |
+
"grad_norm": 0.07051504136128764,
|
| 1330 |
+
"learning_rate": 3.934664047846157e-06,
|
| 1331 |
+
"loss": 0.5685,
|
| 1332 |
+
"step": 189
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 1.8289156626506025,
|
| 1336 |
+
"grad_norm": 0.05730039419508112,
|
| 1337 |
+
"learning_rate": 3.879527239898962e-06,
|
| 1338 |
+
"loss": 0.5875,
|
| 1339 |
+
"step": 190
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 1.83855421686747,
|
| 1343 |
+
"grad_norm": 0.0578158196725697,
|
| 1344 |
+
"learning_rate": 3.8245335211448404e-06,
|
| 1345 |
+
"loss": 0.5635,
|
| 1346 |
+
"step": 191
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 1.8481927710843373,
|
| 1350 |
+
"grad_norm": 0.07225016497571735,
|
| 1351 |
+
"learning_rate": 3.769689914518326e-06,
|
| 1352 |
+
"loss": 0.6155,
|
| 1353 |
+
"step": 192
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 1.8578313253012049,
|
| 1357 |
+
"grad_norm": 0.0549804131963855,
|
| 1358 |
+
"learning_rate": 3.715003423783986e-06,
|
| 1359 |
+
"loss": 0.624,
|
| 1360 |
+
"step": 193
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 1.8674698795180724,
|
| 1364 |
+
"grad_norm": 0.06298514519756715,
|
| 1365 |
+
"learning_rate": 3.660481032642016e-06,
|
| 1366 |
+
"loss": 0.6506,
|
| 1367 |
+
"step": 194
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 1.8771084337349397,
|
| 1371 |
+
"grad_norm": 0.0686500519258368,
|
| 1372 |
+
"learning_rate": 3.6061297038363853e-06,
|
| 1373 |
+
"loss": 0.602,
|
| 1374 |
+
"step": 195
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 1.886746987951807,
|
| 1378 |
+
"grad_norm": 0.06368295604671866,
|
| 1379 |
+
"learning_rate": 3.5519563782656642e-06,
|
| 1380 |
+
"loss": 0.6721,
|
| 1381 |
+
"step": 196
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 1.8963855421686748,
|
| 1385 |
+
"grad_norm": 0.06096396381363934,
|
| 1386 |
+
"learning_rate": 3.497967974096647e-06,
|
| 1387 |
+
"loss": 0.6107,
|
| 1388 |
+
"step": 197
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 1.9060240963855422,
|
| 1392 |
+
"grad_norm": 0.06013183394905736,
|
| 1393 |
+
"learning_rate": 3.4441713858808684e-06,
|
| 1394 |
+
"loss": 0.582,
|
| 1395 |
+
"step": 198
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 1.9156626506024095,
|
| 1399 |
+
"grad_norm": 0.05869263235641963,
|
| 1400 |
+
"learning_rate": 3.3905734836741415e-06,
|
| 1401 |
+
"loss": 0.5896,
|
| 1402 |
+
"step": 199
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 1.9253012048192772,
|
| 1406 |
+
"grad_norm": 0.06144735421932478,
|
| 1407 |
+
"learning_rate": 3.3371811121592203e-06,
|
| 1408 |
+
"loss": 0.551,
|
| 1409 |
+
"step": 200
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 1.9349397590361446,
|
| 1413 |
+
"grad_norm": 0.07092447055189571,
|
| 1414 |
+
"learning_rate": 3.2840010897717045e-06,
|
| 1415 |
+
"loss": 0.603,
|
| 1416 |
+
"step": 201
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 1.944578313253012,
|
| 1420 |
+
"grad_norm": 0.0608882656847333,
|
| 1421 |
+
"learning_rate": 3.2310402078292956e-06,
|
| 1422 |
+
"loss": 0.6072,
|
| 1423 |
+
"step": 202
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 1.9542168674698797,
|
| 1427 |
+
"grad_norm": 0.06498984091309146,
|
| 1428 |
+
"learning_rate": 3.178305229664519e-06,
|
| 1429 |
+
"loss": 0.5848,
|
| 1430 |
+
"step": 203
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 1.963855421686747,
|
| 1434 |
+
"grad_norm": 0.06597558210840206,
|
| 1435 |
+
"learning_rate": 3.125802889761016e-06,
|
| 1436 |
+
"loss": 0.6262,
|
| 1437 |
+
"step": 204
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 1.9734939759036143,
|
| 1441 |
+
"grad_norm": 0.05675086313446132,
|
| 1442 |
+
"learning_rate": 3.073539892893519e-06,
|
| 1443 |
+
"loss": 0.5991,
|
| 1444 |
+
"step": 205
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 1.983132530120482,
|
| 1448 |
+
"grad_norm": 0.062196722016115806,
|
| 1449 |
+
"learning_rate": 3.021522913271627e-06,
|
| 1450 |
+
"loss": 0.601,
|
| 1451 |
+
"step": 206
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 1.9927710843373494,
|
| 1455 |
+
"grad_norm": 0.054930108170390506,
|
| 1456 |
+
"learning_rate": 2.969758593687475e-06,
|
| 1457 |
+
"loss": 0.5751,
|
| 1458 |
+
"step": 207
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 2.0,
|
| 1462 |
+
"grad_norm": 0.08111810625267253,
|
| 1463 |
+
"learning_rate": 2.9182535446674244e-06,
|
| 1464 |
+
"loss": 0.5888,
|
| 1465 |
+
"step": 208
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 2.0096385542168673,
|
| 1469 |
+
"grad_norm": 0.08388412876294439,
|
| 1470 |
+
"learning_rate": 2.8670143436278757e-06,
|
| 1471 |
+
"loss": 0.5468,
|
| 1472 |
+
"step": 209
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 2.019277108433735,
|
| 1476 |
+
"grad_norm": 0.05934240328760545,
|
| 1477 |
+
"learning_rate": 2.8160475340352913e-06,
|
| 1478 |
+
"loss": 0.5644,
|
| 1479 |
+
"step": 210
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 2.0289156626506024,
|
| 1483 |
+
"grad_norm": 0.06787576120463162,
|
| 1484 |
+
"learning_rate": 2.765359624570574e-06,
|
| 1485 |
+
"loss": 0.5886,
|
| 1486 |
+
"step": 211
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 2.0385542168674697,
|
| 1490 |
+
"grad_norm": 0.0633591218554768,
|
| 1491 |
+
"learning_rate": 2.714957088297886e-06,
|
| 1492 |
+
"loss": 0.527,
|
| 1493 |
+
"step": 212
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 2.0481927710843375,
|
| 1497 |
+
"grad_norm": 0.06017979720395379,
|
| 1498 |
+
"learning_rate": 2.664846361837997e-06,
|
| 1499 |
+
"loss": 0.5562,
|
| 1500 |
+
"step": 213
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 2.057831325301205,
|
| 1504 |
+
"grad_norm": 0.0680365289599211,
|
| 1505 |
+
"learning_rate": 2.6150338445463146e-06,
|
| 1506 |
+
"loss": 0.558,
|
| 1507 |
+
"step": 214
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 2.067469879518072,
|
| 1511 |
+
"grad_norm": 0.05321779366151024,
|
| 1512 |
+
"learning_rate": 2.565525897695651e-06,
|
| 1513 |
+
"loss": 0.5618,
|
| 1514 |
+
"step": 215
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 2.07710843373494,
|
| 1518 |
+
"grad_norm": 0.06537975532750284,
|
| 1519 |
+
"learning_rate": 2.51632884366386e-06,
|
| 1520 |
+
"loss": 0.5539,
|
| 1521 |
+
"step": 216
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 2.0867469879518072,
|
| 1525 |
+
"grad_norm": 0.0611607010768512,
|
| 1526 |
+
"learning_rate": 2.4674489651264433e-06,
|
| 1527 |
+
"loss": 0.5085,
|
| 1528 |
+
"step": 217
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 2.0963855421686746,
|
| 1532 |
+
"grad_norm": 0.05053608558411659,
|
| 1533 |
+
"learning_rate": 2.418892504254231e-06,
|
| 1534 |
+
"loss": 0.5368,
|
| 1535 |
+
"step": 218
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 2.1060240963855423,
|
| 1539 |
+
"grad_norm": 0.06189373201667931,
|
| 1540 |
+
"learning_rate": 2.3706656619162278e-06,
|
| 1541 |
+
"loss": 0.5227,
|
| 1542 |
+
"step": 219
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 2.1156626506024097,
|
| 1546 |
+
"grad_norm": 0.061459908604694274,
|
| 1547 |
+
"learning_rate": 2.322774596887726e-06,
|
| 1548 |
+
"loss": 0.6102,
|
| 1549 |
+
"step": 220
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 2.125301204819277,
|
| 1553 |
+
"grad_norm": 0.07071845948302857,
|
| 1554 |
+
"learning_rate": 2.275225425063813e-06,
|
| 1555 |
+
"loss": 0.5673,
|
| 1556 |
+
"step": 221
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 2.1349397590361447,
|
| 1560 |
+
"grad_norm": 0.059265035233080764,
|
| 1561 |
+
"learning_rate": 2.2280242186783473e-06,
|
| 1562 |
+
"loss": 0.5475,
|
| 1563 |
+
"step": 222
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 2.144578313253012,
|
| 1567 |
+
"grad_norm": 0.05927854864357284,
|
| 1568 |
+
"learning_rate": 2.1811770055284968e-06,
|
| 1569 |
+
"loss": 0.566,
|
| 1570 |
+
"step": 223
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 2.1542168674698794,
|
| 1574 |
+
"grad_norm": 0.05843386609477244,
|
| 1575 |
+
"learning_rate": 2.134689768204975e-06,
|
| 1576 |
+
"loss": 0.5263,
|
| 1577 |
+
"step": 224
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 2.163855421686747,
|
| 1581 |
+
"grad_norm": 0.054515077941895744,
|
| 1582 |
+
"learning_rate": 2.0885684433280336e-06,
|
| 1583 |
+
"loss": 0.5706,
|
| 1584 |
+
"step": 225
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 2.1734939759036145,
|
| 1588 |
+
"grad_norm": 0.060254043551932875,
|
| 1589 |
+
"learning_rate": 2.042818920789326e-06,
|
| 1590 |
+
"loss": 0.4764,
|
| 1591 |
+
"step": 226
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 2.183132530120482,
|
| 1595 |
+
"grad_norm": 0.0570869413051247,
|
| 1596 |
+
"learning_rate": 1.9974470429997482e-06,
|
| 1597 |
+
"loss": 0.5465,
|
| 1598 |
+
"step": 227
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 2.1927710843373496,
|
| 1602 |
+
"grad_norm": 0.055562628054217,
|
| 1603 |
+
"learning_rate": 1.9524586041433393e-06,
|
| 1604 |
+
"loss": 0.5472,
|
| 1605 |
+
"step": 228
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 2.202409638554217,
|
| 1609 |
+
"grad_norm": 0.06059513133262611,
|
| 1610 |
+
"learning_rate": 1.907859349437336e-06,
|
| 1611 |
+
"loss": 0.5484,
|
| 1612 |
+
"step": 229
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 2.212048192771084,
|
| 1616 |
+
"grad_norm": 0.06123920490330747,
|
| 1617 |
+
"learning_rate": 1.8636549743984815e-06,
|
| 1618 |
+
"loss": 0.5507,
|
| 1619 |
+
"step": 230
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 2.221686746987952,
|
| 1623 |
+
"grad_norm": 0.05837754169240895,
|
| 1624 |
+
"learning_rate": 1.8198511241156902e-06,
|
| 1625 |
+
"loss": 0.5172,
|
| 1626 |
+
"step": 231
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 2.2313253012048193,
|
| 1630 |
+
"grad_norm": 0.06444730736010255,
|
| 1631 |
+
"learning_rate": 1.776453392529139e-06,
|
| 1632 |
+
"loss": 0.5473,
|
| 1633 |
+
"step": 232
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 2.2409638554216866,
|
| 1637 |
+
"grad_norm": 0.057036642865459466,
|
| 1638 |
+
"learning_rate": 1.7334673217158976e-06,
|
| 1639 |
+
"loss": 0.5243,
|
| 1640 |
+
"step": 233
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 2.2506024096385544,
|
| 1644 |
+
"grad_norm": 0.05541742317802861,
|
| 1645 |
+
"learning_rate": 1.6908984011821883e-06,
|
| 1646 |
+
"loss": 0.5755,
|
| 1647 |
+
"step": 234
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 2.2602409638554217,
|
| 1651 |
+
"grad_norm": 0.059707658549255194,
|
| 1652 |
+
"learning_rate": 1.6487520671623469e-06,
|
| 1653 |
+
"loss": 0.5463,
|
| 1654 |
+
"step": 235
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 2.269879518072289,
|
| 1658 |
+
"grad_norm": 0.0652488770688145,
|
| 1659 |
+
"learning_rate": 1.6070337019245896e-06,
|
| 1660 |
+
"loss": 0.5434,
|
| 1661 |
+
"step": 236
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 2.279518072289157,
|
| 1665 |
+
"grad_norm": 0.062427610958618734,
|
| 1666 |
+
"learning_rate": 1.5657486330836786e-06,
|
| 1667 |
+
"loss": 0.5404,
|
| 1668 |
+
"step": 237
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 2.289156626506024,
|
| 1672 |
+
"grad_norm": 0.066481990615362,
|
| 1673 |
+
"learning_rate": 1.5249021329205638e-06,
|
| 1674 |
+
"loss": 0.5886,
|
| 1675 |
+
"step": 238
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 2.2987951807228915,
|
| 1679 |
+
"grad_norm": 0.06075798833103501,
|
| 1680 |
+
"learning_rate": 1.4844994177090871e-06,
|
| 1681 |
+
"loss": 0.5906,
|
| 1682 |
+
"step": 239
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 2.3084337349397592,
|
| 1686 |
+
"grad_norm": 0.060457408593263186,
|
| 1687 |
+
"learning_rate": 1.4445456470498392e-06,
|
| 1688 |
+
"loss": 0.5201,
|
| 1689 |
+
"step": 240
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 2.3180722891566266,
|
| 1693 |
+
"grad_norm": 0.06189290727406132,
|
| 1694 |
+
"learning_rate": 1.4050459232112652e-06,
|
| 1695 |
+
"loss": 0.5486,
|
| 1696 |
+
"step": 241
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 2.327710843373494,
|
| 1700 |
+
"grad_norm": 0.05446525602984187,
|
| 1701 |
+
"learning_rate": 1.3660052904780707e-06,
|
| 1702 |
+
"loss": 0.5667,
|
| 1703 |
+
"step": 242
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 2.337349397590361,
|
| 1707 |
+
"grad_norm": 0.05680759175644314,
|
| 1708 |
+
"learning_rate": 1.3274287345070564e-06,
|
| 1709 |
+
"loss": 0.4991,
|
| 1710 |
+
"step": 243
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 2.346987951807229,
|
| 1714 |
+
"grad_norm": 0.07369531885761765,
|
| 1715 |
+
"learning_rate": 1.2893211816904243e-06,
|
| 1716 |
+
"loss": 0.5635,
|
| 1717 |
+
"step": 244
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 2.3566265060240963,
|
| 1721 |
+
"grad_norm": 0.047802386111405644,
|
| 1722 |
+
"learning_rate": 1.2516874985266508e-06,
|
| 1723 |
+
"loss": 0.5458,
|
| 1724 |
+
"step": 245
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 2.3662650602409636,
|
| 1728 |
+
"grad_norm": 0.056195785484295456,
|
| 1729 |
+
"learning_rate": 1.2145324909990202e-06,
|
| 1730 |
+
"loss": 0.5458,
|
| 1731 |
+
"step": 246
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 2.3759036144578314,
|
| 1735 |
+
"grad_norm": 0.05563528805443362,
|
| 1736 |
+
"learning_rate": 1.1778609039618804e-06,
|
| 1737 |
+
"loss": 0.4986,
|
| 1738 |
+
"step": 247
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 2.3855421686746987,
|
| 1742 |
+
"grad_norm": 0.06250095823475356,
|
| 1743 |
+
"learning_rate": 1.1416774205347015e-06,
|
| 1744 |
+
"loss": 0.5358,
|
| 1745 |
+
"step": 248
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 2.395180722891566,
|
| 1749 |
+
"grad_norm": 0.05734992662747759,
|
| 1750 |
+
"learning_rate": 1.1059866615040205e-06,
|
| 1751 |
+
"loss": 0.5315,
|
| 1752 |
+
"step": 249
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 2.404819277108434,
|
| 1756 |
+
"grad_norm": 0.06007313643233539,
|
| 1757 |
+
"learning_rate": 1.0707931847333487e-06,
|
| 1758 |
+
"loss": 0.5265,
|
| 1759 |
+
"step": 250
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 2.414457831325301,
|
| 1763 |
+
"grad_norm": 0.055511713433456615,
|
| 1764 |
+
"learning_rate": 1.036101484581117e-06,
|
| 1765 |
+
"loss": 0.5207,
|
| 1766 |
+
"step": 251
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 2.4240963855421684,
|
| 1770 |
+
"grad_norm": 0.06459562675579453,
|
| 1771 |
+
"learning_rate": 1.0019159913267156e-06,
|
| 1772 |
+
"loss": 0.5467,
|
| 1773 |
+
"step": 252
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 2.433734939759036,
|
| 1777 |
+
"grad_norm": 0.0616817710533823,
|
| 1778 |
+
"learning_rate": 9.68241070604743e-07,
|
| 1779 |
+
"loss": 0.5487,
|
| 1780 |
+
"step": 253
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 2.4433734939759035,
|
| 1784 |
+
"grad_norm": 0.05982170620400226,
|
| 1785 |
+
"learning_rate": 9.350810228474855e-07,
|
| 1786 |
+
"loss": 0.5894,
|
| 1787 |
+
"step": 254
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 2.453012048192771,
|
| 1791 |
+
"grad_norm": 0.06404688237048371,
|
| 1792 |
+
"learning_rate": 9.024400827357344e-07,
|
| 1793 |
+
"loss": 0.5273,
|
| 1794 |
+
"step": 255
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 2.4626506024096386,
|
| 1798 |
+
"grad_norm": 0.055361287946098334,
|
| 1799 |
+
"learning_rate": 8.703224186580012e-07,
|
| 1800 |
+
"loss": 0.5576,
|
| 1801 |
+
"step": 256
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 2.472289156626506,
|
| 1805 |
+
"grad_norm": 0.0563022702076071,
|
| 1806 |
+
"learning_rate": 8.387321321781977e-07,
|
| 1807 |
+
"loss": 0.5291,
|
| 1808 |
+
"step": 257
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 2.4819277108433733,
|
| 1812 |
+
"grad_norm": 0.06713390666603862,
|
| 1813 |
+
"learning_rate": 8.07673257511849e-07,
|
| 1814 |
+
"loss": 0.5465,
|
| 1815 |
+
"step": 258
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 2.491566265060241,
|
| 1819 |
+
"grad_norm": 0.049523569414613715,
|
| 1820 |
+
"learning_rate": 7.771497610108981e-07,
|
| 1821 |
+
"loss": 0.5626,
|
| 1822 |
+
"step": 259
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 2.5012048192771084,
|
| 1826 |
+
"grad_norm": 0.06250857829748814,
|
| 1827 |
+
"learning_rate": 7.471655406572003e-07,
|
| 1828 |
+
"loss": 0.5076,
|
| 1829 |
+
"step": 260
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 2.5108433734939757,
|
| 1833 |
+
"grad_norm": 0.05831361066226844,
|
| 1834 |
+
"learning_rate": 7.177244255647209e-07,
|
| 1835 |
+
"loss": 0.5583,
|
| 1836 |
+
"step": 261
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 2.5204819277108435,
|
| 1840 |
+
"grad_norm": 0.06401512676119323,
|
| 1841 |
+
"learning_rate": 6.888301754905469e-07,
|
| 1842 |
+
"loss": 0.5406,
|
| 1843 |
+
"step": 262
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 2.5301204819277108,
|
| 1847 |
+
"grad_norm": 0.05000225920367488,
|
| 1848 |
+
"learning_rate": 6.604864803547511e-07,
|
| 1849 |
+
"loss": 0.5643,
|
| 1850 |
+
"step": 263
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 2.539759036144578,
|
| 1854 |
+
"grad_norm": 0.06148033172818441,
|
| 1855 |
+
"learning_rate": 6.326969597691724e-07,
|
| 1856 |
+
"loss": 0.5538,
|
| 1857 |
+
"step": 264
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 2.549397590361446,
|
| 1861 |
+
"grad_norm": 0.06706799429689687,
|
| 1862 |
+
"learning_rate": 6.054651625751717e-07,
|
| 1863 |
+
"loss": 0.5409,
|
| 1864 |
+
"step": 265
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 2.559036144578313,
|
| 1868 |
+
"grad_norm": 0.060888258525890235,
|
| 1869 |
+
"learning_rate": 5.787945663904332e-07,
|
| 1870 |
+
"loss": 0.5391,
|
| 1871 |
+
"step": 266
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 2.5686746987951805,
|
| 1875 |
+
"grad_norm": 0.06253452716963322,
|
| 1876 |
+
"learning_rate": 5.526885771648599e-07,
|
| 1877 |
+
"loss": 0.545,
|
| 1878 |
+
"step": 267
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 2.5783132530120483,
|
| 1882 |
+
"grad_norm": 0.06343440419288733,
|
| 1883 |
+
"learning_rate": 5.271505287456153e-07,
|
| 1884 |
+
"loss": 0.5832,
|
| 1885 |
+
"step": 268
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 2.5879518072289156,
|
| 1889 |
+
"grad_norm": 0.05543612918268993,
|
| 1890 |
+
"learning_rate": 5.021836824513759e-07,
|
| 1891 |
+
"loss": 0.5834,
|
| 1892 |
+
"step": 269
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 2.597590361445783,
|
| 1896 |
+
"grad_norm": 0.058093169583808696,
|
| 1897 |
+
"learning_rate": 4.777912266558532e-07,
|
| 1898 |
+
"loss": 0.5272,
|
| 1899 |
+
"step": 270
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 2.6072289156626507,
|
| 1903 |
+
"grad_norm": 0.057594604691082484,
|
| 1904 |
+
"learning_rate": 4.5397627638061604e-07,
|
| 1905 |
+
"loss": 0.5327,
|
| 1906 |
+
"step": 271
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 2.616867469879518,
|
| 1910 |
+
"grad_norm": 0.05690543891633744,
|
| 1911 |
+
"learning_rate": 4.307418728972934e-07,
|
| 1912 |
+
"loss": 0.5503,
|
| 1913 |
+
"step": 272
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 2.6265060240963853,
|
| 1917 |
+
"grad_norm": 0.06235909893309931,
|
| 1918 |
+
"learning_rate": 4.080909833391944e-07,
|
| 1919 |
+
"loss": 0.529,
|
| 1920 |
+
"step": 273
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 2.636144578313253,
|
| 1924 |
+
"grad_norm": 0.05905101458291052,
|
| 1925 |
+
"learning_rate": 3.8602650032238675e-07,
|
| 1926 |
+
"loss": 0.5298,
|
| 1927 |
+
"step": 274
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 2.6457831325301204,
|
| 1931 |
+
"grad_norm": 0.06563516042706921,
|
| 1932 |
+
"learning_rate": 3.6455124157629805e-07,
|
| 1933 |
+
"loss": 0.5732,
|
| 1934 |
+
"step": 275
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 2.6554216867469878,
|
| 1938 |
+
"grad_norm": 0.056133471314432144,
|
| 1939 |
+
"learning_rate": 3.436679495838835e-07,
|
| 1940 |
+
"loss": 0.5369,
|
| 1941 |
+
"step": 276
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 2.6650602409638555,
|
| 1945 |
+
"grad_norm": 0.07043862639489791,
|
| 1946 |
+
"learning_rate": 3.233792912313943e-07,
|
| 1947 |
+
"loss": 0.5324,
|
| 1948 |
+
"step": 277
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 2.674698795180723,
|
| 1952 |
+
"grad_norm": 0.05655745002115051,
|
| 1953 |
+
"learning_rate": 3.0368785746780925e-07,
|
| 1954 |
+
"loss": 0.5449,
|
| 1955 |
+
"step": 278
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 2.68433734939759,
|
| 1959 |
+
"grad_norm": 0.06000478684007564,
|
| 1960 |
+
"learning_rate": 2.8459616297395464e-07,
|
| 1961 |
+
"loss": 0.5378,
|
| 1962 |
+
"step": 279
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 2.693975903614458,
|
| 1966 |
+
"grad_norm": 0.06143997342419319,
|
| 1967 |
+
"learning_rate": 2.6610664584137413e-07,
|
| 1968 |
+
"loss": 0.5496,
|
| 1969 |
+
"step": 280
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 2.7036144578313253,
|
| 1973 |
+
"grad_norm": 0.0665566566789724,
|
| 1974 |
+
"learning_rate": 2.482216672609677e-07,
|
| 1975 |
+
"loss": 0.5115,
|
| 1976 |
+
"step": 281
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 2.7132530120481926,
|
| 1980 |
+
"grad_norm": 0.06198933331334414,
|
| 1981 |
+
"learning_rate": 2.3094351122146307e-07,
|
| 1982 |
+
"loss": 0.5563,
|
| 1983 |
+
"step": 282
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 2.7228915662650603,
|
| 1987 |
+
"grad_norm": 0.06294960442358041,
|
| 1988 |
+
"learning_rate": 2.142743842177386e-07,
|
| 1989 |
+
"loss": 0.5321,
|
| 1990 |
+
"step": 283
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 2.7325301204819277,
|
| 1994 |
+
"grad_norm": 0.059684494964788816,
|
| 1995 |
+
"learning_rate": 1.98216414969043e-07,
|
| 1996 |
+
"loss": 0.49,
|
| 1997 |
+
"step": 284
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 2.742168674698795,
|
| 2001 |
+
"grad_norm": 0.05972378233034468,
|
| 2002 |
+
"learning_rate": 1.8277165414714858e-07,
|
| 2003 |
+
"loss": 0.5189,
|
| 2004 |
+
"step": 285
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 2.7518072289156628,
|
| 2008 |
+
"grad_norm": 0.06503508352439365,
|
| 2009 |
+
"learning_rate": 1.6794207411447548e-07,
|
| 2010 |
+
"loss": 0.5701,
|
| 2011 |
+
"step": 286
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 2.76144578313253,
|
| 2015 |
+
"grad_norm": 0.05713319020861861,
|
| 2016 |
+
"learning_rate": 1.5372956867220678e-07,
|
| 2017 |
+
"loss": 0.576,
|
| 2018 |
+
"step": 287
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 2.7710843373493974,
|
| 2022 |
+
"grad_norm": 0.06601360939759876,
|
| 2023 |
+
"learning_rate": 1.4013595281844872e-07,
|
| 2024 |
+
"loss": 0.5726,
|
| 2025 |
+
"step": 288
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 2.780722891566265,
|
| 2029 |
+
"grad_norm": 0.052203739381335726,
|
| 2030 |
+
"learning_rate": 1.2716296251644e-07,
|
| 2031 |
+
"loss": 0.5328,
|
| 2032 |
+
"step": 289
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 2.7903614457831325,
|
| 2036 |
+
"grad_norm": 0.06980077697889618,
|
| 2037 |
+
"learning_rate": 1.1481225447286803e-07,
|
| 2038 |
+
"loss": 0.5655,
|
| 2039 |
+
"step": 290
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 2.8,
|
| 2043 |
+
"grad_norm": 0.055551268853618295,
|
| 2044 |
+
"learning_rate": 1.0308540592629756e-07,
|
| 2045 |
+
"loss": 0.5792,
|
| 2046 |
+
"step": 291
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 2.8096385542168676,
|
| 2050 |
+
"grad_norm": 0.06646819340479895,
|
| 2051 |
+
"learning_rate": 9.198391444575072e-08,
|
| 2052 |
+
"loss": 0.548,
|
| 2053 |
+
"step": 292
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 2.819277108433735,
|
| 2057 |
+
"grad_norm": 0.06410763337476913,
|
| 2058 |
+
"learning_rate": 8.150919773946165e-08,
|
| 2059 |
+
"loss": 0.5871,
|
| 2060 |
+
"step": 293
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 2.8289156626506022,
|
| 2064 |
+
"grad_norm": 0.05712141832748347,
|
| 2065 |
+
"learning_rate": 7.166259347382854e-08,
|
| 2066 |
+
"loss": 0.5526,
|
| 2067 |
+
"step": 294
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 2.83855421686747,
|
| 2071 |
+
"grad_norm": 0.06542740373895323,
|
| 2072 |
+
"learning_rate": 6.244535910258697e-08,
|
| 2073 |
+
"loss": 0.5502,
|
| 2074 |
+
"step": 295
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 2.8481927710843373,
|
| 2078 |
+
"grad_norm": 0.053860816649099204,
|
| 2079 |
+
"learning_rate": 5.3858671706230605e-08,
|
| 2080 |
+
"loss": 0.5654,
|
| 2081 |
+
"step": 296
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 2.8578313253012047,
|
| 2085 |
+
"grad_norm": 0.057862897174585454,
|
| 2086 |
+
"learning_rate": 4.590362784169022e-08,
|
| 2087 |
+
"loss": 0.5248,
|
| 2088 |
+
"step": 297
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 2.8674698795180724,
|
| 2092 |
+
"grad_norm": 0.04998420980952799,
|
| 2093 |
+
"learning_rate": 3.858124340229863e-08,
|
| 2094 |
+
"loss": 0.5558,
|
| 2095 |
+
"step": 298
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 2.8771084337349397,
|
| 2099 |
+
"grad_norm": 0.07024515673340713,
|
| 2100 |
+
"learning_rate": 3.1892453488058803e-08,
|
| 2101 |
+
"loss": 0.5335,
|
| 2102 |
+
"step": 299
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 2.886746987951807,
|
| 2106 |
+
"grad_norm": 0.05961311900445455,
|
| 2107 |
+
"learning_rate": 2.5838112286226123e-08,
|
| 2108 |
+
"loss": 0.5456,
|
| 2109 |
+
"step": 300
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 2.896385542168675,
|
| 2113 |
+
"grad_norm": 0.056751540967121886,
|
| 2114 |
+
"learning_rate": 2.0418992962224495e-08,
|
| 2115 |
+
"loss": 0.5273,
|
| 2116 |
+
"step": 301
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 2.906024096385542,
|
| 2120 |
+
"grad_norm": 0.05998655290019423,
|
| 2121 |
+
"learning_rate": 1.563578756091144e-08,
|
| 2122 |
+
"loss": 0.542,
|
| 2123 |
+
"step": 302
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 2.9156626506024095,
|
| 2127 |
+
"grad_norm": 0.05626179494647873,
|
| 2128 |
+
"learning_rate": 1.1489106918200487e-08,
|
| 2129 |
+
"loss": 0.5507,
|
| 2130 |
+
"step": 303
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 2.9253012048192772,
|
| 2134 |
+
"grad_norm": 0.0588808013463836,
|
| 2135 |
+
"learning_rate": 7.979480583052423e-09,
|
| 2136 |
+
"loss": 0.5572,
|
| 2137 |
+
"step": 304
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 2.9349397590361446,
|
| 2141 |
+
"grad_norm": 0.06020531165656935,
|
| 2142 |
+
"learning_rate": 5.107356749853298e-09,
|
| 2143 |
+
"loss": 0.5345,
|
| 2144 |
+
"step": 305
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 2.944578313253012,
|
| 2148 |
+
"grad_norm": 0.06194492848591381,
|
| 2149 |
+
"learning_rate": 2.8731022011757593e-09,
|
| 2150 |
+
"loss": 0.5595,
|
| 2151 |
+
"step": 306
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 2.9542168674698797,
|
| 2155 |
+
"grad_norm": 0.06174931334891141,
|
| 2156 |
+
"learning_rate": 1.2770022609409628e-09,
|
| 2157 |
+
"loss": 0.5209,
|
| 2158 |
+
"step": 307
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 2.963855421686747,
|
| 2162 |
+
"grad_norm": 0.0627769874173011,
|
| 2163 |
+
"learning_rate": 3.1926075797827914e-10,
|
| 2164 |
+
"loss": 0.5404,
|
| 2165 |
+
"step": 308
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 2.9734939759036143,
|
| 2169 |
+
"grad_norm": 0.056693084454889416,
|
| 2170 |
+
"learning_rate": 0.0,
|
| 2171 |
+
"loss": 0.5376,
|
| 2172 |
+
"step": 309
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 2.9734939759036143,
|
| 2176 |
+
"step": 309,
|
| 2177 |
+
"total_flos": 1.0191470808845517e+17,
|
| 2178 |
+
"train_loss": 0.6618778399857889,
|
| 2179 |
+
"train_runtime": 3030.6803,
|
| 2180 |
+
"train_samples_per_second": 4.928,
|
| 2181 |
+
"train_steps_per_second": 0.102
|
| 2182 |
+
}
|
| 2183 |
+
],
|
| 2184 |
+
"logging_steps": 1,
|
| 2185 |
+
"max_steps": 309,
|
| 2186 |
+
"num_input_tokens_seen": 0,
|
| 2187 |
+
"num_train_epochs": 3,
|
| 2188 |
+
"save_steps": 200,
|
| 2189 |
+
"stateful_callbacks": {
|
| 2190 |
+
"TrainerControl": {
|
| 2191 |
+
"args": {
|
| 2192 |
+
"should_epoch_stop": false,
|
| 2193 |
+
"should_evaluate": false,
|
| 2194 |
+
"should_log": false,
|
| 2195 |
+
"should_save": true,
|
| 2196 |
+
"should_training_stop": true
|
| 2197 |
+
},
|
| 2198 |
+
"attributes": {}
|
| 2199 |
+
}
|
| 2200 |
+
},
|
| 2201 |
+
"total_flos": 1.0191470808845517e+17,
|
| 2202 |
+
"train_batch_size": 1,
|
| 2203 |
+
"trial_name": null,
|
| 2204 |
+
"trial_params": null
|
| 2205 |
+
}
|
training_loss.png
ADDED
|