End of training
Browse files- README.md +2 -1
- all_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +3675 -0
- training_loss.png +0 -0
README.md
CHANGED
|
@@ -4,6 +4,7 @@ license: apache-2.0
|
|
| 4 |
base_model: Qwen/Qwen2.5-7B-Instruct
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
|
|
|
| 7 |
- generated_from_trainer
|
| 8 |
model-index:
|
| 9 |
- name: traceback
|
|
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
# traceback
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct) on
|
| 19 |
|
| 20 |
## Model description
|
| 21 |
|
|
|
|
| 4 |
base_model: Qwen/Qwen2.5-7B-Instruct
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
| 7 |
+
- full
|
| 8 |
- generated_from_trainer
|
| 9 |
model-index:
|
| 10 |
- name: traceback
|
|
|
|
| 16 |
|
| 17 |
# traceback
|
| 18 |
|
| 19 |
+
This model is a fine-tuned version of [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct) on the secemp9/instruction_solution_thought dataset.
|
| 20 |
|
| 21 |
## Model description
|
| 22 |
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.995670995670996,
|
| 3 |
+
"total_flos": 1.0686069936733815e+19,
|
| 4 |
+
"train_loss": 0.4367897748258073,
|
| 5 |
+
"train_runtime": 27788.7927,
|
| 6 |
+
"train_samples_per_second": 9.571,
|
| 7 |
+
"train_steps_per_second": 0.019
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.995670995670996,
|
| 3 |
+
"total_flos": 1.0686069936733815e+19,
|
| 4 |
+
"train_loss": 0.4367897748258073,
|
| 5 |
+
"train_runtime": 27788.7927,
|
| 6 |
+
"train_samples_per_second": 9.571,
|
| 7 |
+
"train_steps_per_second": 0.019
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3675 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.995670995670996,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 519,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.005772005772005772,
|
| 13 |
+
"grad_norm": 6.591821989460518,
|
| 14 |
+
"learning_rate": 1.5384615384615387e-06,
|
| 15 |
+
"loss": 0.8904,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.011544011544011544,
|
| 20 |
+
"grad_norm": 6.701083163841649,
|
| 21 |
+
"learning_rate": 3.0769230769230774e-06,
|
| 22 |
+
"loss": 0.9046,
|
| 23 |
+
"step": 2
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.017316017316017316,
|
| 27 |
+
"grad_norm": 6.536145770386053,
|
| 28 |
+
"learning_rate": 4.615384615384616e-06,
|
| 29 |
+
"loss": 0.8942,
|
| 30 |
+
"step": 3
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.023088023088023088,
|
| 34 |
+
"grad_norm": 5.047333842456431,
|
| 35 |
+
"learning_rate": 6.153846153846155e-06,
|
| 36 |
+
"loss": 0.8572,
|
| 37 |
+
"step": 4
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.02886002886002886,
|
| 41 |
+
"grad_norm": 2.468734509084157,
|
| 42 |
+
"learning_rate": 7.692307692307694e-06,
|
| 43 |
+
"loss": 0.793,
|
| 44 |
+
"step": 5
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.03463203463203463,
|
| 48 |
+
"grad_norm": 2.2050657522791153,
|
| 49 |
+
"learning_rate": 9.230769230769232e-06,
|
| 50 |
+
"loss": 0.7611,
|
| 51 |
+
"step": 6
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.04040404040404041,
|
| 55 |
+
"grad_norm": 4.090781002458348,
|
| 56 |
+
"learning_rate": 1.076923076923077e-05,
|
| 57 |
+
"loss": 0.7895,
|
| 58 |
+
"step": 7
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.046176046176046176,
|
| 62 |
+
"grad_norm": 4.080399864777092,
|
| 63 |
+
"learning_rate": 1.230769230769231e-05,
|
| 64 |
+
"loss": 0.7438,
|
| 65 |
+
"step": 8
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.05194805194805195,
|
| 69 |
+
"grad_norm": 4.592810986419584,
|
| 70 |
+
"learning_rate": 1.3846153846153847e-05,
|
| 71 |
+
"loss": 0.7438,
|
| 72 |
+
"step": 9
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.05772005772005772,
|
| 76 |
+
"grad_norm": 2.9887502837956124,
|
| 77 |
+
"learning_rate": 1.5384615384615387e-05,
|
| 78 |
+
"loss": 0.7188,
|
| 79 |
+
"step": 10
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.06349206349206349,
|
| 83 |
+
"grad_norm": 2.0334736259374444,
|
| 84 |
+
"learning_rate": 1.6923076923076924e-05,
|
| 85 |
+
"loss": 0.6873,
|
| 86 |
+
"step": 11
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.06926406926406926,
|
| 90 |
+
"grad_norm": 2.2441138864875905,
|
| 91 |
+
"learning_rate": 1.8461538461538465e-05,
|
| 92 |
+
"loss": 0.6738,
|
| 93 |
+
"step": 12
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.07503607503607504,
|
| 97 |
+
"grad_norm": 1.309205359684577,
|
| 98 |
+
"learning_rate": 2e-05,
|
| 99 |
+
"loss": 0.6549,
|
| 100 |
+
"step": 13
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.08080808080808081,
|
| 104 |
+
"grad_norm": 1.6968217490459554,
|
| 105 |
+
"learning_rate": 2.153846153846154e-05,
|
| 106 |
+
"loss": 0.6572,
|
| 107 |
+
"step": 14
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.08658008658008658,
|
| 111 |
+
"grad_norm": 1.3166881257009861,
|
| 112 |
+
"learning_rate": 2.3076923076923076e-05,
|
| 113 |
+
"loss": 0.6267,
|
| 114 |
+
"step": 15
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.09235209235209235,
|
| 118 |
+
"grad_norm": 1.058656168608237,
|
| 119 |
+
"learning_rate": 2.461538461538462e-05,
|
| 120 |
+
"loss": 0.6148,
|
| 121 |
+
"step": 16
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.09812409812409813,
|
| 125 |
+
"grad_norm": 0.9684271054203041,
|
| 126 |
+
"learning_rate": 2.6153846153846157e-05,
|
| 127 |
+
"loss": 0.6063,
|
| 128 |
+
"step": 17
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.1038961038961039,
|
| 132 |
+
"grad_norm": 0.9909925970261345,
|
| 133 |
+
"learning_rate": 2.7692307692307694e-05,
|
| 134 |
+
"loss": 0.6029,
|
| 135 |
+
"step": 18
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.10966810966810966,
|
| 139 |
+
"grad_norm": 0.9842365088549295,
|
| 140 |
+
"learning_rate": 2.923076923076923e-05,
|
| 141 |
+
"loss": 0.6049,
|
| 142 |
+
"step": 19
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.11544011544011544,
|
| 146 |
+
"grad_norm": 1.059427338447146,
|
| 147 |
+
"learning_rate": 3.0769230769230774e-05,
|
| 148 |
+
"loss": 0.5893,
|
| 149 |
+
"step": 20
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.12121212121212122,
|
| 153 |
+
"grad_norm": 0.9044592617829416,
|
| 154 |
+
"learning_rate": 3.230769230769231e-05,
|
| 155 |
+
"loss": 0.5789,
|
| 156 |
+
"step": 21
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.12698412698412698,
|
| 160 |
+
"grad_norm": 0.9839678076732425,
|
| 161 |
+
"learning_rate": 3.384615384615385e-05,
|
| 162 |
+
"loss": 0.5653,
|
| 163 |
+
"step": 22
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.13275613275613277,
|
| 167 |
+
"grad_norm": 1.3019517260461804,
|
| 168 |
+
"learning_rate": 3.538461538461539e-05,
|
| 169 |
+
"loss": 0.5784,
|
| 170 |
+
"step": 23
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.13852813852813853,
|
| 174 |
+
"grad_norm": 1.1863713203167559,
|
| 175 |
+
"learning_rate": 3.692307692307693e-05,
|
| 176 |
+
"loss": 0.5657,
|
| 177 |
+
"step": 24
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.1443001443001443,
|
| 181 |
+
"grad_norm": 1.0984531627337475,
|
| 182 |
+
"learning_rate": 3.846153846153846e-05,
|
| 183 |
+
"loss": 0.5628,
|
| 184 |
+
"step": 25
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.15007215007215008,
|
| 188 |
+
"grad_norm": 1.361248597465037,
|
| 189 |
+
"learning_rate": 4e-05,
|
| 190 |
+
"loss": 0.5589,
|
| 191 |
+
"step": 26
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.15584415584415584,
|
| 195 |
+
"grad_norm": 0.9699622745779369,
|
| 196 |
+
"learning_rate": 4.1538461538461544e-05,
|
| 197 |
+
"loss": 0.5587,
|
| 198 |
+
"step": 27
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.16161616161616163,
|
| 202 |
+
"grad_norm": 1.6250628750726255,
|
| 203 |
+
"learning_rate": 4.307692307692308e-05,
|
| 204 |
+
"loss": 0.5533,
|
| 205 |
+
"step": 28
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.1673881673881674,
|
| 209 |
+
"grad_norm": 0.8965839274541476,
|
| 210 |
+
"learning_rate": 4.461538461538462e-05,
|
| 211 |
+
"loss": 0.5532,
|
| 212 |
+
"step": 29
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.17316017316017315,
|
| 216 |
+
"grad_norm": 2.2172027740776237,
|
| 217 |
+
"learning_rate": 4.615384615384615e-05,
|
| 218 |
+
"loss": 0.5516,
|
| 219 |
+
"step": 30
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.17893217893217894,
|
| 223 |
+
"grad_norm": 1.4206471782118077,
|
| 224 |
+
"learning_rate": 4.76923076923077e-05,
|
| 225 |
+
"loss": 0.5443,
|
| 226 |
+
"step": 31
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.1847041847041847,
|
| 230 |
+
"grad_norm": 1.953518854356112,
|
| 231 |
+
"learning_rate": 4.923076923076924e-05,
|
| 232 |
+
"loss": 0.5457,
|
| 233 |
+
"step": 32
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.19047619047619047,
|
| 237 |
+
"grad_norm": 1.7035803319314056,
|
| 238 |
+
"learning_rate": 5.076923076923077e-05,
|
| 239 |
+
"loss": 0.5452,
|
| 240 |
+
"step": 33
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.19624819624819625,
|
| 244 |
+
"grad_norm": 1.3778506162803026,
|
| 245 |
+
"learning_rate": 5.230769230769231e-05,
|
| 246 |
+
"loss": 0.5386,
|
| 247 |
+
"step": 34
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.20202020202020202,
|
| 251 |
+
"grad_norm": 2.2007791090752273,
|
| 252 |
+
"learning_rate": 5.3846153846153853e-05,
|
| 253 |
+
"loss": 0.5326,
|
| 254 |
+
"step": 35
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.2077922077922078,
|
| 258 |
+
"grad_norm": 1.3929576225978784,
|
| 259 |
+
"learning_rate": 5.538461538461539e-05,
|
| 260 |
+
"loss": 0.5445,
|
| 261 |
+
"step": 36
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.21356421356421357,
|
| 265 |
+
"grad_norm": 2.5179570562309475,
|
| 266 |
+
"learning_rate": 5.692307692307693e-05,
|
| 267 |
+
"loss": 0.5395,
|
| 268 |
+
"step": 37
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.21933621933621933,
|
| 272 |
+
"grad_norm": 1.958838600657758,
|
| 273 |
+
"learning_rate": 5.846153846153846e-05,
|
| 274 |
+
"loss": 0.5288,
|
| 275 |
+
"step": 38
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.22510822510822512,
|
| 279 |
+
"grad_norm": 2.2628335323973,
|
| 280 |
+
"learning_rate": 6.000000000000001e-05,
|
| 281 |
+
"loss": 0.5204,
|
| 282 |
+
"step": 39
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.23088023088023088,
|
| 286 |
+
"grad_norm": 1.834622802154548,
|
| 287 |
+
"learning_rate": 6.153846153846155e-05,
|
| 288 |
+
"loss": 0.5261,
|
| 289 |
+
"step": 40
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.23665223665223664,
|
| 293 |
+
"grad_norm": 2.0831656414023443,
|
| 294 |
+
"learning_rate": 6.307692307692308e-05,
|
| 295 |
+
"loss": 0.5316,
|
| 296 |
+
"step": 41
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.24242424242424243,
|
| 300 |
+
"grad_norm": 1.6649570041240862,
|
| 301 |
+
"learning_rate": 6.461538461538462e-05,
|
| 302 |
+
"loss": 0.5298,
|
| 303 |
+
"step": 42
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.2481962481962482,
|
| 307 |
+
"grad_norm": 1.8210200420604277,
|
| 308 |
+
"learning_rate": 6.615384615384616e-05,
|
| 309 |
+
"loss": 0.5289,
|
| 310 |
+
"step": 43
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.25396825396825395,
|
| 314 |
+
"grad_norm": 1.232981446929873,
|
| 315 |
+
"learning_rate": 6.76923076923077e-05,
|
| 316 |
+
"loss": 0.5155,
|
| 317 |
+
"step": 44
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.2597402597402597,
|
| 321 |
+
"grad_norm": 1.7708070269877163,
|
| 322 |
+
"learning_rate": 6.923076923076924e-05,
|
| 323 |
+
"loss": 0.5311,
|
| 324 |
+
"step": 45
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.26551226551226553,
|
| 328 |
+
"grad_norm": 1.2424713252981978,
|
| 329 |
+
"learning_rate": 7.076923076923078e-05,
|
| 330 |
+
"loss": 0.5244,
|
| 331 |
+
"step": 46
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.2712842712842713,
|
| 335 |
+
"grad_norm": 1.917850479797542,
|
| 336 |
+
"learning_rate": 7.230769230769232e-05,
|
| 337 |
+
"loss": 0.5229,
|
| 338 |
+
"step": 47
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.27705627705627706,
|
| 342 |
+
"grad_norm": 1.4820579510034553,
|
| 343 |
+
"learning_rate": 7.384615384615386e-05,
|
| 344 |
+
"loss": 0.5222,
|
| 345 |
+
"step": 48
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.2828282828282828,
|
| 349 |
+
"grad_norm": 1.3751360045514713,
|
| 350 |
+
"learning_rate": 7.538461538461539e-05,
|
| 351 |
+
"loss": 0.5127,
|
| 352 |
+
"step": 49
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.2886002886002886,
|
| 356 |
+
"grad_norm": 2.4703954526048792,
|
| 357 |
+
"learning_rate": 7.692307692307693e-05,
|
| 358 |
+
"loss": 0.5358,
|
| 359 |
+
"step": 50
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.2943722943722944,
|
| 363 |
+
"grad_norm": 1.1276824883035388,
|
| 364 |
+
"learning_rate": 7.846153846153847e-05,
|
| 365 |
+
"loss": 0.5202,
|
| 366 |
+
"step": 51
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.30014430014430016,
|
| 370 |
+
"grad_norm": 1.9518381050342926,
|
| 371 |
+
"learning_rate": 8e-05,
|
| 372 |
+
"loss": 0.5231,
|
| 373 |
+
"step": 52
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.3059163059163059,
|
| 377 |
+
"grad_norm": 1.6826277240274867,
|
| 378 |
+
"learning_rate": 7.999909490463248e-05,
|
| 379 |
+
"loss": 0.5266,
|
| 380 |
+
"step": 53
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.3116883116883117,
|
| 384 |
+
"grad_norm": 1.738831578332151,
|
| 385 |
+
"learning_rate": 7.999637965948977e-05,
|
| 386 |
+
"loss": 0.5092,
|
| 387 |
+
"step": 54
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.31746031746031744,
|
| 391 |
+
"grad_norm": 1.4323939653994584,
|
| 392 |
+
"learning_rate": 7.999185438744968e-05,
|
| 393 |
+
"loss": 0.5095,
|
| 394 |
+
"step": 55
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.32323232323232326,
|
| 398 |
+
"grad_norm": 1.3664334433498815,
|
| 399 |
+
"learning_rate": 7.998551929330234e-05,
|
| 400 |
+
"loss": 0.5114,
|
| 401 |
+
"step": 56
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.329004329004329,
|
| 405 |
+
"grad_norm": 1.1542720966817734,
|
| 406 |
+
"learning_rate": 7.997737466374096e-05,
|
| 407 |
+
"loss": 0.5091,
|
| 408 |
+
"step": 57
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.3347763347763348,
|
| 412 |
+
"grad_norm": 1.807835646103163,
|
| 413 |
+
"learning_rate": 7.99674208673489e-05,
|
| 414 |
+
"loss": 0.502,
|
| 415 |
+
"step": 58
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.34054834054834054,
|
| 419 |
+
"grad_norm": 1.4816405608909524,
|
| 420 |
+
"learning_rate": 7.995565835458286e-05,
|
| 421 |
+
"loss": 0.5007,
|
| 422 |
+
"step": 59
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.3463203463203463,
|
| 426 |
+
"grad_norm": 2.072657075856209,
|
| 427 |
+
"learning_rate": 7.994208765775267e-05,
|
| 428 |
+
"loss": 0.5142,
|
| 429 |
+
"step": 60
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.35209235209235207,
|
| 433 |
+
"grad_norm": 1.3835777391967627,
|
| 434 |
+
"learning_rate": 7.992670939099704e-05,
|
| 435 |
+
"loss": 0.5039,
|
| 436 |
+
"step": 61
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.3578643578643579,
|
| 440 |
+
"grad_norm": 1.315993786091659,
|
| 441 |
+
"learning_rate": 7.99095242502559e-05,
|
| 442 |
+
"loss": 0.5018,
|
| 443 |
+
"step": 62
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.36363636363636365,
|
| 447 |
+
"grad_norm": 1.350155329631968,
|
| 448 |
+
"learning_rate": 7.989053301323881e-05,
|
| 449 |
+
"loss": 0.4986,
|
| 450 |
+
"step": 63
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.3694083694083694,
|
| 454 |
+
"grad_norm": 1.5446010928098293,
|
| 455 |
+
"learning_rate": 7.986973653938977e-05,
|
| 456 |
+
"loss": 0.4983,
|
| 457 |
+
"step": 64
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.37518037518037517,
|
| 461 |
+
"grad_norm": 1.04496723786831,
|
| 462 |
+
"learning_rate": 7.984713576984842e-05,
|
| 463 |
+
"loss": 0.509,
|
| 464 |
+
"step": 65
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.38095238095238093,
|
| 468 |
+
"grad_norm": 1.8739311708739077,
|
| 469 |
+
"learning_rate": 7.982273172740734e-05,
|
| 470 |
+
"loss": 0.5126,
|
| 471 |
+
"step": 66
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.38672438672438675,
|
| 475 |
+
"grad_norm": 1.3652541433364316,
|
| 476 |
+
"learning_rate": 7.979652551646583e-05,
|
| 477 |
+
"loss": 0.5,
|
| 478 |
+
"step": 67
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.3924963924963925,
|
| 482 |
+
"grad_norm": 1.9130220059566032,
|
| 483 |
+
"learning_rate": 7.976851832297987e-05,
|
| 484 |
+
"loss": 0.5132,
|
| 485 |
+
"step": 68
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.39826839826839827,
|
| 489 |
+
"grad_norm": 1.1729154565908322,
|
| 490 |
+
"learning_rate": 7.973871141440853e-05,
|
| 491 |
+
"loss": 0.5052,
|
| 492 |
+
"step": 69
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.40404040404040403,
|
| 496 |
+
"grad_norm": 1.7055313247779371,
|
| 497 |
+
"learning_rate": 7.970710613965657e-05,
|
| 498 |
+
"loss": 0.5078,
|
| 499 |
+
"step": 70
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.4098124098124098,
|
| 503 |
+
"grad_norm": 1.3935437350999336,
|
| 504 |
+
"learning_rate": 7.967370392901334e-05,
|
| 505 |
+
"loss": 0.5023,
|
| 506 |
+
"step": 71
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.4155844155844156,
|
| 510 |
+
"grad_norm": 245.851096759516,
|
| 511 |
+
"learning_rate": 7.963850629408817e-05,
|
| 512 |
+
"loss": 7.192,
|
| 513 |
+
"step": 72
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.4213564213564214,
|
| 517 |
+
"grad_norm": 3.5843288436560656,
|
| 518 |
+
"learning_rate": 7.960151482774188e-05,
|
| 519 |
+
"loss": 0.5501,
|
| 520 |
+
"step": 73
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.42712842712842713,
|
| 524 |
+
"grad_norm": 1.8159995585862374,
|
| 525 |
+
"learning_rate": 7.95627312040147e-05,
|
| 526 |
+
"loss": 0.5197,
|
| 527 |
+
"step": 74
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.4329004329004329,
|
| 531 |
+
"grad_norm": 1.7593611646948464,
|
| 532 |
+
"learning_rate": 7.952215717805055e-05,
|
| 533 |
+
"loss": 0.5043,
|
| 534 |
+
"step": 75
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.43867243867243866,
|
| 538 |
+
"grad_norm": 2.554507256709062,
|
| 539 |
+
"learning_rate": 7.947979458601756e-05,
|
| 540 |
+
"loss": 0.5184,
|
| 541 |
+
"step": 76
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.4444444444444444,
|
| 545 |
+
"grad_norm": 1.7433980613956555,
|
| 546 |
+
"learning_rate": 7.943564534502503e-05,
|
| 547 |
+
"loss": 0.5184,
|
| 548 |
+
"step": 77
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.45021645021645024,
|
| 552 |
+
"grad_norm": 0.9438409442305278,
|
| 553 |
+
"learning_rate": 7.938971145303666e-05,
|
| 554 |
+
"loss": 0.5054,
|
| 555 |
+
"step": 78
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.455988455988456,
|
| 559 |
+
"grad_norm": 1.1198133821796519,
|
| 560 |
+
"learning_rate": 7.934199498878005e-05,
|
| 561 |
+
"loss": 0.4974,
|
| 562 |
+
"step": 79
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.46176046176046176,
|
| 566 |
+
"grad_norm": 1.2641541874296718,
|
| 567 |
+
"learning_rate": 7.929249811165274e-05,
|
| 568 |
+
"loss": 0.4979,
|
| 569 |
+
"step": 80
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.4675324675324675,
|
| 573 |
+
"grad_norm": 4.739955304345111,
|
| 574 |
+
"learning_rate": 7.924122306162448e-05,
|
| 575 |
+
"loss": 0.5227,
|
| 576 |
+
"step": 81
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.4733044733044733,
|
| 580 |
+
"grad_norm": 3.777145023653992,
|
| 581 |
+
"learning_rate": 7.918817215913574e-05,
|
| 582 |
+
"loss": 0.5384,
|
| 583 |
+
"step": 82
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.4790764790764791,
|
| 587 |
+
"grad_norm": 1.897498471819791,
|
| 588 |
+
"learning_rate": 7.913334780499284e-05,
|
| 589 |
+
"loss": 0.5193,
|
| 590 |
+
"step": 83
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.48484848484848486,
|
| 594 |
+
"grad_norm": 1.9166167580469424,
|
| 595 |
+
"learning_rate": 7.907675248025926e-05,
|
| 596 |
+
"loss": 0.5178,
|
| 597 |
+
"step": 84
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.4906204906204906,
|
| 601 |
+
"grad_norm": 111.71899143551498,
|
| 602 |
+
"learning_rate": 7.901838874614325e-05,
|
| 603 |
+
"loss": 0.565,
|
| 604 |
+
"step": 85
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.4963924963924964,
|
| 608 |
+
"grad_norm": 12.497569359247084,
|
| 609 |
+
"learning_rate": 7.895825924388214e-05,
|
| 610 |
+
"loss": 0.574,
|
| 611 |
+
"step": 86
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.5021645021645021,
|
| 615 |
+
"grad_norm": 2.309806827904168,
|
| 616 |
+
"learning_rate": 7.889636669462257e-05,
|
| 617 |
+
"loss": 0.5636,
|
| 618 |
+
"step": 87
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.5079365079365079,
|
| 622 |
+
"grad_norm": 1.3442519653064227,
|
| 623 |
+
"learning_rate": 7.883271389929755e-05,
|
| 624 |
+
"loss": 0.5266,
|
| 625 |
+
"step": 88
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.5137085137085137,
|
| 629 |
+
"grad_norm": 1.5607960630672684,
|
| 630 |
+
"learning_rate": 7.87673037384996e-05,
|
| 631 |
+
"loss": 0.5151,
|
| 632 |
+
"step": 89
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.5194805194805194,
|
| 636 |
+
"grad_norm": 1.2317264148870677,
|
| 637 |
+
"learning_rate": 7.870013917235039e-05,
|
| 638 |
+
"loss": 0.5138,
|
| 639 |
+
"step": 90
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.5252525252525253,
|
| 643 |
+
"grad_norm": 1.2174210025618513,
|
| 644 |
+
"learning_rate": 7.86312232403668e-05,
|
| 645 |
+
"loss": 0.5084,
|
| 646 |
+
"step": 91
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.5310245310245311,
|
| 650 |
+
"grad_norm": 0.9844891141990263,
|
| 651 |
+
"learning_rate": 7.856055906132337e-05,
|
| 652 |
+
"loss": 0.5013,
|
| 653 |
+
"step": 92
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.5367965367965368,
|
| 657 |
+
"grad_norm": 1.6249846234942358,
|
| 658 |
+
"learning_rate": 7.848814983311114e-05,
|
| 659 |
+
"loss": 0.4875,
|
| 660 |
+
"step": 93
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.5425685425685426,
|
| 664 |
+
"grad_norm": 1.110842702049686,
|
| 665 |
+
"learning_rate": 7.841399883259298e-05,
|
| 666 |
+
"loss": 0.5072,
|
| 667 |
+
"step": 94
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.5483405483405484,
|
| 671 |
+
"grad_norm": 1.2409317588868078,
|
| 672 |
+
"learning_rate": 7.833810941545525e-05,
|
| 673 |
+
"loss": 0.4981,
|
| 674 |
+
"step": 95
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.5541125541125541,
|
| 678 |
+
"grad_norm": 0.9216863812635485,
|
| 679 |
+
"learning_rate": 7.826048501605592e-05,
|
| 680 |
+
"loss": 0.4976,
|
| 681 |
+
"step": 96
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.5598845598845599,
|
| 685 |
+
"grad_norm": 1.1689690422801344,
|
| 686 |
+
"learning_rate": 7.818112914726922e-05,
|
| 687 |
+
"loss": 0.4939,
|
| 688 |
+
"step": 97
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.5656565656565656,
|
| 692 |
+
"grad_norm": 1.3912487511676008,
|
| 693 |
+
"learning_rate": 7.81000454003266e-05,
|
| 694 |
+
"loss": 0.4973,
|
| 695 |
+
"step": 98
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.5714285714285714,
|
| 699 |
+
"grad_norm": 1.5119375553594363,
|
| 700 |
+
"learning_rate": 7.801723744465427e-05,
|
| 701 |
+
"loss": 0.4877,
|
| 702 |
+
"step": 99
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.5772005772005772,
|
| 706 |
+
"grad_norm": 1.1451986843068807,
|
| 707 |
+
"learning_rate": 7.793270902770707e-05,
|
| 708 |
+
"loss": 0.4847,
|
| 709 |
+
"step": 100
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.5829725829725829,
|
| 713 |
+
"grad_norm": 0.8253514693262355,
|
| 714 |
+
"learning_rate": 7.784646397479892e-05,
|
| 715 |
+
"loss": 0.4789,
|
| 716 |
+
"step": 101
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.5887445887445888,
|
| 720 |
+
"grad_norm": 1.0501024034599673,
|
| 721 |
+
"learning_rate": 7.775850618892973e-05,
|
| 722 |
+
"loss": 0.4826,
|
| 723 |
+
"step": 102
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.5945165945165946,
|
| 727 |
+
"grad_norm": 1.6363531551977177,
|
| 728 |
+
"learning_rate": 7.766883965060871e-05,
|
| 729 |
+
"loss": 0.4801,
|
| 730 |
+
"step": 103
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.6002886002886003,
|
| 734 |
+
"grad_norm": 0.9154587109673673,
|
| 735 |
+
"learning_rate": 7.757746841767431e-05,
|
| 736 |
+
"loss": 0.4871,
|
| 737 |
+
"step": 104
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.6060606060606061,
|
| 741 |
+
"grad_norm": 1.5192770236375652,
|
| 742 |
+
"learning_rate": 7.748439662511049e-05,
|
| 743 |
+
"loss": 0.4835,
|
| 744 |
+
"step": 105
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.6118326118326118,
|
| 748 |
+
"grad_norm": 0.9122272724371713,
|
| 749 |
+
"learning_rate": 7.738962848485967e-05,
|
| 750 |
+
"loss": 0.4768,
|
| 751 |
+
"step": 106
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.6176046176046176,
|
| 755 |
+
"grad_norm": 1.2673986949755514,
|
| 756 |
+
"learning_rate": 7.729316828563207e-05,
|
| 757 |
+
"loss": 0.4824,
|
| 758 |
+
"step": 107
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.6233766233766234,
|
| 762 |
+
"grad_norm": 1.6724258053369745,
|
| 763 |
+
"learning_rate": 7.71950203927117e-05,
|
| 764 |
+
"loss": 0.4826,
|
| 765 |
+
"step": 108
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.6291486291486291,
|
| 769 |
+
"grad_norm": 0.8735759562195867,
|
| 770 |
+
"learning_rate": 7.709518924775869e-05,
|
| 771 |
+
"loss": 0.4903,
|
| 772 |
+
"step": 109
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.6349206349206349,
|
| 776 |
+
"grad_norm": 1.9881565552085574,
|
| 777 |
+
"learning_rate": 7.69936793686084e-05,
|
| 778 |
+
"loss": 0.4827,
|
| 779 |
+
"step": 110
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.6406926406926406,
|
| 783 |
+
"grad_norm": 1.1525310802346136,
|
| 784 |
+
"learning_rate": 7.689049534906688e-05,
|
| 785 |
+
"loss": 0.4839,
|
| 786 |
+
"step": 111
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.6464646464646465,
|
| 790 |
+
"grad_norm": 1.6737242634010223,
|
| 791 |
+
"learning_rate": 7.678564185870306e-05,
|
| 792 |
+
"loss": 0.4905,
|
| 793 |
+
"step": 112
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.6522366522366523,
|
| 797 |
+
"grad_norm": 1.4430692554643452,
|
| 798 |
+
"learning_rate": 7.667912364263734e-05,
|
| 799 |
+
"loss": 0.4935,
|
| 800 |
+
"step": 113
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.658008658008658,
|
| 804 |
+
"grad_norm": 1.0555012525229592,
|
| 805 |
+
"learning_rate": 7.657094552132692e-05,
|
| 806 |
+
"loss": 0.4753,
|
| 807 |
+
"step": 114
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.6637806637806638,
|
| 811 |
+
"grad_norm": 1.2792144841892175,
|
| 812 |
+
"learning_rate": 7.646111239034762e-05,
|
| 813 |
+
"loss": 0.4847,
|
| 814 |
+
"step": 115
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.6695526695526696,
|
| 818 |
+
"grad_norm": 0.8673391499660823,
|
| 819 |
+
"learning_rate": 7.634962922017234e-05,
|
| 820 |
+
"loss": 0.4812,
|
| 821 |
+
"step": 116
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.6753246753246753,
|
| 825 |
+
"grad_norm": 1.450882719103234,
|
| 826 |
+
"learning_rate": 7.623650105594616e-05,
|
| 827 |
+
"loss": 0.4789,
|
| 828 |
+
"step": 117
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.6810966810966811,
|
| 832 |
+
"grad_norm": 1.1189212432889974,
|
| 833 |
+
"learning_rate": 7.612173301725791e-05,
|
| 834 |
+
"loss": 0.4782,
|
| 835 |
+
"step": 118
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.6868686868686869,
|
| 839 |
+
"grad_norm": 1.2500032375967438,
|
| 840 |
+
"learning_rate": 7.600533029790859e-05,
|
| 841 |
+
"loss": 0.4766,
|
| 842 |
+
"step": 119
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.6926406926406926,
|
| 846 |
+
"grad_norm": 0.9200876953926456,
|
| 847 |
+
"learning_rate": 7.588729816567634e-05,
|
| 848 |
+
"loss": 0.4742,
|
| 849 |
+
"step": 120
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.6984126984126984,
|
| 853 |
+
"grad_norm": 1.1553568867731325,
|
| 854 |
+
"learning_rate": 7.576764196207793e-05,
|
| 855 |
+
"loss": 0.4715,
|
| 856 |
+
"step": 121
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.7041847041847041,
|
| 860 |
+
"grad_norm": 0.8934619182669754,
|
| 861 |
+
"learning_rate": 7.564636710212717e-05,
|
| 862 |
+
"loss": 0.4727,
|
| 863 |
+
"step": 122
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.70995670995671,
|
| 867 |
+
"grad_norm": 1.0904046636453852,
|
| 868 |
+
"learning_rate": 7.552347907408973e-05,
|
| 869 |
+
"loss": 0.4679,
|
| 870 |
+
"step": 123
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.7157287157287158,
|
| 874 |
+
"grad_norm": 1.3764671359522456,
|
| 875 |
+
"learning_rate": 7.539898343923487e-05,
|
| 876 |
+
"loss": 0.4634,
|
| 877 |
+
"step": 124
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.7215007215007215,
|
| 881 |
+
"grad_norm": 1.171521245459155,
|
| 882 |
+
"learning_rate": 7.527288583158371e-05,
|
| 883 |
+
"loss": 0.4685,
|
| 884 |
+
"step": 125
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.7272727272727273,
|
| 888 |
+
"grad_norm": 0.7696078257367441,
|
| 889 |
+
"learning_rate": 7.514519195765428e-05,
|
| 890 |
+
"loss": 0.4673,
|
| 891 |
+
"step": 126
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.733044733044733,
|
| 895 |
+
"grad_norm": 1.1883468674609674,
|
| 896 |
+
"learning_rate": 7.501590759620324e-05,
|
| 897 |
+
"loss": 0.4725,
|
| 898 |
+
"step": 127
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.7388167388167388,
|
| 902 |
+
"grad_norm": 0.9291536438970882,
|
| 903 |
+
"learning_rate": 7.488503859796446e-05,
|
| 904 |
+
"loss": 0.4621,
|
| 905 |
+
"step": 128
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.7445887445887446,
|
| 909 |
+
"grad_norm": 1.0627658395705477,
|
| 910 |
+
"learning_rate": 7.475259088538412e-05,
|
| 911 |
+
"loss": 0.4565,
|
| 912 |
+
"step": 129
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.7503607503607503,
|
| 916 |
+
"grad_norm": 1.54642074651037,
|
| 917 |
+
"learning_rate": 7.461857045235278e-05,
|
| 918 |
+
"loss": 0.4643,
|
| 919 |
+
"step": 130
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.7561327561327561,
|
| 923 |
+
"grad_norm": 0.6810343830586086,
|
| 924 |
+
"learning_rate": 7.448298336393409e-05,
|
| 925 |
+
"loss": 0.4661,
|
| 926 |
+
"step": 131
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.7619047619047619,
|
| 930 |
+
"grad_norm": 0.6372542725369494,
|
| 931 |
+
"learning_rate": 7.434583575609034e-05,
|
| 932 |
+
"loss": 0.4653,
|
| 933 |
+
"step": 132
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.7676767676767676,
|
| 937 |
+
"grad_norm": 1.0126413662175853,
|
| 938 |
+
"learning_rate": 7.420713383540476e-05,
|
| 939 |
+
"loss": 0.4693,
|
| 940 |
+
"step": 133
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.7734487734487735,
|
| 944 |
+
"grad_norm": 1.3737542427125524,
|
| 945 |
+
"learning_rate": 7.406688387880064e-05,
|
| 946 |
+
"loss": 0.4623,
|
| 947 |
+
"step": 134
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.7792207792207793,
|
| 951 |
+
"grad_norm": 0.9130609282834156,
|
| 952 |
+
"learning_rate": 7.392509223325727e-05,
|
| 953 |
+
"loss": 0.4625,
|
| 954 |
+
"step": 135
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.784992784992785,
|
| 958 |
+
"grad_norm": 1.9579393793658253,
|
| 959 |
+
"learning_rate": 7.378176531552272e-05,
|
| 960 |
+
"loss": 0.4695,
|
| 961 |
+
"step": 136
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.7907647907647908,
|
| 965 |
+
"grad_norm": 0.8632201170827392,
|
| 966 |
+
"learning_rate": 7.363690961182348e-05,
|
| 967 |
+
"loss": 0.4603,
|
| 968 |
+
"step": 137
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.7965367965367965,
|
| 972 |
+
"grad_norm": 2.814682600145235,
|
| 973 |
+
"learning_rate": 7.349053167757087e-05,
|
| 974 |
+
"loss": 0.4729,
|
| 975 |
+
"step": 138
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.8023088023088023,
|
| 979 |
+
"grad_norm": 2.2606064260809635,
|
| 980 |
+
"learning_rate": 7.334263813706438e-05,
|
| 981 |
+
"loss": 0.4901,
|
| 982 |
+
"step": 139
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.8080808080808081,
|
| 986 |
+
"grad_norm": 1.7214477605860958,
|
| 987 |
+
"learning_rate": 7.319323568319195e-05,
|
| 988 |
+
"loss": 0.4643,
|
| 989 |
+
"step": 140
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.8138528138528138,
|
| 993 |
+
"grad_norm": 1.7622962795748711,
|
| 994 |
+
"learning_rate": 7.304233107712699e-05,
|
| 995 |
+
"loss": 0.4732,
|
| 996 |
+
"step": 141
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.8196248196248196,
|
| 1000 |
+
"grad_norm": 1.5507278602835506,
|
| 1001 |
+
"learning_rate": 7.288993114802251e-05,
|
| 1002 |
+
"loss": 0.4805,
|
| 1003 |
+
"step": 142
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.8253968253968254,
|
| 1007 |
+
"grad_norm": 1.0461957792765249,
|
| 1008 |
+
"learning_rate": 7.273604279270203e-05,
|
| 1009 |
+
"loss": 0.4575,
|
| 1010 |
+
"step": 143
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.8311688311688312,
|
| 1014 |
+
"grad_norm": 1.4147827443269227,
|
| 1015 |
+
"learning_rate": 7.25806729753474e-05,
|
| 1016 |
+
"loss": 0.461,
|
| 1017 |
+
"step": 144
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.836940836940837,
|
| 1021 |
+
"grad_norm": 0.8903624685904831,
|
| 1022 |
+
"learning_rate": 7.242382872718371e-05,
|
| 1023 |
+
"loss": 0.4587,
|
| 1024 |
+
"step": 145
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.8427128427128427,
|
| 1028 |
+
"grad_norm": 1.534917997797476,
|
| 1029 |
+
"learning_rate": 7.226551714616111e-05,
|
| 1030 |
+
"loss": 0.4647,
|
| 1031 |
+
"step": 146
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.8484848484848485,
|
| 1035 |
+
"grad_norm": 0.857263805930584,
|
| 1036 |
+
"learning_rate": 7.210574539663351e-05,
|
| 1037 |
+
"loss": 0.4595,
|
| 1038 |
+
"step": 147
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.8542568542568543,
|
| 1042 |
+
"grad_norm": 1.4926745213805535,
|
| 1043 |
+
"learning_rate": 7.194452070903443e-05,
|
| 1044 |
+
"loss": 0.4592,
|
| 1045 |
+
"step": 148
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.86002886002886,
|
| 1049 |
+
"grad_norm": 0.9521702084080935,
|
| 1050 |
+
"learning_rate": 7.178185037954977e-05,
|
| 1051 |
+
"loss": 0.4556,
|
| 1052 |
+
"step": 149
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.8658008658008658,
|
| 1056 |
+
"grad_norm": 1.5854019697451536,
|
| 1057 |
+
"learning_rate": 7.16177417697876e-05,
|
| 1058 |
+
"loss": 0.461,
|
| 1059 |
+
"step": 150
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.8715728715728716,
|
| 1063 |
+
"grad_norm": 1.0922004596100605,
|
| 1064 |
+
"learning_rate": 7.145220230644507e-05,
|
| 1065 |
+
"loss": 0.4597,
|
| 1066 |
+
"step": 151
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.8773448773448773,
|
| 1070 |
+
"grad_norm": 1.369601810507737,
|
| 1071 |
+
"learning_rate": 7.128523948097222e-05,
|
| 1072 |
+
"loss": 0.4626,
|
| 1073 |
+
"step": 152
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.8831168831168831,
|
| 1077 |
+
"grad_norm": 1.2602330173138279,
|
| 1078 |
+
"learning_rate": 7.111686084923306e-05,
|
| 1079 |
+
"loss": 0.4599,
|
| 1080 |
+
"step": 153
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.8888888888888888,
|
| 1084 |
+
"grad_norm": 1.194998174450651,
|
| 1085 |
+
"learning_rate": 7.094707403116356e-05,
|
| 1086 |
+
"loss": 0.4595,
|
| 1087 |
+
"step": 154
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.8946608946608947,
|
| 1091 |
+
"grad_norm": 0.9745804896707817,
|
| 1092 |
+
"learning_rate": 7.077588671042686e-05,
|
| 1093 |
+
"loss": 0.4544,
|
| 1094 |
+
"step": 155
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.9004329004329005,
|
| 1098 |
+
"grad_norm": 1.0884187236190401,
|
| 1099 |
+
"learning_rate": 7.06033066340655e-05,
|
| 1100 |
+
"loss": 0.4435,
|
| 1101 |
+
"step": 156
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.9062049062049062,
|
| 1105 |
+
"grad_norm": 0.8051773982771812,
|
| 1106 |
+
"learning_rate": 7.042934161215085e-05,
|
| 1107 |
+
"loss": 0.4608,
|
| 1108 |
+
"step": 157
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.911976911976912,
|
| 1112 |
+
"grad_norm": 0.9378153819388926,
|
| 1113 |
+
"learning_rate": 7.02539995174297e-05,
|
| 1114 |
+
"loss": 0.4583,
|
| 1115 |
+
"step": 158
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.9177489177489178,
|
| 1119 |
+
"grad_norm": 0.6989550093651844,
|
| 1120 |
+
"learning_rate": 7.007728828496793e-05,
|
| 1121 |
+
"loss": 0.4524,
|
| 1122 |
+
"step": 159
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.9235209235209235,
|
| 1126 |
+
"grad_norm": 0.8390345572841903,
|
| 1127 |
+
"learning_rate": 6.989921591179145e-05,
|
| 1128 |
+
"loss": 0.4519,
|
| 1129 |
+
"step": 160
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.9292929292929293,
|
| 1133 |
+
"grad_norm": 0.6441569982951852,
|
| 1134 |
+
"learning_rate": 6.97197904565242e-05,
|
| 1135 |
+
"loss": 0.4504,
|
| 1136 |
+
"step": 161
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.935064935064935,
|
| 1140 |
+
"grad_norm": 0.9443858881471812,
|
| 1141 |
+
"learning_rate": 6.953902003902368e-05,
|
| 1142 |
+
"loss": 0.4535,
|
| 1143 |
+
"step": 162
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.9408369408369408,
|
| 1147 |
+
"grad_norm": 0.798028444293174,
|
| 1148 |
+
"learning_rate": 6.935691284001322e-05,
|
| 1149 |
+
"loss": 0.4495,
|
| 1150 |
+
"step": 163
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.9466089466089466,
|
| 1154 |
+
"grad_norm": 0.8082815856339027,
|
| 1155 |
+
"learning_rate": 6.917347710071193e-05,
|
| 1156 |
+
"loss": 0.4466,
|
| 1157 |
+
"step": 164
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.9523809523809523,
|
| 1161 |
+
"grad_norm": 1.1854187052850362,
|
| 1162 |
+
"learning_rate": 6.898872112246172e-05,
|
| 1163 |
+
"loss": 0.4458,
|
| 1164 |
+
"step": 165
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.9581529581529582,
|
| 1168 |
+
"grad_norm": 1.4406146951543184,
|
| 1169 |
+
"learning_rate": 6.880265326635159e-05,
|
| 1170 |
+
"loss": 0.4537,
|
| 1171 |
+
"step": 166
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.963924963924964,
|
| 1175 |
+
"grad_norm": 0.5540098885616462,
|
| 1176 |
+
"learning_rate": 6.861528195283924e-05,
|
| 1177 |
+
"loss": 0.4392,
|
| 1178 |
+
"step": 167
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.9696969696969697,
|
| 1182 |
+
"grad_norm": 1.4158419971388627,
|
| 1183 |
+
"learning_rate": 6.842661566137011e-05,
|
| 1184 |
+
"loss": 0.4453,
|
| 1185 |
+
"step": 168
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.9754689754689755,
|
| 1189 |
+
"grad_norm": 0.9692734629530391,
|
| 1190 |
+
"learning_rate": 6.82366629299935e-05,
|
| 1191 |
+
"loss": 0.4534,
|
| 1192 |
+
"step": 169
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.9812409812409812,
|
| 1196 |
+
"grad_norm": 1.0422752052381763,
|
| 1197 |
+
"learning_rate": 6.804543235497625e-05,
|
| 1198 |
+
"loss": 0.4427,
|
| 1199 |
+
"step": 170
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.987012987012987,
|
| 1203 |
+
"grad_norm": 1.0057520882099593,
|
| 1204 |
+
"learning_rate": 6.785293259041376e-05,
|
| 1205 |
+
"loss": 0.4388,
|
| 1206 |
+
"step": 171
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.9927849927849928,
|
| 1210 |
+
"grad_norm": 0.5628295830650956,
|
| 1211 |
+
"learning_rate": 6.76591723478383e-05,
|
| 1212 |
+
"loss": 0.4407,
|
| 1213 |
+
"step": 172
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.9985569985569985,
|
| 1217 |
+
"grad_norm": 0.6701556147889078,
|
| 1218 |
+
"learning_rate": 6.746416039582474e-05,
|
| 1219 |
+
"loss": 0.4422,
|
| 1220 |
+
"step": 173
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 1.0043290043290043,
|
| 1224 |
+
"grad_norm": 1.1221935027853214,
|
| 1225 |
+
"learning_rate": 6.726790555959383e-05,
|
| 1226 |
+
"loss": 0.7579,
|
| 1227 |
+
"step": 174
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 1.0101010101010102,
|
| 1231 |
+
"grad_norm": 1.5081319057253486,
|
| 1232 |
+
"learning_rate": 6.707041672061271e-05,
|
| 1233 |
+
"loss": 0.4294,
|
| 1234 |
+
"step": 175
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 1.0158730158730158,
|
| 1238 |
+
"grad_norm": 0.5264438763222767,
|
| 1239 |
+
"learning_rate": 6.687170281619302e-05,
|
| 1240 |
+
"loss": 0.4276,
|
| 1241 |
+
"step": 176
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 1.0216450216450217,
|
| 1245 |
+
"grad_norm": 1.2609191944611602,
|
| 1246 |
+
"learning_rate": 6.667177283908654e-05,
|
| 1247 |
+
"loss": 0.4262,
|
| 1248 |
+
"step": 177
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 1.0274170274170273,
|
| 1252 |
+
"grad_norm": 0.8861386082370325,
|
| 1253 |
+
"learning_rate": 6.647063583707802e-05,
|
| 1254 |
+
"loss": 0.4242,
|
| 1255 |
+
"step": 178
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 1.0331890331890332,
|
| 1259 |
+
"grad_norm": 1.1725958779338623,
|
| 1260 |
+
"learning_rate": 6.62683009125759e-05,
|
| 1261 |
+
"loss": 0.4237,
|
| 1262 |
+
"step": 179
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 1.0389610389610389,
|
| 1266 |
+
"grad_norm": 0.9929961855942879,
|
| 1267 |
+
"learning_rate": 6.606477722220036e-05,
|
| 1268 |
+
"loss": 0.4195,
|
| 1269 |
+
"step": 180
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 1.0447330447330447,
|
| 1273 |
+
"grad_norm": 0.9200828863123336,
|
| 1274 |
+
"learning_rate": 6.586007397636885e-05,
|
| 1275 |
+
"loss": 0.418,
|
| 1276 |
+
"step": 181
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 1.0505050505050506,
|
| 1280 |
+
"grad_norm": 0.9908709522642545,
|
| 1281 |
+
"learning_rate": 6.565420043887931e-05,
|
| 1282 |
+
"loss": 0.4257,
|
| 1283 |
+
"step": 182
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 1.0562770562770563,
|
| 1287 |
+
"grad_norm": 0.9523415844130142,
|
| 1288 |
+
"learning_rate": 6.544716592649105e-05,
|
| 1289 |
+
"loss": 0.4236,
|
| 1290 |
+
"step": 183
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 1.0620490620490621,
|
| 1294 |
+
"grad_norm": 0.7301914835828962,
|
| 1295 |
+
"learning_rate": 6.523897980850294e-05,
|
| 1296 |
+
"loss": 0.4232,
|
| 1297 |
+
"step": 184
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 1.0678210678210678,
|
| 1301 |
+
"grad_norm": 0.6165866879818018,
|
| 1302 |
+
"learning_rate": 6.502965150632955e-05,
|
| 1303 |
+
"loss": 0.4259,
|
| 1304 |
+
"step": 185
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 1.0735930735930737,
|
| 1308 |
+
"grad_norm": 0.8435043150433259,
|
| 1309 |
+
"learning_rate": 6.48191904930747e-05,
|
| 1310 |
+
"loss": 0.4258,
|
| 1311 |
+
"step": 186
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 1.0793650793650793,
|
| 1315 |
+
"grad_norm": 1.0653357822717338,
|
| 1316 |
+
"learning_rate": 6.460760629310277e-05,
|
| 1317 |
+
"loss": 0.421,
|
| 1318 |
+
"step": 187
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 1.0851370851370852,
|
| 1322 |
+
"grad_norm": 0.896889134293885,
|
| 1323 |
+
"learning_rate": 6.439490848160778e-05,
|
| 1324 |
+
"loss": 0.4185,
|
| 1325 |
+
"step": 188
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 1.0909090909090908,
|
| 1329 |
+
"grad_norm": 0.445405759766472,
|
| 1330 |
+
"learning_rate": 6.418110668417987e-05,
|
| 1331 |
+
"loss": 0.4285,
|
| 1332 |
+
"step": 189
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 1.0966810966810967,
|
| 1336 |
+
"grad_norm": 1.091734393127673,
|
| 1337 |
+
"learning_rate": 6.396621057636989e-05,
|
| 1338 |
+
"loss": 0.4179,
|
| 1339 |
+
"step": 190
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 1.1024531024531024,
|
| 1343 |
+
"grad_norm": 0.8153012892445496,
|
| 1344 |
+
"learning_rate": 6.375022988325141e-05,
|
| 1345 |
+
"loss": 0.4189,
|
| 1346 |
+
"step": 191
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 1.1082251082251082,
|
| 1350 |
+
"grad_norm": 0.47339196574510367,
|
| 1351 |
+
"learning_rate": 6.353317437898067e-05,
|
| 1352 |
+
"loss": 0.4219,
|
| 1353 |
+
"step": 192
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 1.113997113997114,
|
| 1357 |
+
"grad_norm": 0.7949439049041327,
|
| 1358 |
+
"learning_rate": 6.331505388635425e-05,
|
| 1359 |
+
"loss": 0.4209,
|
| 1360 |
+
"step": 193
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 1.1197691197691197,
|
| 1364 |
+
"grad_norm": 0.6343684319179256,
|
| 1365 |
+
"learning_rate": 6.309587827636452e-05,
|
| 1366 |
+
"loss": 0.416,
|
| 1367 |
+
"step": 194
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 1.1255411255411256,
|
| 1371 |
+
"grad_norm": 0.4991129395667874,
|
| 1372 |
+
"learning_rate": 6.287565746775295e-05,
|
| 1373 |
+
"loss": 0.4173,
|
| 1374 |
+
"step": 195
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 1.1313131313131313,
|
| 1378 |
+
"grad_norm": 0.8121276548673446,
|
| 1379 |
+
"learning_rate": 6.265440142656122e-05,
|
| 1380 |
+
"loss": 0.4202,
|
| 1381 |
+
"step": 196
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 1.1370851370851371,
|
| 1385 |
+
"grad_norm": 0.6074410669253114,
|
| 1386 |
+
"learning_rate": 6.243212016568022e-05,
|
| 1387 |
+
"loss": 0.4131,
|
| 1388 |
+
"step": 197
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 1.1428571428571428,
|
| 1392 |
+
"grad_norm": 0.44314411552468735,
|
| 1393 |
+
"learning_rate": 6.220882374439692e-05,
|
| 1394 |
+
"loss": 0.4163,
|
| 1395 |
+
"step": 198
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 1.1486291486291487,
|
| 1399 |
+
"grad_norm": 0.729137709710534,
|
| 1400 |
+
"learning_rate": 6.198452226793918e-05,
|
| 1401 |
+
"loss": 0.4107,
|
| 1402 |
+
"step": 199
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 1.1544011544011543,
|
| 1406 |
+
"grad_norm": 0.5568871901766357,
|
| 1407 |
+
"learning_rate": 6.175922588701831e-05,
|
| 1408 |
+
"loss": 0.4154,
|
| 1409 |
+
"step": 200
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 1.1601731601731602,
|
| 1413 |
+
"grad_norm": 0.4394583517605951,
|
| 1414 |
+
"learning_rate": 6.153294479736988e-05,
|
| 1415 |
+
"loss": 0.4173,
|
| 1416 |
+
"step": 201
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 1.1659451659451658,
|
| 1420 |
+
"grad_norm": 0.5781221410704105,
|
| 1421 |
+
"learning_rate": 6.130568923929218e-05,
|
| 1422 |
+
"loss": 0.4185,
|
| 1423 |
+
"step": 202
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 1.1717171717171717,
|
| 1427 |
+
"grad_norm": 0.4445795099202651,
|
| 1428 |
+
"learning_rate": 6.107746949718286e-05,
|
| 1429 |
+
"loss": 0.4159,
|
| 1430 |
+
"step": 203
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 1.1774891774891776,
|
| 1434 |
+
"grad_norm": 0.29486481158921624,
|
| 1435 |
+
"learning_rate": 6.084829589907348e-05,
|
| 1436 |
+
"loss": 0.4108,
|
| 1437 |
+
"step": 204
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 1.1832611832611832,
|
| 1441 |
+
"grad_norm": 0.4361187764956284,
|
| 1442 |
+
"learning_rate": 6.061817881616215e-05,
|
| 1443 |
+
"loss": 0.4091,
|
| 1444 |
+
"step": 205
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 1.189033189033189,
|
| 1448 |
+
"grad_norm": 0.3800065537217578,
|
| 1449 |
+
"learning_rate": 6.038712866234415e-05,
|
| 1450 |
+
"loss": 0.4122,
|
| 1451 |
+
"step": 206
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 1.1948051948051948,
|
| 1455 |
+
"grad_norm": 0.3347247020532211,
|
| 1456 |
+
"learning_rate": 6.015515589374066e-05,
|
| 1457 |
+
"loss": 0.4157,
|
| 1458 |
+
"step": 207
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 1.2005772005772006,
|
| 1462 |
+
"grad_norm": 0.5905935850780871,
|
| 1463 |
+
"learning_rate": 5.992227100822562e-05,
|
| 1464 |
+
"loss": 0.4151,
|
| 1465 |
+
"step": 208
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 1.2063492063492063,
|
| 1469 |
+
"grad_norm": 0.6199555792896543,
|
| 1470 |
+
"learning_rate": 5.968848454495058e-05,
|
| 1471 |
+
"loss": 0.4114,
|
| 1472 |
+
"step": 209
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 1.2121212121212122,
|
| 1476 |
+
"grad_norm": 0.4743837488073826,
|
| 1477 |
+
"learning_rate": 5.945380708386776e-05,
|
| 1478 |
+
"loss": 0.4111,
|
| 1479 |
+
"step": 210
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 1.2178932178932178,
|
| 1483 |
+
"grad_norm": 0.5588814047789137,
|
| 1484 |
+
"learning_rate": 5.921824924525132e-05,
|
| 1485 |
+
"loss": 0.4199,
|
| 1486 |
+
"step": 211
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 1.2236652236652237,
|
| 1490 |
+
"grad_norm": 0.7590505057556534,
|
| 1491 |
+
"learning_rate": 5.898182168921669e-05,
|
| 1492 |
+
"loss": 0.413,
|
| 1493 |
+
"step": 212
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 1.2294372294372296,
|
| 1497 |
+
"grad_norm": 0.9817733675700191,
|
| 1498 |
+
"learning_rate": 5.874453511523815e-05,
|
| 1499 |
+
"loss": 0.4055,
|
| 1500 |
+
"step": 213
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 1.2352092352092352,
|
| 1504 |
+
"grad_norm": 1.1447333189993498,
|
| 1505 |
+
"learning_rate": 5.8506400261664645e-05,
|
| 1506 |
+
"loss": 0.4181,
|
| 1507 |
+
"step": 214
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 1.240981240981241,
|
| 1511 |
+
"grad_norm": 0.643085872578358,
|
| 1512 |
+
"learning_rate": 5.826742790523382e-05,
|
| 1513 |
+
"loss": 0.4178,
|
| 1514 |
+
"step": 215
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 1.2467532467532467,
|
| 1518 |
+
"grad_norm": 0.6054604348523384,
|
| 1519 |
+
"learning_rate": 5.80276288605843e-05,
|
| 1520 |
+
"loss": 0.4069,
|
| 1521 |
+
"step": 216
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 1.2525252525252526,
|
| 1525 |
+
"grad_norm": 1.182426250184434,
|
| 1526 |
+
"learning_rate": 5.7787013979766334e-05,
|
| 1527 |
+
"loss": 0.4185,
|
| 1528 |
+
"step": 217
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 1.2582972582972582,
|
| 1532 |
+
"grad_norm": 0.8047691901798822,
|
| 1533 |
+
"learning_rate": 5.75455941517506e-05,
|
| 1534 |
+
"loss": 0.4118,
|
| 1535 |
+
"step": 218
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 1.2640692640692641,
|
| 1539 |
+
"grad_norm": 0.2930520495485002,
|
| 1540 |
+
"learning_rate": 5.7303380301935495e-05,
|
| 1541 |
+
"loss": 0.4095,
|
| 1542 |
+
"step": 219
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 1.2698412698412698,
|
| 1546 |
+
"grad_norm": 0.5888144741471648,
|
| 1547 |
+
"learning_rate": 5.70603833916527e-05,
|
| 1548 |
+
"loss": 0.4093,
|
| 1549 |
+
"step": 220
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 1.2756132756132756,
|
| 1553 |
+
"grad_norm": 0.6886762220666399,
|
| 1554 |
+
"learning_rate": 5.68166144176711e-05,
|
| 1555 |
+
"loss": 0.4044,
|
| 1556 |
+
"step": 221
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 1.2813852813852815,
|
| 1560 |
+
"grad_norm": 0.6308802163826126,
|
| 1561 |
+
"learning_rate": 5.657208441169916e-05,
|
| 1562 |
+
"loss": 0.4125,
|
| 1563 |
+
"step": 222
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 1.2871572871572872,
|
| 1567 |
+
"grad_norm": 0.42058805295199864,
|
| 1568 |
+
"learning_rate": 5.6326804439885645e-05,
|
| 1569 |
+
"loss": 0.4101,
|
| 1570 |
+
"step": 223
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 1.2929292929292928,
|
| 1574 |
+
"grad_norm": 0.3592378091851368,
|
| 1575 |
+
"learning_rate": 5.608078560231888e-05,
|
| 1576 |
+
"loss": 0.4041,
|
| 1577 |
+
"step": 224
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 1.2987012987012987,
|
| 1581 |
+
"grad_norm": 0.43029512729600117,
|
| 1582 |
+
"learning_rate": 5.583403903252437e-05,
|
| 1583 |
+
"loss": 0.4121,
|
| 1584 |
+
"step": 225
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 1.3044733044733046,
|
| 1588 |
+
"grad_norm": 0.3889055430906675,
|
| 1589 |
+
"learning_rate": 5.558657589696098e-05,
|
| 1590 |
+
"loss": 0.4059,
|
| 1591 |
+
"step": 226
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 1.3102453102453102,
|
| 1595 |
+
"grad_norm": 0.3440494840478949,
|
| 1596 |
+
"learning_rate": 5.533840739451559e-05,
|
| 1597 |
+
"loss": 0.4034,
|
| 1598 |
+
"step": 227
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 1.316017316017316,
|
| 1602 |
+
"grad_norm": 0.2571934157256294,
|
| 1603 |
+
"learning_rate": 5.5089544755996294e-05,
|
| 1604 |
+
"loss": 0.41,
|
| 1605 |
+
"step": 228
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 1.3217893217893217,
|
| 1609 |
+
"grad_norm": 0.3824710170652386,
|
| 1610 |
+
"learning_rate": 5.483999924362417e-05,
|
| 1611 |
+
"loss": 0.406,
|
| 1612 |
+
"step": 229
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 1.3275613275613276,
|
| 1616 |
+
"grad_norm": 0.41843958170342843,
|
| 1617 |
+
"learning_rate": 5.458978215052357e-05,
|
| 1618 |
+
"loss": 0.4046,
|
| 1619 |
+
"step": 230
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 1.3333333333333333,
|
| 1623 |
+
"grad_norm": 0.3498561520298379,
|
| 1624 |
+
"learning_rate": 5.433890480021107e-05,
|
| 1625 |
+
"loss": 0.4036,
|
| 1626 |
+
"step": 231
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 1.3391053391053391,
|
| 1630 |
+
"grad_norm": 0.42164963063157984,
|
| 1631 |
+
"learning_rate": 5.408737854608307e-05,
|
| 1632 |
+
"loss": 0.4027,
|
| 1633 |
+
"step": 232
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 1.3448773448773448,
|
| 1637 |
+
"grad_norm": 0.48195135845030257,
|
| 1638 |
+
"learning_rate": 5.3835214770901945e-05,
|
| 1639 |
+
"loss": 0.4018,
|
| 1640 |
+
"step": 233
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 1.3506493506493507,
|
| 1644 |
+
"grad_norm": 0.4852917535336072,
|
| 1645 |
+
"learning_rate": 5.358242488628092e-05,
|
| 1646 |
+
"loss": 0.4083,
|
| 1647 |
+
"step": 234
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 1.3564213564213565,
|
| 1651 |
+
"grad_norm": 0.41225922180900215,
|
| 1652 |
+
"learning_rate": 5.332902033216766e-05,
|
| 1653 |
+
"loss": 0.4058,
|
| 1654 |
+
"step": 235
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 1.3621933621933622,
|
| 1658 |
+
"grad_norm": 0.3010982188590577,
|
| 1659 |
+
"learning_rate": 5.3075012576326607e-05,
|
| 1660 |
+
"loss": 0.4051,
|
| 1661 |
+
"step": 236
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 1.3679653679653678,
|
| 1665 |
+
"grad_norm": 0.2433157254360926,
|
| 1666 |
+
"learning_rate": 5.2820413113819885e-05,
|
| 1667 |
+
"loss": 0.3998,
|
| 1668 |
+
"step": 237
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 1.3737373737373737,
|
| 1672 |
+
"grad_norm": 0.2833835472451347,
|
| 1673 |
+
"learning_rate": 5.256523346648721e-05,
|
| 1674 |
+
"loss": 0.404,
|
| 1675 |
+
"step": 238
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 1.3795093795093796,
|
| 1679 |
+
"grad_norm": 0.27072179535322477,
|
| 1680 |
+
"learning_rate": 5.230948518242439e-05,
|
| 1681 |
+
"loss": 0.4019,
|
| 1682 |
+
"step": 239
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 1.3852813852813852,
|
| 1686 |
+
"grad_norm": 0.24437603340900985,
|
| 1687 |
+
"learning_rate": 5.205317983546083e-05,
|
| 1688 |
+
"loss": 0.4076,
|
| 1689 |
+
"step": 240
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 1.391053391053391,
|
| 1693 |
+
"grad_norm": 0.22711332037700857,
|
| 1694 |
+
"learning_rate": 5.179632902463562e-05,
|
| 1695 |
+
"loss": 0.4,
|
| 1696 |
+
"step": 241
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 1.3968253968253967,
|
| 1700 |
+
"grad_norm": 0.29929194968825346,
|
| 1701 |
+
"learning_rate": 5.1538944373672694e-05,
|
| 1702 |
+
"loss": 0.4179,
|
| 1703 |
+
"step": 242
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 1.4025974025974026,
|
| 1707 |
+
"grad_norm": 0.284448857778488,
|
| 1708 |
+
"learning_rate": 5.128103753045484e-05,
|
| 1709 |
+
"loss": 0.4097,
|
| 1710 |
+
"step": 243
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 1.4083694083694085,
|
| 1714 |
+
"grad_norm": 0.32649263835005693,
|
| 1715 |
+
"learning_rate": 5.1022620166496486e-05,
|
| 1716 |
+
"loss": 0.4046,
|
| 1717 |
+
"step": 244
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 1.4141414141414141,
|
| 1721 |
+
"grad_norm": 0.2601804107726681,
|
| 1722 |
+
"learning_rate": 5.07637039764156e-05,
|
| 1723 |
+
"loss": 0.4007,
|
| 1724 |
+
"step": 245
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 1.4199134199134198,
|
| 1728 |
+
"grad_norm": 0.2985437577875243,
|
| 1729 |
+
"learning_rate": 5.05043006774044e-05,
|
| 1730 |
+
"loss": 0.4113,
|
| 1731 |
+
"step": 246
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 1.4256854256854257,
|
| 1735 |
+
"grad_norm": 0.3479198216565732,
|
| 1736 |
+
"learning_rate": 5.024442200869907e-05,
|
| 1737 |
+
"loss": 0.4063,
|
| 1738 |
+
"step": 247
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 1.4314574314574315,
|
| 1742 |
+
"grad_norm": 0.2934502700017442,
|
| 1743 |
+
"learning_rate": 4.9984079731048594e-05,
|
| 1744 |
+
"loss": 0.4044,
|
| 1745 |
+
"step": 248
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 1.4372294372294372,
|
| 1749 |
+
"grad_norm": 0.284443761108607,
|
| 1750 |
+
"learning_rate": 4.972328562618244e-05,
|
| 1751 |
+
"loss": 0.4021,
|
| 1752 |
+
"step": 249
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 1.443001443001443,
|
| 1756 |
+
"grad_norm": 0.2686380250003549,
|
| 1757 |
+
"learning_rate": 4.946205149627742e-05,
|
| 1758 |
+
"loss": 0.3967,
|
| 1759 |
+
"step": 250
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 1.4487734487734487,
|
| 1763 |
+
"grad_norm": 0.20965782391240942,
|
| 1764 |
+
"learning_rate": 4.920038916342355e-05,
|
| 1765 |
+
"loss": 0.4055,
|
| 1766 |
+
"step": 251
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 1.4545454545454546,
|
| 1770 |
+
"grad_norm": 0.26593807250010953,
|
| 1771 |
+
"learning_rate": 4.893831046908913e-05,
|
| 1772 |
+
"loss": 0.3919,
|
| 1773 |
+
"step": 252
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 1.4603174603174602,
|
| 1777 |
+
"grad_norm": 0.3244764958086802,
|
| 1778 |
+
"learning_rate": 4.8675827273584764e-05,
|
| 1779 |
+
"loss": 0.4066,
|
| 1780 |
+
"step": 253
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 1.466089466089466,
|
| 1784 |
+
"grad_norm": 0.4519123025012951,
|
| 1785 |
+
"learning_rate": 4.841295145552664e-05,
|
| 1786 |
+
"loss": 0.409,
|
| 1787 |
+
"step": 254
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 1.4718614718614718,
|
| 1791 |
+
"grad_norm": 0.47539799585538184,
|
| 1792 |
+
"learning_rate": 4.8149694911299046e-05,
|
| 1793 |
+
"loss": 0.4068,
|
| 1794 |
+
"step": 255
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 1.4776334776334776,
|
| 1798 |
+
"grad_norm": 0.3163497152972903,
|
| 1799 |
+
"learning_rate": 4.788606955451589e-05,
|
| 1800 |
+
"loss": 0.4008,
|
| 1801 |
+
"step": 256
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 1.4834054834054835,
|
| 1805 |
+
"grad_norm": 0.2530368981149338,
|
| 1806 |
+
"learning_rate": 4.762208731548166e-05,
|
| 1807 |
+
"loss": 0.4045,
|
| 1808 |
+
"step": 257
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 1.4891774891774892,
|
| 1812 |
+
"grad_norm": 0.23038203037988386,
|
| 1813 |
+
"learning_rate": 4.735776014065142e-05,
|
| 1814 |
+
"loss": 0.3988,
|
| 1815 |
+
"step": 258
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 1.494949494949495,
|
| 1819 |
+
"grad_norm": 0.21695905607050836,
|
| 1820 |
+
"learning_rate": 4.709309999209023e-05,
|
| 1821 |
+
"loss": 0.3978,
|
| 1822 |
+
"step": 259
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 1.5007215007215007,
|
| 1826 |
+
"grad_norm": 0.2567325295594068,
|
| 1827 |
+
"learning_rate": 4.682811884693185e-05,
|
| 1828 |
+
"loss": 0.3936,
|
| 1829 |
+
"step": 260
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 1.5064935064935066,
|
| 1833 |
+
"grad_norm": 0.2692904261100886,
|
| 1834 |
+
"learning_rate": 4.656282869683659e-05,
|
| 1835 |
+
"loss": 0.3969,
|
| 1836 |
+
"step": 261
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 1.5122655122655124,
|
| 1840 |
+
"grad_norm": 0.2636411848036993,
|
| 1841 |
+
"learning_rate": 4.629724154744876e-05,
|
| 1842 |
+
"loss": 0.3992,
|
| 1843 |
+
"step": 262
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 1.518037518037518,
|
| 1847 |
+
"grad_norm": 0.20496250107354846,
|
| 1848 |
+
"learning_rate": 4.603136941785328e-05,
|
| 1849 |
+
"loss": 0.3947,
|
| 1850 |
+
"step": 263
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 1.5238095238095237,
|
| 1854 |
+
"grad_norm": 0.24102557044701634,
|
| 1855 |
+
"learning_rate": 4.576522434003183e-05,
|
| 1856 |
+
"loss": 0.3984,
|
| 1857 |
+
"step": 264
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 1.5295815295815296,
|
| 1861 |
+
"grad_norm": 0.270197841622627,
|
| 1862 |
+
"learning_rate": 4.549881835831821e-05,
|
| 1863 |
+
"loss": 0.3915,
|
| 1864 |
+
"step": 265
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 1.5353535353535355,
|
| 1868 |
+
"grad_norm": 0.31636039416656614,
|
| 1869 |
+
"learning_rate": 4.523216352885345e-05,
|
| 1870 |
+
"loss": 0.3952,
|
| 1871 |
+
"step": 266
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 1.5411255411255411,
|
| 1875 |
+
"grad_norm": 0.3077596806884647,
|
| 1876 |
+
"learning_rate": 4.496527191904009e-05,
|
| 1877 |
+
"loss": 0.3976,
|
| 1878 |
+
"step": 267
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 1.5468975468975468,
|
| 1882 |
+
"grad_norm": 0.2675132016090662,
|
| 1883 |
+
"learning_rate": 4.4698155606996104e-05,
|
| 1884 |
+
"loss": 0.4057,
|
| 1885 |
+
"step": 268
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 1.5526695526695526,
|
| 1889 |
+
"grad_norm": 0.21038804294274088,
|
| 1890 |
+
"learning_rate": 4.4430826681008316e-05,
|
| 1891 |
+
"loss": 0.4074,
|
| 1892 |
+
"step": 269
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 1.5584415584415585,
|
| 1896 |
+
"grad_norm": 0.23735821005743604,
|
| 1897 |
+
"learning_rate": 4.416329723898536e-05,
|
| 1898 |
+
"loss": 0.3951,
|
| 1899 |
+
"step": 270
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 1.5642135642135642,
|
| 1903 |
+
"grad_norm": 0.18061314585892582,
|
| 1904 |
+
"learning_rate": 4.389557938791019e-05,
|
| 1905 |
+
"loss": 0.394,
|
| 1906 |
+
"step": 271
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 1.56998556998557,
|
| 1910 |
+
"grad_norm": 0.23191872327293342,
|
| 1911 |
+
"learning_rate": 4.362768524329212e-05,
|
| 1912 |
+
"loss": 0.3993,
|
| 1913 |
+
"step": 272
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 1.5757575757575757,
|
| 1917 |
+
"grad_norm": 0.20523251504710777,
|
| 1918 |
+
"learning_rate": 4.335962692861862e-05,
|
| 1919 |
+
"loss": 0.4013,
|
| 1920 |
+
"step": 273
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 1.5815295815295816,
|
| 1924 |
+
"grad_norm": 0.21213920806551073,
|
| 1925 |
+
"learning_rate": 4.3091416574806644e-05,
|
| 1926 |
+
"loss": 0.3932,
|
| 1927 |
+
"step": 274
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 1.5873015873015874,
|
| 1931 |
+
"grad_norm": 0.20822870111060587,
|
| 1932 |
+
"learning_rate": 4.2823066319653606e-05,
|
| 1933 |
+
"loss": 0.4066,
|
| 1934 |
+
"step": 275
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 1.593073593073593,
|
| 1938 |
+
"grad_norm": 0.2677158742340245,
|
| 1939 |
+
"learning_rate": 4.2554588307288166e-05,
|
| 1940 |
+
"loss": 0.3953,
|
| 1941 |
+
"step": 276
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 1.5988455988455987,
|
| 1945 |
+
"grad_norm": 0.24447097986602137,
|
| 1946 |
+
"learning_rate": 4.2285994687620586e-05,
|
| 1947 |
+
"loss": 0.3908,
|
| 1948 |
+
"step": 277
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 1.6046176046176046,
|
| 1952 |
+
"grad_norm": 0.211196522709894,
|
| 1953 |
+
"learning_rate": 4.20172976157929e-05,
|
| 1954 |
+
"loss": 0.4048,
|
| 1955 |
+
"step": 278
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 1.6103896103896105,
|
| 1959 |
+
"grad_norm": 0.2035651604206692,
|
| 1960 |
+
"learning_rate": 4.174850925162887e-05,
|
| 1961 |
+
"loss": 0.3994,
|
| 1962 |
+
"step": 279
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 1.6161616161616161,
|
| 1966 |
+
"grad_norm": 0.28087618115996654,
|
| 1967 |
+
"learning_rate": 4.1479641759083653e-05,
|
| 1968 |
+
"loss": 0.3954,
|
| 1969 |
+
"step": 280
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 1.6219336219336218,
|
| 1973 |
+
"grad_norm": 0.2869550764537825,
|
| 1974 |
+
"learning_rate": 4.121070730569336e-05,
|
| 1975 |
+
"loss": 0.3934,
|
| 1976 |
+
"step": 281
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 1.6277056277056277,
|
| 1980 |
+
"grad_norm": 0.2627243894997915,
|
| 1981 |
+
"learning_rate": 4.094171806202436e-05,
|
| 1982 |
+
"loss": 0.3929,
|
| 1983 |
+
"step": 282
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 1.6334776334776335,
|
| 1987 |
+
"grad_norm": 0.18504635663976976,
|
| 1988 |
+
"learning_rate": 4.067268620112259e-05,
|
| 1989 |
+
"loss": 0.3921,
|
| 1990 |
+
"step": 283
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 1.6392496392496394,
|
| 1994 |
+
"grad_norm": 0.24273499891124686,
|
| 1995 |
+
"learning_rate": 4.04036238979626e-05,
|
| 1996 |
+
"loss": 0.3863,
|
| 1997 |
+
"step": 284
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 1.645021645021645,
|
| 2001 |
+
"grad_norm": 0.21825468272936616,
|
| 2002 |
+
"learning_rate": 4.01345433288966e-05,
|
| 2003 |
+
"loss": 0.3903,
|
| 2004 |
+
"step": 285
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 1.6507936507936507,
|
| 2008 |
+
"grad_norm": 0.21301467745169061,
|
| 2009 |
+
"learning_rate": 3.9865456671103404e-05,
|
| 2010 |
+
"loss": 0.3941,
|
| 2011 |
+
"step": 286
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 1.6565656565656566,
|
| 2015 |
+
"grad_norm": 0.24240452578983904,
|
| 2016 |
+
"learning_rate": 3.959637610203741e-05,
|
| 2017 |
+
"loss": 0.3914,
|
| 2018 |
+
"step": 287
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 1.6623376623376624,
|
| 2022 |
+
"grad_norm": 0.24775047125742394,
|
| 2023 |
+
"learning_rate": 3.932731379887742e-05,
|
| 2024 |
+
"loss": 0.3865,
|
| 2025 |
+
"step": 288
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 1.668109668109668,
|
| 2029 |
+
"grad_norm": 0.20875238150820688,
|
| 2030 |
+
"learning_rate": 3.905828193797566e-05,
|
| 2031 |
+
"loss": 0.3946,
|
| 2032 |
+
"step": 289
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 1.6738816738816737,
|
| 2036 |
+
"grad_norm": 0.2092017685246258,
|
| 2037 |
+
"learning_rate": 3.878929269430667e-05,
|
| 2038 |
+
"loss": 0.3905,
|
| 2039 |
+
"step": 290
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 1.6796536796536796,
|
| 2043 |
+
"grad_norm": 0.21363231113184575,
|
| 2044 |
+
"learning_rate": 3.852035824091635e-05,
|
| 2045 |
+
"loss": 0.3893,
|
| 2046 |
+
"step": 291
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 1.6854256854256855,
|
| 2050 |
+
"grad_norm": 0.21569898511300334,
|
| 2051 |
+
"learning_rate": 3.8251490748371136e-05,
|
| 2052 |
+
"loss": 0.3942,
|
| 2053 |
+
"step": 292
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 1.6911976911976911,
|
| 2057 |
+
"grad_norm": 0.2018232998507949,
|
| 2058 |
+
"learning_rate": 3.798270238420711e-05,
|
| 2059 |
+
"loss": 0.3901,
|
| 2060 |
+
"step": 293
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 1.696969696969697,
|
| 2064 |
+
"grad_norm": 0.1925594320139884,
|
| 2065 |
+
"learning_rate": 3.771400531237942e-05,
|
| 2066 |
+
"loss": 0.3906,
|
| 2067 |
+
"step": 294
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 1.7027417027417027,
|
| 2071 |
+
"grad_norm": 0.18974891989147144,
|
| 2072 |
+
"learning_rate": 3.744541169271184e-05,
|
| 2073 |
+
"loss": 0.3954,
|
| 2074 |
+
"step": 295
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 1.7085137085137085,
|
| 2078 |
+
"grad_norm": 0.20798653934785333,
|
| 2079 |
+
"learning_rate": 3.7176933680346414e-05,
|
| 2080 |
+
"loss": 0.3943,
|
| 2081 |
+
"step": 296
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 1.7142857142857144,
|
| 2085 |
+
"grad_norm": 0.22892716154764753,
|
| 2086 |
+
"learning_rate": 3.6908583425193376e-05,
|
| 2087 |
+
"loss": 0.3927,
|
| 2088 |
+
"step": 297
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 1.72005772005772,
|
| 2092 |
+
"grad_norm": 0.20442563522892104,
|
| 2093 |
+
"learning_rate": 3.66403730713814e-05,
|
| 2094 |
+
"loss": 0.3899,
|
| 2095 |
+
"step": 298
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 1.7258297258297257,
|
| 2099 |
+
"grad_norm": 0.2290558666884635,
|
| 2100 |
+
"learning_rate": 3.637231475670789e-05,
|
| 2101 |
+
"loss": 0.3893,
|
| 2102 |
+
"step": 299
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 1.7316017316017316,
|
| 2106 |
+
"grad_norm": 0.24205910207531414,
|
| 2107 |
+
"learning_rate": 3.6104420612089814e-05,
|
| 2108 |
+
"loss": 0.3785,
|
| 2109 |
+
"step": 300
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 1.7373737373737375,
|
| 2113 |
+
"grad_norm": 0.25593097184058905,
|
| 2114 |
+
"learning_rate": 3.583670276101464e-05,
|
| 2115 |
+
"loss": 0.3933,
|
| 2116 |
+
"step": 301
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 1.743145743145743,
|
| 2120 |
+
"grad_norm": 0.17738260043921314,
|
| 2121 |
+
"learning_rate": 3.556917331899169e-05,
|
| 2122 |
+
"loss": 0.3972,
|
| 2123 |
+
"step": 302
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 1.7489177489177488,
|
| 2127 |
+
"grad_norm": 0.22220255965346145,
|
| 2128 |
+
"learning_rate": 3.530184439300392e-05,
|
| 2129 |
+
"loss": 0.3932,
|
| 2130 |
+
"step": 303
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 1.7546897546897546,
|
| 2134 |
+
"grad_norm": 0.2123289870960355,
|
| 2135 |
+
"learning_rate": 3.503472808095993e-05,
|
| 2136 |
+
"loss": 0.388,
|
| 2137 |
+
"step": 304
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 1.7604617604617605,
|
| 2141 |
+
"grad_norm": 0.21684056813791056,
|
| 2142 |
+
"learning_rate": 3.476783647114656e-05,
|
| 2143 |
+
"loss": 0.3871,
|
| 2144 |
+
"step": 305
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 1.7662337662337664,
|
| 2148 |
+
"grad_norm": 0.2986703879447781,
|
| 2149 |
+
"learning_rate": 3.4501181641681806e-05,
|
| 2150 |
+
"loss": 0.4022,
|
| 2151 |
+
"step": 306
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 1.772005772005772,
|
| 2155 |
+
"grad_norm": 0.22910905968331996,
|
| 2156 |
+
"learning_rate": 3.4234775659968185e-05,
|
| 2157 |
+
"loss": 0.387,
|
| 2158 |
+
"step": 307
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 1.7777777777777777,
|
| 2162 |
+
"grad_norm": 0.30310896794339226,
|
| 2163 |
+
"learning_rate": 3.396863058214671e-05,
|
| 2164 |
+
"loss": 0.3954,
|
| 2165 |
+
"step": 308
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 1.7835497835497836,
|
| 2169 |
+
"grad_norm": 0.28241561357778666,
|
| 2170 |
+
"learning_rate": 3.370275845255125e-05,
|
| 2171 |
+
"loss": 0.3876,
|
| 2172 |
+
"step": 309
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 1.7893217893217894,
|
| 2176 |
+
"grad_norm": 0.2509414521780602,
|
| 2177 |
+
"learning_rate": 3.343717130316344e-05,
|
| 2178 |
+
"loss": 0.3844,
|
| 2179 |
+
"step": 310
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 1.795093795093795,
|
| 2183 |
+
"grad_norm": 0.23323065292973938,
|
| 2184 |
+
"learning_rate": 3.317188115306817e-05,
|
| 2185 |
+
"loss": 0.3869,
|
| 2186 |
+
"step": 311
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 1.8008658008658007,
|
| 2190 |
+
"grad_norm": 0.26047360683879345,
|
| 2191 |
+
"learning_rate": 3.290690000790978e-05,
|
| 2192 |
+
"loss": 0.3953,
|
| 2193 |
+
"step": 312
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 1.8066378066378066,
|
| 2197 |
+
"grad_norm": 0.23595979586126703,
|
| 2198 |
+
"learning_rate": 3.2642239859348594e-05,
|
| 2199 |
+
"loss": 0.3893,
|
| 2200 |
+
"step": 313
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 1.8124098124098125,
|
| 2204 |
+
"grad_norm": 0.22572346653298958,
|
| 2205 |
+
"learning_rate": 3.237791268451835e-05,
|
| 2206 |
+
"loss": 0.3847,
|
| 2207 |
+
"step": 314
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 1.8181818181818183,
|
| 2211 |
+
"grad_norm": 0.2098525856576488,
|
| 2212 |
+
"learning_rate": 3.211393044548411e-05,
|
| 2213 |
+
"loss": 0.4011,
|
| 2214 |
+
"step": 315
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 1.823953823953824,
|
| 2218 |
+
"grad_norm": 0.2116216292884672,
|
| 2219 |
+
"learning_rate": 3.185030508870096e-05,
|
| 2220 |
+
"loss": 0.3806,
|
| 2221 |
+
"step": 316
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 1.8297258297258296,
|
| 2225 |
+
"grad_norm": 0.21095945535091404,
|
| 2226 |
+
"learning_rate": 3.158704854447337e-05,
|
| 2227 |
+
"loss": 0.385,
|
| 2228 |
+
"step": 317
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 1.8354978354978355,
|
| 2232 |
+
"grad_norm": 0.20877744581550697,
|
| 2233 |
+
"learning_rate": 3.1324172726415256e-05,
|
| 2234 |
+
"loss": 0.3879,
|
| 2235 |
+
"step": 318
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 1.8412698412698414,
|
| 2239 |
+
"grad_norm": 0.19527613530959578,
|
| 2240 |
+
"learning_rate": 3.106168953091088e-05,
|
| 2241 |
+
"loss": 0.3892,
|
| 2242 |
+
"step": 319
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 1.847041847041847,
|
| 2246 |
+
"grad_norm": 0.20996890512994054,
|
| 2247 |
+
"learning_rate": 3.079961083657646e-05,
|
| 2248 |
+
"loss": 0.3821,
|
| 2249 |
+
"step": 320
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 1.8528138528138527,
|
| 2253 |
+
"grad_norm": 0.19908534919754334,
|
| 2254 |
+
"learning_rate": 3.0537948503722595e-05,
|
| 2255 |
+
"loss": 0.391,
|
| 2256 |
+
"step": 321
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 1.8585858585858586,
|
| 2260 |
+
"grad_norm": 0.2069424165481144,
|
| 2261 |
+
"learning_rate": 3.027671437381757e-05,
|
| 2262 |
+
"loss": 0.3896,
|
| 2263 |
+
"step": 322
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 1.8643578643578644,
|
| 2267 |
+
"grad_norm": 0.18991524333796642,
|
| 2268 |
+
"learning_rate": 3.0015920268951412e-05,
|
| 2269 |
+
"loss": 0.3927,
|
| 2270 |
+
"step": 323
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 1.87012987012987,
|
| 2274 |
+
"grad_norm": 0.202639438723347,
|
| 2275 |
+
"learning_rate": 2.975557799130094e-05,
|
| 2276 |
+
"loss": 0.3837,
|
| 2277 |
+
"step": 324
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 1.8759018759018757,
|
| 2281 |
+
"grad_norm": 0.17613137029703754,
|
| 2282 |
+
"learning_rate": 2.9495699322595615e-05,
|
| 2283 |
+
"loss": 0.3975,
|
| 2284 |
+
"step": 325
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 1.8816738816738816,
|
| 2288 |
+
"grad_norm": 0.21155619568693387,
|
| 2289 |
+
"learning_rate": 2.923629602358441e-05,
|
| 2290 |
+
"loss": 0.387,
|
| 2291 |
+
"step": 326
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 1.8874458874458875,
|
| 2295 |
+
"grad_norm": 0.1725011702032577,
|
| 2296 |
+
"learning_rate": 2.8977379833503524e-05,
|
| 2297 |
+
"loss": 0.391,
|
| 2298 |
+
"step": 327
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 1.8932178932178934,
|
| 2302 |
+
"grad_norm": 0.22393946086021008,
|
| 2303 |
+
"learning_rate": 2.871896246954518e-05,
|
| 2304 |
+
"loss": 0.3924,
|
| 2305 |
+
"step": 328
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 1.898989898989899,
|
| 2309 |
+
"grad_norm": 0.19876543898525306,
|
| 2310 |
+
"learning_rate": 2.8461055626327313e-05,
|
| 2311 |
+
"loss": 0.3861,
|
| 2312 |
+
"step": 329
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 1.9047619047619047,
|
| 2316 |
+
"grad_norm": 0.2245112603739307,
|
| 2317 |
+
"learning_rate": 2.8203670975364395e-05,
|
| 2318 |
+
"loss": 0.3923,
|
| 2319 |
+
"step": 330
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 1.9105339105339105,
|
| 2323 |
+
"grad_norm": 0.16602895720578406,
|
| 2324 |
+
"learning_rate": 2.7946820164539182e-05,
|
| 2325 |
+
"loss": 0.3891,
|
| 2326 |
+
"step": 331
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 1.9163059163059164,
|
| 2330 |
+
"grad_norm": 0.21870332861694683,
|
| 2331 |
+
"learning_rate": 2.7690514817575615e-05,
|
| 2332 |
+
"loss": 0.3924,
|
| 2333 |
+
"step": 332
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 1.922077922077922,
|
| 2337 |
+
"grad_norm": 0.17060774656580643,
|
| 2338 |
+
"learning_rate": 2.7434766533512806e-05,
|
| 2339 |
+
"loss": 0.3737,
|
| 2340 |
+
"step": 333
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 1.9278499278499277,
|
| 2344 |
+
"grad_norm": 0.21300688928118505,
|
| 2345 |
+
"learning_rate": 2.7179586886180128e-05,
|
| 2346 |
+
"loss": 0.3948,
|
| 2347 |
+
"step": 334
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 1.9336219336219336,
|
| 2351 |
+
"grad_norm": 0.175788452853784,
|
| 2352 |
+
"learning_rate": 2.69249874236734e-05,
|
| 2353 |
+
"loss": 0.3866,
|
| 2354 |
+
"step": 335
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 1.9393939393939394,
|
| 2358 |
+
"grad_norm": 0.18460721760511187,
|
| 2359 |
+
"learning_rate": 2.667097966783234e-05,
|
| 2360 |
+
"loss": 0.3862,
|
| 2361 |
+
"step": 336
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 1.9451659451659453,
|
| 2365 |
+
"grad_norm": 0.17444022188650732,
|
| 2366 |
+
"learning_rate": 2.6417575113719087e-05,
|
| 2367 |
+
"loss": 0.3802,
|
| 2368 |
+
"step": 337
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 1.950937950937951,
|
| 2372 |
+
"grad_norm": 0.1484989021731224,
|
| 2373 |
+
"learning_rate": 2.616478522909807e-05,
|
| 2374 |
+
"loss": 0.392,
|
| 2375 |
+
"step": 338
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 1.9567099567099566,
|
| 2379 |
+
"grad_norm": 0.16268258501003877,
|
| 2380 |
+
"learning_rate": 2.5912621453916944e-05,
|
| 2381 |
+
"loss": 0.3885,
|
| 2382 |
+
"step": 339
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 1.9624819624819625,
|
| 2386 |
+
"grad_norm": 0.16171751920146682,
|
| 2387 |
+
"learning_rate": 2.566109519978894e-05,
|
| 2388 |
+
"loss": 0.3808,
|
| 2389 |
+
"step": 340
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 1.9682539682539684,
|
| 2393 |
+
"grad_norm": 0.16382873859878613,
|
| 2394 |
+
"learning_rate": 2.5410217849476447e-05,
|
| 2395 |
+
"loss": 0.3752,
|
| 2396 |
+
"step": 341
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 1.974025974025974,
|
| 2400 |
+
"grad_norm": 0.17482617424613112,
|
| 2401 |
+
"learning_rate": 2.516000075637584e-05,
|
| 2402 |
+
"loss": 0.3763,
|
| 2403 |
+
"step": 342
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 1.9797979797979797,
|
| 2407 |
+
"grad_norm": 0.1702896815624782,
|
| 2408 |
+
"learning_rate": 2.4910455244003702e-05,
|
| 2409 |
+
"loss": 0.3814,
|
| 2410 |
+
"step": 343
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 1.9855699855699855,
|
| 2414 |
+
"grad_norm": 0.16803526970105556,
|
| 2415 |
+
"learning_rate": 2.4661592605484422e-05,
|
| 2416 |
+
"loss": 0.3886,
|
| 2417 |
+
"step": 344
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 1.9913419913419914,
|
| 2421 |
+
"grad_norm": 0.17404030012234012,
|
| 2422 |
+
"learning_rate": 2.4413424103039035e-05,
|
| 2423 |
+
"loss": 0.3856,
|
| 2424 |
+
"step": 345
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 1.997113997113997,
|
| 2428 |
+
"grad_norm": 0.16281429029619196,
|
| 2429 |
+
"learning_rate": 2.416596096747564e-05,
|
| 2430 |
+
"loss": 0.3812,
|
| 2431 |
+
"step": 346
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 2.0028860028860027,
|
| 2435 |
+
"grad_norm": 0.34376210625654163,
|
| 2436 |
+
"learning_rate": 2.3919214397681137e-05,
|
| 2437 |
+
"loss": 0.6364,
|
| 2438 |
+
"step": 347
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 2.0086580086580086,
|
| 2442 |
+
"grad_norm": 0.25772199023989556,
|
| 2443 |
+
"learning_rate": 2.3673195560114365e-05,
|
| 2444 |
+
"loss": 0.3528,
|
| 2445 |
+
"step": 348
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 2.0144300144300145,
|
| 2449 |
+
"grad_norm": 0.276112829230668,
|
| 2450 |
+
"learning_rate": 2.3427915588300846e-05,
|
| 2451 |
+
"loss": 0.3535,
|
| 2452 |
+
"step": 349
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 2.0202020202020203,
|
| 2456 |
+
"grad_norm": 0.31992574049691747,
|
| 2457 |
+
"learning_rate": 2.31833855823289e-05,
|
| 2458 |
+
"loss": 0.3491,
|
| 2459 |
+
"step": 350
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 2.0259740259740258,
|
| 2463 |
+
"grad_norm": 0.2608841192334598,
|
| 2464 |
+
"learning_rate": 2.2939616608347316e-05,
|
| 2465 |
+
"loss": 0.3554,
|
| 2466 |
+
"step": 351
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 2.0317460317460316,
|
| 2470 |
+
"grad_norm": 0.2653214329011091,
|
| 2471 |
+
"learning_rate": 2.2696619698064515e-05,
|
| 2472 |
+
"loss": 0.3493,
|
| 2473 |
+
"step": 352
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 2.0375180375180375,
|
| 2477 |
+
"grad_norm": 0.2472838788658597,
|
| 2478 |
+
"learning_rate": 2.245440584824941e-05,
|
| 2479 |
+
"loss": 0.347,
|
| 2480 |
+
"step": 353
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 2.0432900432900434,
|
| 2484 |
+
"grad_norm": 0.26371573358746464,
|
| 2485 |
+
"learning_rate": 2.2212986020233683e-05,
|
| 2486 |
+
"loss": 0.3493,
|
| 2487 |
+
"step": 354
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 2.0490620490620493,
|
| 2491 |
+
"grad_norm": 0.20878293083270738,
|
| 2492 |
+
"learning_rate": 2.1972371139415705e-05,
|
| 2493 |
+
"loss": 0.3553,
|
| 2494 |
+
"step": 355
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 2.0548340548340547,
|
| 2498 |
+
"grad_norm": 0.23273233249130984,
|
| 2499 |
+
"learning_rate": 2.1732572094766188e-05,
|
| 2500 |
+
"loss": 0.3532,
|
| 2501 |
+
"step": 356
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 2.0606060606060606,
|
| 2505 |
+
"grad_norm": 0.23527543008138327,
|
| 2506 |
+
"learning_rate": 2.1493599738335358e-05,
|
| 2507 |
+
"loss": 0.3553,
|
| 2508 |
+
"step": 357
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 2.0663780663780664,
|
| 2512 |
+
"grad_norm": 0.20970708352117728,
|
| 2513 |
+
"learning_rate": 2.1255464884761863e-05,
|
| 2514 |
+
"loss": 0.3574,
|
| 2515 |
+
"step": 358
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 2.0721500721500723,
|
| 2519 |
+
"grad_norm": 0.2083525361419124,
|
| 2520 |
+
"learning_rate": 2.1018178310783322e-05,
|
| 2521 |
+
"loss": 0.3467,
|
| 2522 |
+
"step": 359
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 2.0779220779220777,
|
| 2526 |
+
"grad_norm": 0.20295702822985864,
|
| 2527 |
+
"learning_rate": 2.07817507547487e-05,
|
| 2528 |
+
"loss": 0.3518,
|
| 2529 |
+
"step": 360
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 2.0836940836940836,
|
| 2533 |
+
"grad_norm": 0.17820081287215864,
|
| 2534 |
+
"learning_rate": 2.054619291613226e-05,
|
| 2535 |
+
"loss": 0.3454,
|
| 2536 |
+
"step": 361
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 2.0894660894660895,
|
| 2540 |
+
"grad_norm": 0.21352202172185789,
|
| 2541 |
+
"learning_rate": 2.031151545504944e-05,
|
| 2542 |
+
"loss": 0.3509,
|
| 2543 |
+
"step": 362
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 2.0952380952380953,
|
| 2547 |
+
"grad_norm": 0.16570061254078286,
|
| 2548 |
+
"learning_rate": 2.007772899177438e-05,
|
| 2549 |
+
"loss": 0.3591,
|
| 2550 |
+
"step": 363
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 2.101010101010101,
|
| 2554 |
+
"grad_norm": 0.17922618028941195,
|
| 2555 |
+
"learning_rate": 1.984484410625935e-05,
|
| 2556 |
+
"loss": 0.3462,
|
| 2557 |
+
"step": 364
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 2.1067821067821066,
|
| 2561 |
+
"grad_norm": 0.15633838545430417,
|
| 2562 |
+
"learning_rate": 1.9612871337655882e-05,
|
| 2563 |
+
"loss": 0.3474,
|
| 2564 |
+
"step": 365
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 2.1125541125541125,
|
| 2568 |
+
"grad_norm": 0.15747516923085514,
|
| 2569 |
+
"learning_rate": 1.938182118383787e-05,
|
| 2570 |
+
"loss": 0.3532,
|
| 2571 |
+
"step": 366
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 2.1183261183261184,
|
| 2575 |
+
"grad_norm": 0.15811521645472146,
|
| 2576 |
+
"learning_rate": 1.9151704100926536e-05,
|
| 2577 |
+
"loss": 0.348,
|
| 2578 |
+
"step": 367
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 2.1240981240981243,
|
| 2582 |
+
"grad_norm": 0.14644722471559174,
|
| 2583 |
+
"learning_rate": 1.892253050281715e-05,
|
| 2584 |
+
"loss": 0.349,
|
| 2585 |
+
"step": 368
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 2.1298701298701297,
|
| 2589 |
+
"grad_norm": 0.1548058583190097,
|
| 2590 |
+
"learning_rate": 1.8694310760707824e-05,
|
| 2591 |
+
"loss": 0.346,
|
| 2592 |
+
"step": 369
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 2.1356421356421356,
|
| 2596 |
+
"grad_norm": 0.13350674475881225,
|
| 2597 |
+
"learning_rate": 1.8467055202630137e-05,
|
| 2598 |
+
"loss": 0.3472,
|
| 2599 |
+
"step": 370
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 2.1414141414141414,
|
| 2603 |
+
"grad_norm": 0.1390929667594732,
|
| 2604 |
+
"learning_rate": 1.8240774112981702e-05,
|
| 2605 |
+
"loss": 0.3482,
|
| 2606 |
+
"step": 371
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 2.1471861471861473,
|
| 2610 |
+
"grad_norm": 0.13904631097945175,
|
| 2611 |
+
"learning_rate": 1.8015477732060853e-05,
|
| 2612 |
+
"loss": 0.3461,
|
| 2613 |
+
"step": 372
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 2.1529581529581527,
|
| 2617 |
+
"grad_norm": 0.1388842546531878,
|
| 2618 |
+
"learning_rate": 1.779117625560309e-05,
|
| 2619 |
+
"loss": 0.3534,
|
| 2620 |
+
"step": 373
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 2.1587301587301586,
|
| 2624 |
+
"grad_norm": 0.1329557989872869,
|
| 2625 |
+
"learning_rate": 1.7567879834319793e-05,
|
| 2626 |
+
"loss": 0.3541,
|
| 2627 |
+
"step": 374
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 2.1645021645021645,
|
| 2631 |
+
"grad_norm": 0.13708881477559062,
|
| 2632 |
+
"learning_rate": 1.734559857343878e-05,
|
| 2633 |
+
"loss": 0.3512,
|
| 2634 |
+
"step": 375
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 2.1702741702741704,
|
| 2638 |
+
"grad_norm": 0.13046281719011577,
|
| 2639 |
+
"learning_rate": 1.7124342532247057e-05,
|
| 2640 |
+
"loss": 0.3453,
|
| 2641 |
+
"step": 376
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 2.1760461760461762,
|
| 2645 |
+
"grad_norm": 0.14597316841401312,
|
| 2646 |
+
"learning_rate": 1.6904121723635477e-05,
|
| 2647 |
+
"loss": 0.3513,
|
| 2648 |
+
"step": 377
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 2.1818181818181817,
|
| 2652 |
+
"grad_norm": 0.15098750108014802,
|
| 2653 |
+
"learning_rate": 1.6684946113645746e-05,
|
| 2654 |
+
"loss": 0.3406,
|
| 2655 |
+
"step": 378
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 2.1875901875901875,
|
| 2659 |
+
"grad_norm": 0.1341433244109453,
|
| 2660 |
+
"learning_rate": 1.6466825621019337e-05,
|
| 2661 |
+
"loss": 0.3469,
|
| 2662 |
+
"step": 379
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 2.1933621933621934,
|
| 2666 |
+
"grad_norm": 0.14986397828347373,
|
| 2667 |
+
"learning_rate": 1.6249770116748604e-05,
|
| 2668 |
+
"loss": 0.3485,
|
| 2669 |
+
"step": 380
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 2.1991341991341993,
|
| 2673 |
+
"grad_norm": 0.13918404166032933,
|
| 2674 |
+
"learning_rate": 1.6033789423630134e-05,
|
| 2675 |
+
"loss": 0.3499,
|
| 2676 |
+
"step": 381
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 2.2049062049062047,
|
| 2680 |
+
"grad_norm": 0.12359065581631808,
|
| 2681 |
+
"learning_rate": 1.5818893315820144e-05,
|
| 2682 |
+
"loss": 0.3467,
|
| 2683 |
+
"step": 382
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 2.2106782106782106,
|
| 2687 |
+
"grad_norm": 0.1513201157889788,
|
| 2688 |
+
"learning_rate": 1.5605091518392234e-05,
|
| 2689 |
+
"loss": 0.3431,
|
| 2690 |
+
"step": 383
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 2.2164502164502164,
|
| 2694 |
+
"grad_norm": 0.12485709624581334,
|
| 2695 |
+
"learning_rate": 1.5392393706897224e-05,
|
| 2696 |
+
"loss": 0.3483,
|
| 2697 |
+
"step": 384
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 2.2222222222222223,
|
| 2701 |
+
"grad_norm": 0.13602588524238346,
|
| 2702 |
+
"learning_rate": 1.5180809506925306e-05,
|
| 2703 |
+
"loss": 0.3532,
|
| 2704 |
+
"step": 385
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 2.227994227994228,
|
| 2708 |
+
"grad_norm": 0.14562695629291877,
|
| 2709 |
+
"learning_rate": 1.4970348493670454e-05,
|
| 2710 |
+
"loss": 0.3483,
|
| 2711 |
+
"step": 386
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 2.2337662337662336,
|
| 2715 |
+
"grad_norm": 0.11828582123138484,
|
| 2716 |
+
"learning_rate": 1.476102019149707e-05,
|
| 2717 |
+
"loss": 0.3516,
|
| 2718 |
+
"step": 387
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 2.2395382395382395,
|
| 2722 |
+
"grad_norm": 0.12945968901614865,
|
| 2723 |
+
"learning_rate": 1.4552834073508958e-05,
|
| 2724 |
+
"loss": 0.3475,
|
| 2725 |
+
"step": 388
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 2.2453102453102454,
|
| 2729 |
+
"grad_norm": 0.1170563090280043,
|
| 2730 |
+
"learning_rate": 1.434579956112069e-05,
|
| 2731 |
+
"loss": 0.3453,
|
| 2732 |
+
"step": 389
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 2.2510822510822512,
|
| 2736 |
+
"grad_norm": 0.1287506811756111,
|
| 2737 |
+
"learning_rate": 1.4139926023631168e-05,
|
| 2738 |
+
"loss": 0.3405,
|
| 2739 |
+
"step": 390
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 2.2568542568542567,
|
| 2743 |
+
"grad_norm": 0.11142792038606325,
|
| 2744 |
+
"learning_rate": 1.3935222777799652e-05,
|
| 2745 |
+
"loss": 0.3467,
|
| 2746 |
+
"step": 391
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 2.2626262626262625,
|
| 2750 |
+
"grad_norm": 0.1274164884676621,
|
| 2751 |
+
"learning_rate": 1.3731699087424115e-05,
|
| 2752 |
+
"loss": 0.3493,
|
| 2753 |
+
"step": 392
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 2.2683982683982684,
|
| 2757 |
+
"grad_norm": 0.11590691394723747,
|
| 2758 |
+
"learning_rate": 1.3529364162922004e-05,
|
| 2759 |
+
"loss": 0.3441,
|
| 2760 |
+
"step": 393
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 2.2741702741702743,
|
| 2764 |
+
"grad_norm": 0.13145419286429477,
|
| 2765 |
+
"learning_rate": 1.3328227160913478e-05,
|
| 2766 |
+
"loss": 0.3432,
|
| 2767 |
+
"step": 394
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 2.27994227994228,
|
| 2771 |
+
"grad_norm": 0.10726667067770969,
|
| 2772 |
+
"learning_rate": 1.3128297183806976e-05,
|
| 2773 |
+
"loss": 0.3454,
|
| 2774 |
+
"step": 395
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 2.2857142857142856,
|
| 2778 |
+
"grad_norm": 0.1144655280850949,
|
| 2779 |
+
"learning_rate": 1.2929583279387296e-05,
|
| 2780 |
+
"loss": 0.3437,
|
| 2781 |
+
"step": 396
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 2.2914862914862915,
|
| 2785 |
+
"grad_norm": 0.11276212616579144,
|
| 2786 |
+
"learning_rate": 1.2732094440406177e-05,
|
| 2787 |
+
"loss": 0.352,
|
| 2788 |
+
"step": 397
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 2.2972582972582973,
|
| 2792 |
+
"grad_norm": 0.11880977448596589,
|
| 2793 |
+
"learning_rate": 1.2535839604175255e-05,
|
| 2794 |
+
"loss": 0.3486,
|
| 2795 |
+
"step": 398
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 2.303030303030303,
|
| 2799 |
+
"grad_norm": 0.10962217727818492,
|
| 2800 |
+
"learning_rate": 1.2340827652161713e-05,
|
| 2801 |
+
"loss": 0.3452,
|
| 2802 |
+
"step": 399
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 2.3088023088023086,
|
| 2806 |
+
"grad_norm": 0.11274713005381089,
|
| 2807 |
+
"learning_rate": 1.2147067409586243e-05,
|
| 2808 |
+
"loss": 0.348,
|
| 2809 |
+
"step": 400
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 2.3145743145743145,
|
| 2813 |
+
"grad_norm": 0.11693801240672108,
|
| 2814 |
+
"learning_rate": 1.1954567645023759e-05,
|
| 2815 |
+
"loss": 0.3535,
|
| 2816 |
+
"step": 401
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 2.3203463203463204,
|
| 2820 |
+
"grad_norm": 0.12532346971561012,
|
| 2821 |
+
"learning_rate": 1.1763337070006515e-05,
|
| 2822 |
+
"loss": 0.3495,
|
| 2823 |
+
"step": 402
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 2.3261183261183263,
|
| 2827 |
+
"grad_norm": 0.10893628624682762,
|
| 2828 |
+
"learning_rate": 1.1573384338629908e-05,
|
| 2829 |
+
"loss": 0.3518,
|
| 2830 |
+
"step": 403
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 2.3318903318903317,
|
| 2834 |
+
"grad_norm": 0.10920863452315147,
|
| 2835 |
+
"learning_rate": 1.138471804716076e-05,
|
| 2836 |
+
"loss": 0.349,
|
| 2837 |
+
"step": 404
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 2.3376623376623376,
|
| 2841 |
+
"grad_norm": 0.12529050366314204,
|
| 2842 |
+
"learning_rate": 1.1197346733648424e-05,
|
| 2843 |
+
"loss": 0.3472,
|
| 2844 |
+
"step": 405
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 2.3434343434343434,
|
| 2848 |
+
"grad_norm": 0.10496748545263929,
|
| 2849 |
+
"learning_rate": 1.1011278877538292e-05,
|
| 2850 |
+
"loss": 0.3517,
|
| 2851 |
+
"step": 406
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 2.3492063492063493,
|
| 2855 |
+
"grad_norm": 0.11719451671854277,
|
| 2856 |
+
"learning_rate": 1.0826522899288077e-05,
|
| 2857 |
+
"loss": 0.3447,
|
| 2858 |
+
"step": 407
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 2.354978354978355,
|
| 2862 |
+
"grad_norm": 0.10969963263744975,
|
| 2863 |
+
"learning_rate": 1.0643087159986796e-05,
|
| 2864 |
+
"loss": 0.3497,
|
| 2865 |
+
"step": 408
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 2.3607503607503606,
|
| 2869 |
+
"grad_norm": 0.10835055548857855,
|
| 2870 |
+
"learning_rate": 1.046097996097633e-05,
|
| 2871 |
+
"loss": 0.3399,
|
| 2872 |
+
"step": 409
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 2.3665223665223665,
|
| 2876 |
+
"grad_norm": 0.10687762232224175,
|
| 2877 |
+
"learning_rate": 1.0280209543475799e-05,
|
| 2878 |
+
"loss": 0.3469,
|
| 2879 |
+
"step": 410
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 2.3722943722943723,
|
| 2883 |
+
"grad_norm": 0.10737382891175314,
|
| 2884 |
+
"learning_rate": 1.0100784088208569e-05,
|
| 2885 |
+
"loss": 0.3498,
|
| 2886 |
+
"step": 411
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 2.378066378066378,
|
| 2890 |
+
"grad_norm": 0.11196662282124958,
|
| 2891 |
+
"learning_rate": 9.922711715032069e-06,
|
| 2892 |
+
"loss": 0.3479,
|
| 2893 |
+
"step": 412
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 2.3838383838383836,
|
| 2897 |
+
"grad_norm": 0.10249258673290118,
|
| 2898 |
+
"learning_rate": 9.746000482570306e-06,
|
| 2899 |
+
"loss": 0.3507,
|
| 2900 |
+
"step": 413
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 2.3896103896103895,
|
| 2904 |
+
"grad_norm": 0.10682683103804237,
|
| 2905 |
+
"learning_rate": 9.570658387849163e-06,
|
| 2906 |
+
"loss": 0.3396,
|
| 2907 |
+
"step": 414
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 2.3953823953823954,
|
| 2911 |
+
"grad_norm": 0.11318251935991214,
|
| 2912 |
+
"learning_rate": 9.396693365934517e-06,
|
| 2913 |
+
"loss": 0.3409,
|
| 2914 |
+
"step": 415
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 2.4011544011544013,
|
| 2918 |
+
"grad_norm": 0.1019527978859083,
|
| 2919 |
+
"learning_rate": 9.22411328957315e-06,
|
| 2920 |
+
"loss": 0.3476,
|
| 2921 |
+
"step": 416
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 2.4069264069264067,
|
| 2925 |
+
"grad_norm": 0.10749236145185602,
|
| 2926 |
+
"learning_rate": 9.052925968836442e-06,
|
| 2927 |
+
"loss": 0.343,
|
| 2928 |
+
"step": 417
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 2.4126984126984126,
|
| 2932 |
+
"grad_norm": 0.0986457791725115,
|
| 2933 |
+
"learning_rate": 8.883139150766941e-06,
|
| 2934 |
+
"loss": 0.3457,
|
| 2935 |
+
"step": 418
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 2.4184704184704184,
|
| 2939 |
+
"grad_norm": 0.10213111766143514,
|
| 2940 |
+
"learning_rate": 8.714760519027789e-06,
|
| 2941 |
+
"loss": 0.3491,
|
| 2942 |
+
"step": 419
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 2.4242424242424243,
|
| 2946 |
+
"grad_norm": 0.10746607521482601,
|
| 2947 |
+
"learning_rate": 8.547797693554946e-06,
|
| 2948 |
+
"loss": 0.3486,
|
| 2949 |
+
"step": 420
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 2.43001443001443,
|
| 2953 |
+
"grad_norm": 0.10441851840468046,
|
| 2954 |
+
"learning_rate": 8.3822582302124e-06,
|
| 2955 |
+
"loss": 0.3456,
|
| 2956 |
+
"step": 421
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 2.4357864357864356,
|
| 2960 |
+
"grad_norm": 0.10831025638581726,
|
| 2961 |
+
"learning_rate": 8.218149620450235e-06,
|
| 2962 |
+
"loss": 0.3411,
|
| 2963 |
+
"step": 422
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 2.4415584415584415,
|
| 2967 |
+
"grad_norm": 0.10028072114634466,
|
| 2968 |
+
"learning_rate": 8.055479290965569e-06,
|
| 2969 |
+
"loss": 0.3414,
|
| 2970 |
+
"step": 423
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 2.4473304473304474,
|
| 2974 |
+
"grad_norm": 0.10675866856021239,
|
| 2975 |
+
"learning_rate": 7.894254603366497e-06,
|
| 2976 |
+
"loss": 0.3382,
|
| 2977 |
+
"step": 424
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 2.4531024531024532,
|
| 2981 |
+
"grad_norm": 0.10790874855042172,
|
| 2982 |
+
"learning_rate": 7.734482853838892e-06,
|
| 2983 |
+
"loss": 0.35,
|
| 2984 |
+
"step": 425
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 2.458874458874459,
|
| 2988 |
+
"grad_norm": 0.10662476000191072,
|
| 2989 |
+
"learning_rate": 7.576171272816286e-06,
|
| 2990 |
+
"loss": 0.3414,
|
| 2991 |
+
"step": 426
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 2.4646464646464645,
|
| 2995 |
+
"grad_norm": 0.10947969228036052,
|
| 2996 |
+
"learning_rate": 7.4193270246526135e-06,
|
| 2997 |
+
"loss": 0.3412,
|
| 2998 |
+
"step": 427
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 2.4704184704184704,
|
| 3002 |
+
"grad_norm": 0.10274131450450585,
|
| 3003 |
+
"learning_rate": 7.263957207297978e-06,
|
| 3004 |
+
"loss": 0.3433,
|
| 3005 |
+
"step": 428
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 2.4761904761904763,
|
| 3009 |
+
"grad_norm": 0.10154033582796251,
|
| 3010 |
+
"learning_rate": 7.1100688519774876e-06,
|
| 3011 |
+
"loss": 0.3356,
|
| 3012 |
+
"step": 429
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 2.481962481962482,
|
| 3016 |
+
"grad_norm": 0.10986813684307772,
|
| 3017 |
+
"learning_rate": 6.957668922873031e-06,
|
| 3018 |
+
"loss": 0.3526,
|
| 3019 |
+
"step": 430
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 2.4877344877344876,
|
| 3023 |
+
"grad_norm": 0.11268326800105738,
|
| 3024 |
+
"learning_rate": 6.8067643168080766e-06,
|
| 3025 |
+
"loss": 0.3488,
|
| 3026 |
+
"step": 431
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 2.4935064935064934,
|
| 3030 |
+
"grad_norm": 0.09840769427350186,
|
| 3031 |
+
"learning_rate": 6.6573618629356274e-06,
|
| 3032 |
+
"loss": 0.3452,
|
| 3033 |
+
"step": 432
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 2.4992784992784993,
|
| 3037 |
+
"grad_norm": 0.11269175164918041,
|
| 3038 |
+
"learning_rate": 6.509468322429131e-06,
|
| 3039 |
+
"loss": 0.347,
|
| 3040 |
+
"step": 433
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 2.505050505050505,
|
| 3044 |
+
"grad_norm": 0.10999136780654757,
|
| 3045 |
+
"learning_rate": 6.36309038817652e-06,
|
| 3046 |
+
"loss": 0.3455,
|
| 3047 |
+
"step": 434
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 2.5108225108225106,
|
| 3051 |
+
"grad_norm": 0.09840209750950327,
|
| 3052 |
+
"learning_rate": 6.218234684477278e-06,
|
| 3053 |
+
"loss": 0.3439,
|
| 3054 |
+
"step": 435
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 2.5165945165945165,
|
| 3058 |
+
"grad_norm": 0.09447419631866041,
|
| 3059 |
+
"learning_rate": 6.07490776674275e-06,
|
| 3060 |
+
"loss": 0.3378,
|
| 3061 |
+
"step": 436
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 2.5223665223665224,
|
| 3065 |
+
"grad_norm": 0.09706328476200818,
|
| 3066 |
+
"learning_rate": 5.933116121199378e-06,
|
| 3067 |
+
"loss": 0.3452,
|
| 3068 |
+
"step": 437
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 2.5281385281385282,
|
| 3072 |
+
"grad_norm": 0.10231049024580399,
|
| 3073 |
+
"learning_rate": 5.792866164595245e-06,
|
| 3074 |
+
"loss": 0.3368,
|
| 3075 |
+
"step": 438
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 2.533910533910534,
|
| 3079 |
+
"grad_norm": 0.09978034629381005,
|
| 3080 |
+
"learning_rate": 5.654164243909659e-06,
|
| 3081 |
+
"loss": 0.3396,
|
| 3082 |
+
"step": 439
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 2.5396825396825395,
|
| 3086 |
+
"grad_norm": 0.10413041011105151,
|
| 3087 |
+
"learning_rate": 5.517016636065915e-06,
|
| 3088 |
+
"loss": 0.3432,
|
| 3089 |
+
"step": 440
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 2.5454545454545454,
|
| 3093 |
+
"grad_norm": 0.0967954090668441,
|
| 3094 |
+
"learning_rate": 5.381429547647239e-06,
|
| 3095 |
+
"loss": 0.3499,
|
| 3096 |
+
"step": 441
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 2.5512265512265513,
|
| 3100 |
+
"grad_norm": 0.10109502518725756,
|
| 3101 |
+
"learning_rate": 5.247409114615898e-06,
|
| 3102 |
+
"loss": 0.3407,
|
| 3103 |
+
"step": 442
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 2.556998556998557,
|
| 3107 |
+
"grad_norm": 0.09490435025341039,
|
| 3108 |
+
"learning_rate": 5.114961402035552e-06,
|
| 3109 |
+
"loss": 0.3424,
|
| 3110 |
+
"step": 443
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 2.562770562770563,
|
| 3114 |
+
"grad_norm": 0.10229915180017767,
|
| 3115 |
+
"learning_rate": 4.984092403796763e-06,
|
| 3116 |
+
"loss": 0.3413,
|
| 3117 |
+
"step": 444
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 2.5685425685425685,
|
| 3121 |
+
"grad_norm": 0.09805972167431506,
|
| 3122 |
+
"learning_rate": 4.854808042345736e-06,
|
| 3123 |
+
"loss": 0.3373,
|
| 3124 |
+
"step": 445
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 2.5743145743145743,
|
| 3128 |
+
"grad_norm": 0.09581515115465772,
|
| 3129 |
+
"learning_rate": 4.7271141684163e-06,
|
| 3130 |
+
"loss": 0.3486,
|
| 3131 |
+
"step": 446
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 2.58008658008658,
|
| 3135 |
+
"grad_norm": 0.09484063048204838,
|
| 3136 |
+
"learning_rate": 4.601016560765144e-06,
|
| 3137 |
+
"loss": 0.3419,
|
| 3138 |
+
"step": 447
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 2.5858585858585856,
|
| 3142 |
+
"grad_norm": 0.09536057522665425,
|
| 3143 |
+
"learning_rate": 4.476520925910284e-06,
|
| 3144 |
+
"loss": 0.3366,
|
| 3145 |
+
"step": 448
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 2.5916305916305915,
|
| 3149 |
+
"grad_norm": 0.09987541239632843,
|
| 3150 |
+
"learning_rate": 4.353632897872846e-06,
|
| 3151 |
+
"loss": 0.3425,
|
| 3152 |
+
"step": 449
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 2.5974025974025974,
|
| 3156 |
+
"grad_norm": 0.09963505235706265,
|
| 3157 |
+
"learning_rate": 4.2323580379220755e-06,
|
| 3158 |
+
"loss": 0.3484,
|
| 3159 |
+
"step": 450
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 2.6031746031746033,
|
| 3163 |
+
"grad_norm": 0.09253983593459067,
|
| 3164 |
+
"learning_rate": 4.112701834323672e-06,
|
| 3165 |
+
"loss": 0.3429,
|
| 3166 |
+
"step": 451
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 2.608946608946609,
|
| 3170 |
+
"grad_norm": 0.09199990563523672,
|
| 3171 |
+
"learning_rate": 3.994669702091418e-06,
|
| 3172 |
+
"loss": 0.3351,
|
| 3173 |
+
"step": 452
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 2.6147186147186146,
|
| 3177 |
+
"grad_norm": 0.09874324626738527,
|
| 3178 |
+
"learning_rate": 3.878266982742109e-06,
|
| 3179 |
+
"loss": 0.3464,
|
| 3180 |
+
"step": 453
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 2.6204906204906204,
|
| 3184 |
+
"grad_norm": 0.10055349590155505,
|
| 3185 |
+
"learning_rate": 3.7634989440538603e-06,
|
| 3186 |
+
"loss": 0.3493,
|
| 3187 |
+
"step": 454
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 2.6262626262626263,
|
| 3191 |
+
"grad_norm": 0.09935024527889372,
|
| 3192 |
+
"learning_rate": 3.650370779827661e-06,
|
| 3193 |
+
"loss": 0.3407,
|
| 3194 |
+
"step": 455
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 2.632034632034632,
|
| 3198 |
+
"grad_norm": 0.09534517733096698,
|
| 3199 |
+
"learning_rate": 3.538887609652388e-06,
|
| 3200 |
+
"loss": 0.3378,
|
| 3201 |
+
"step": 456
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 2.637806637806638,
|
| 3205 |
+
"grad_norm": 0.09350817574484052,
|
| 3206 |
+
"learning_rate": 3.4290544786730996e-06,
|
| 3207 |
+
"loss": 0.3414,
|
| 3208 |
+
"step": 457
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 2.6435786435786435,
|
| 3212 |
+
"grad_norm": 0.10205262104677329,
|
| 3213 |
+
"learning_rate": 3.3208763573626766e-06,
|
| 3214 |
+
"loss": 0.3423,
|
| 3215 |
+
"step": 458
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 2.6493506493506493,
|
| 3219 |
+
"grad_norm": 0.09409112126481023,
|
| 3220 |
+
"learning_rate": 3.214358141296949e-06,
|
| 3221 |
+
"loss": 0.3454,
|
| 3222 |
+
"step": 459
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 2.655122655122655,
|
| 3226 |
+
"grad_norm": 0.0917867834220046,
|
| 3227 |
+
"learning_rate": 3.1095046509331196e-06,
|
| 3228 |
+
"loss": 0.3406,
|
| 3229 |
+
"step": 460
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 2.6608946608946606,
|
| 3233 |
+
"grad_norm": 0.09278650920624583,
|
| 3234 |
+
"learning_rate": 3.006320631391613e-06,
|
| 3235 |
+
"loss": 0.3476,
|
| 3236 |
+
"step": 461
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 2.6666666666666665,
|
| 3240 |
+
"grad_norm": 0.0897517799081907,
|
| 3241 |
+
"learning_rate": 2.904810752241316e-06,
|
| 3242 |
+
"loss": 0.3351,
|
| 3243 |
+
"step": 462
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 2.6724386724386724,
|
| 3247 |
+
"grad_norm": 0.08778981943962322,
|
| 3248 |
+
"learning_rate": 2.804979607288312e-06,
|
| 3249 |
+
"loss": 0.3364,
|
| 3250 |
+
"step": 463
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 2.6782106782106783,
|
| 3254 |
+
"grad_norm": 0.09570343221719087,
|
| 3255 |
+
"learning_rate": 2.706831714367932e-06,
|
| 3256 |
+
"loss": 0.351,
|
| 3257 |
+
"step": 464
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 2.683982683982684,
|
| 3261 |
+
"grad_norm": 0.09311012426822064,
|
| 3262 |
+
"learning_rate": 2.6103715151403463e-06,
|
| 3263 |
+
"loss": 0.3432,
|
| 3264 |
+
"step": 465
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 2.6897546897546896,
|
| 3268 |
+
"grad_norm": 0.09350413954762375,
|
| 3269 |
+
"learning_rate": 2.515603374889519e-06,
|
| 3270 |
+
"loss": 0.3444,
|
| 3271 |
+
"step": 466
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 2.6955266955266954,
|
| 3275 |
+
"grad_norm": 0.08867208161846349,
|
| 3276 |
+
"learning_rate": 2.4225315823257e-06,
|
| 3277 |
+
"loss": 0.3396,
|
| 3278 |
+
"step": 467
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 2.7012987012987013,
|
| 3282 |
+
"grad_norm": 0.09924016061963505,
|
| 3283 |
+
"learning_rate": 2.3311603493912927e-06,
|
| 3284 |
+
"loss": 0.3452,
|
| 3285 |
+
"step": 468
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 2.707070707070707,
|
| 3289 |
+
"grad_norm": 0.09992707634020694,
|
| 3290 |
+
"learning_rate": 2.241493811070279e-06,
|
| 3291 |
+
"loss": 0.3468,
|
| 3292 |
+
"step": 469
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 2.712842712842713,
|
| 3296 |
+
"grad_norm": 0.09281248365094172,
|
| 3297 |
+
"learning_rate": 2.1535360252010885e-06,
|
| 3298 |
+
"loss": 0.3399,
|
| 3299 |
+
"step": 470
|
| 3300 |
+
},
|
| 3301 |
+
{
|
| 3302 |
+
"epoch": 2.7186147186147185,
|
| 3303 |
+
"grad_norm": 0.08481302767106144,
|
| 3304 |
+
"learning_rate": 2.0672909722929412e-06,
|
| 3305 |
+
"loss": 0.3356,
|
| 3306 |
+
"step": 471
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 2.7243867243867244,
|
| 3310 |
+
"grad_norm": 0.0891639961832306,
|
| 3311 |
+
"learning_rate": 1.982762555345734e-06,
|
| 3312 |
+
"loss": 0.3458,
|
| 3313 |
+
"step": 472
|
| 3314 |
+
},
|
| 3315 |
+
{
|
| 3316 |
+
"epoch": 2.7301587301587302,
|
| 3317 |
+
"grad_norm": 0.0892742474664926,
|
| 3318 |
+
"learning_rate": 1.8999545996734037e-06,
|
| 3319 |
+
"loss": 0.3379,
|
| 3320 |
+
"step": 473
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 2.7359307359307357,
|
| 3324 |
+
"grad_norm": 0.09452191488114642,
|
| 3325 |
+
"learning_rate": 1.8188708527307942e-06,
|
| 3326 |
+
"loss": 0.3419,
|
| 3327 |
+
"step": 474
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 2.741702741702742,
|
| 3331 |
+
"grad_norm": 0.08990569611001864,
|
| 3332 |
+
"learning_rate": 1.7395149839440994e-06,
|
| 3333 |
+
"loss": 0.3531,
|
| 3334 |
+
"step": 475
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 2.7474747474747474,
|
| 3338 |
+
"grad_norm": 0.09350546944764407,
|
| 3339 |
+
"learning_rate": 1.6618905845447697e-06,
|
| 3340 |
+
"loss": 0.339,
|
| 3341 |
+
"step": 476
|
| 3342 |
+
},
|
| 3343 |
+
{
|
| 3344 |
+
"epoch": 2.7532467532467533,
|
| 3345 |
+
"grad_norm": 0.08986097068862921,
|
| 3346 |
+
"learning_rate": 1.5860011674070275e-06,
|
| 3347 |
+
"loss": 0.3501,
|
| 3348 |
+
"step": 477
|
| 3349 |
+
},
|
| 3350 |
+
{
|
| 3351 |
+
"epoch": 2.759018759018759,
|
| 3352 |
+
"grad_norm": 0.09079040328351605,
|
| 3353 |
+
"learning_rate": 1.5118501668888664e-06,
|
| 3354 |
+
"loss": 0.3398,
|
| 3355 |
+
"step": 478
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 2.7647907647907646,
|
| 3359 |
+
"grad_norm": 0.11261442419130217,
|
| 3360 |
+
"learning_rate": 1.4394409386766461e-06,
|
| 3361 |
+
"loss": 0.3463,
|
| 3362 |
+
"step": 479
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 2.7705627705627704,
|
| 3366 |
+
"grad_norm": 0.08862585291124896,
|
| 3367 |
+
"learning_rate": 1.3687767596332058e-06,
|
| 3368 |
+
"loss": 0.3353,
|
| 3369 |
+
"step": 480
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 2.7763347763347763,
|
| 3373 |
+
"grad_norm": 0.08868031485746894,
|
| 3374 |
+
"learning_rate": 1.2998608276496128e-06,
|
| 3375 |
+
"loss": 0.3454,
|
| 3376 |
+
"step": 481
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 2.782106782106782,
|
| 3380 |
+
"grad_norm": 0.08746810675950548,
|
| 3381 |
+
"learning_rate": 1.232696261500399e-06,
|
| 3382 |
+
"loss": 0.3379,
|
| 3383 |
+
"step": 482
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 2.787878787878788,
|
| 3387 |
+
"grad_norm": 0.08636658234472545,
|
| 3388 |
+
"learning_rate": 1.1672861007024472e-06,
|
| 3389 |
+
"loss": 0.3445,
|
| 3390 |
+
"step": 483
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 2.7936507936507935,
|
| 3394 |
+
"grad_norm": 0.08321719005621234,
|
| 3395 |
+
"learning_rate": 1.1036333053774429e-06,
|
| 3396 |
+
"loss": 0.3381,
|
| 3397 |
+
"step": 484
|
| 3398 |
+
},
|
| 3399 |
+
{
|
| 3400 |
+
"epoch": 2.7994227994227994,
|
| 3401 |
+
"grad_norm": 0.08872817366905046,
|
| 3402 |
+
"learning_rate": 1.0417407561178839e-06,
|
| 3403 |
+
"loss": 0.346,
|
| 3404 |
+
"step": 485
|
| 3405 |
+
},
|
| 3406 |
+
{
|
| 3407 |
+
"epoch": 2.8051948051948052,
|
| 3408 |
+
"grad_norm": 0.08920218308026576,
|
| 3409 |
+
"learning_rate": 9.816112538567578e-07,
|
| 3410 |
+
"loss": 0.3417,
|
| 3411 |
+
"step": 486
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"epoch": 2.810966810966811,
|
| 3415 |
+
"grad_norm": 0.08547034523589102,
|
| 3416 |
+
"learning_rate": 9.232475197407553e-07,
|
| 3417 |
+
"loss": 0.3418,
|
| 3418 |
+
"step": 487
|
| 3419 |
+
},
|
| 3420 |
+
{
|
| 3421 |
+
"epoch": 2.816738816738817,
|
| 3422 |
+
"grad_norm": 0.08471208619998576,
|
| 3423 |
+
"learning_rate": 8.666521950071538e-07,
|
| 3424 |
+
"loss": 0.3361,
|
| 3425 |
+
"step": 488
|
| 3426 |
+
},
|
| 3427 |
+
{
|
| 3428 |
+
"epoch": 2.8225108225108224,
|
| 3429 |
+
"grad_norm": 0.08685346174962041,
|
| 3430 |
+
"learning_rate": 8.118278408642655e-07,
|
| 3431 |
+
"loss": 0.3417,
|
| 3432 |
+
"step": 489
|
| 3433 |
+
},
|
| 3434 |
+
{
|
| 3435 |
+
"epoch": 2.8282828282828283,
|
| 3436 |
+
"grad_norm": 0.09031115384378957,
|
| 3437 |
+
"learning_rate": 7.587769383755383e-07,
|
| 3438 |
+
"loss": 0.3476,
|
| 3439 |
+
"step": 490
|
| 3440 |
+
},
|
| 3441 |
+
{
|
| 3442 |
+
"epoch": 2.834054834054834,
|
| 3443 |
+
"grad_norm": 0.08344504301313972,
|
| 3444 |
+
"learning_rate": 7.07501888347264e-07,
|
| 3445 |
+
"loss": 0.3498,
|
| 3446 |
+
"step": 491
|
| 3447 |
+
},
|
| 3448 |
+
{
|
| 3449 |
+
"epoch": 2.8398268398268396,
|
| 3450 |
+
"grad_norm": 0.08370353552469191,
|
| 3451 |
+
"learning_rate": 6.580050112199709e-07,
|
| 3452 |
+
"loss": 0.3358,
|
| 3453 |
+
"step": 492
|
| 3454 |
+
},
|
| 3455 |
+
{
|
| 3456 |
+
"epoch": 2.8455988455988455,
|
| 3457 |
+
"grad_norm": 0.08633408464372305,
|
| 3458 |
+
"learning_rate": 6.10288546963358e-07,
|
| 3459 |
+
"loss": 0.3427,
|
| 3460 |
+
"step": 493
|
| 3461 |
+
},
|
| 3462 |
+
{
|
| 3463 |
+
"epoch": 2.8513708513708513,
|
| 3464 |
+
"grad_norm": 0.08347643597997188,
|
| 3465 |
+
"learning_rate": 5.643546549749656e-07,
|
| 3466 |
+
"loss": 0.3361,
|
| 3467 |
+
"step": 494
|
| 3468 |
+
},
|
| 3469 |
+
{
|
| 3470 |
+
"epoch": 2.857142857142857,
|
| 3471 |
+
"grad_norm": 0.08240574474272953,
|
| 3472 |
+
"learning_rate": 5.202054139824419e-07,
|
| 3473 |
+
"loss": 0.3416,
|
| 3474 |
+
"step": 495
|
| 3475 |
+
},
|
| 3476 |
+
{
|
| 3477 |
+
"epoch": 2.862914862914863,
|
| 3478 |
+
"grad_norm": 0.08206414021141857,
|
| 3479 |
+
"learning_rate": 4.778428219494613e-07,
|
| 3480 |
+
"loss": 0.3431,
|
| 3481 |
+
"step": 496
|
| 3482 |
+
},
|
| 3483 |
+
{
|
| 3484 |
+
"epoch": 2.8686868686868685,
|
| 3485 |
+
"grad_norm": 0.08094241936651854,
|
| 3486 |
+
"learning_rate": 4.372687959853039e-07,
|
| 3487 |
+
"loss": 0.3372,
|
| 3488 |
+
"step": 497
|
| 3489 |
+
},
|
| 3490 |
+
{
|
| 3491 |
+
"epoch": 2.8744588744588744,
|
| 3492 |
+
"grad_norm": 0.08376193752213902,
|
| 3493 |
+
"learning_rate": 3.984851722581251e-07,
|
| 3494 |
+
"loss": 0.3337,
|
| 3495 |
+
"step": 498
|
| 3496 |
+
},
|
| 3497 |
+
{
|
| 3498 |
+
"epoch": 2.8802308802308803,
|
| 3499 |
+
"grad_norm": 0.0833016035487455,
|
| 3500 |
+
"learning_rate": 3.614937059118351e-07,
|
| 3501 |
+
"loss": 0.3453,
|
| 3502 |
+
"step": 499
|
| 3503 |
+
},
|
| 3504 |
+
{
|
| 3505 |
+
"epoch": 2.886002886002886,
|
| 3506 |
+
"grad_norm": 0.08536696753078746,
|
| 3507 |
+
"learning_rate": 3.262960709866692e-07,
|
| 3508 |
+
"loss": 0.3392,
|
| 3509 |
+
"step": 500
|
| 3510 |
+
},
|
| 3511 |
+
{
|
| 3512 |
+
"epoch": 2.891774891774892,
|
| 3513 |
+
"grad_norm": 0.08538387172597513,
|
| 3514 |
+
"learning_rate": 2.9289386034344835e-07,
|
| 3515 |
+
"loss": 0.3441,
|
| 3516 |
+
"step": 501
|
| 3517 |
+
},
|
| 3518 |
+
{
|
| 3519 |
+
"epoch": 2.8975468975468974,
|
| 3520 |
+
"grad_norm": 0.08180946091642226,
|
| 3521 |
+
"learning_rate": 2.6128858559147705e-07,
|
| 3522 |
+
"loss": 0.3376,
|
| 3523 |
+
"step": 502
|
| 3524 |
+
},
|
| 3525 |
+
{
|
| 3526 |
+
"epoch": 2.9033189033189033,
|
| 3527 |
+
"grad_norm": 0.08190246281621259,
|
| 3528 |
+
"learning_rate": 2.3148167702014e-07,
|
| 3529 |
+
"loss": 0.3393,
|
| 3530 |
+
"step": 503
|
| 3531 |
+
},
|
| 3532 |
+
{
|
| 3533 |
+
"epoch": 2.909090909090909,
|
| 3534 |
+
"grad_norm": 0.0829720124973207,
|
| 3535 |
+
"learning_rate": 2.0347448353418508e-07,
|
| 3536 |
+
"loss": 0.34,
|
| 3537 |
+
"step": 504
|
| 3538 |
+
},
|
| 3539 |
+
{
|
| 3540 |
+
"epoch": 2.9148629148629146,
|
| 3541 |
+
"grad_norm": 0.08167916104889678,
|
| 3542 |
+
"learning_rate": 1.7726827259266554e-07,
|
| 3543 |
+
"loss": 0.336,
|
| 3544 |
+
"step": 505
|
| 3545 |
+
},
|
| 3546 |
+
{
|
| 3547 |
+
"epoch": 2.9206349206349205,
|
| 3548 |
+
"grad_norm": 0.08161907646122096,
|
| 3549 |
+
"learning_rate": 1.5286423015158591e-07,
|
| 3550 |
+
"loss": 0.3353,
|
| 3551 |
+
"step": 506
|
| 3552 |
+
},
|
| 3553 |
+
{
|
| 3554 |
+
"epoch": 2.9264069264069263,
|
| 3555 |
+
"grad_norm": 0.08545137866828025,
|
| 3556 |
+
"learning_rate": 1.302634606102382e-07,
|
| 3557 |
+
"loss": 0.3412,
|
| 3558 |
+
"step": 507
|
| 3559 |
+
},
|
| 3560 |
+
{
|
| 3561 |
+
"epoch": 2.932178932178932,
|
| 3562 |
+
"grad_norm": 0.08380300828431841,
|
| 3563 |
+
"learning_rate": 1.0946698676120637e-07,
|
| 3564 |
+
"loss": 0.3451,
|
| 3565 |
+
"step": 508
|
| 3566 |
+
},
|
| 3567 |
+
{
|
| 3568 |
+
"epoch": 2.937950937950938,
|
| 3569 |
+
"grad_norm": 0.08344201696465267,
|
| 3570 |
+
"learning_rate": 9.047574974410555e-08,
|
| 3571 |
+
"loss": 0.3426,
|
| 3572 |
+
"step": 509
|
| 3573 |
+
},
|
| 3574 |
+
{
|
| 3575 |
+
"epoch": 2.9437229437229435,
|
| 3576 |
+
"grad_norm": 0.08265827123035746,
|
| 3577 |
+
"learning_rate": 7.32906090029628e-08,
|
| 3578 |
+
"loss": 0.3424,
|
| 3579 |
+
"step": 510
|
| 3580 |
+
},
|
| 3581 |
+
{
|
| 3582 |
+
"epoch": 2.9494949494949494,
|
| 3583 |
+
"grad_norm": 0.08102461126707187,
|
| 3584 |
+
"learning_rate": 5.791234224734155e-08,
|
| 3585 |
+
"loss": 0.3378,
|
| 3586 |
+
"step": 511
|
| 3587 |
+
},
|
| 3588 |
+
{
|
| 3589 |
+
"epoch": 2.9552669552669553,
|
| 3590 |
+
"grad_norm": 0.08230714376381316,
|
| 3591 |
+
"learning_rate": 4.434164541714747e-08,
|
| 3592 |
+
"loss": 0.344,
|
| 3593 |
+
"step": 512
|
| 3594 |
+
},
|
| 3595 |
+
{
|
| 3596 |
+
"epoch": 2.961038961038961,
|
| 3597 |
+
"grad_norm": 0.08355558229737302,
|
| 3598 |
+
"learning_rate": 3.2579132651111566e-08,
|
| 3599 |
+
"loss": 0.3362,
|
| 3600 |
+
"step": 513
|
| 3601 |
+
},
|
| 3602 |
+
{
|
| 3603 |
+
"epoch": 2.966810966810967,
|
| 3604 |
+
"grad_norm": 0.08254140174865716,
|
| 3605 |
+
"learning_rate": 2.2625336259038955e-08,
|
| 3606 |
+
"loss": 0.3418,
|
| 3607 |
+
"step": 514
|
| 3608 |
+
},
|
| 3609 |
+
{
|
| 3610 |
+
"epoch": 2.9725829725829724,
|
| 3611 |
+
"grad_norm": 0.08243153799305375,
|
| 3612 |
+
"learning_rate": 1.4480706697668213e-08,
|
| 3613 |
+
"loss": 0.3437,
|
| 3614 |
+
"step": 515
|
| 3615 |
+
},
|
| 3616 |
+
{
|
| 3617 |
+
"epoch": 2.9783549783549783,
|
| 3618 |
+
"grad_norm": 0.08625002296494041,
|
| 3619 |
+
"learning_rate": 8.145612550327642e-09,
|
| 3620 |
+
"loss": 0.3529,
|
| 3621 |
+
"step": 516
|
| 3622 |
+
},
|
| 3623 |
+
{
|
| 3624 |
+
"epoch": 2.984126984126984,
|
| 3625 |
+
"grad_norm": 0.08225559535813236,
|
| 3626 |
+
"learning_rate": 3.620340510233078e-09,
|
| 3627 |
+
"loss": 0.3441,
|
| 3628 |
+
"step": 517
|
| 3629 |
+
},
|
| 3630 |
+
{
|
| 3631 |
+
"epoch": 2.98989898989899,
|
| 3632 |
+
"grad_norm": 0.08106046733970695,
|
| 3633 |
+
"learning_rate": 9.050953675293627e-10,
|
| 3634 |
+
"loss": 0.3343,
|
| 3635 |
+
"step": 518
|
| 3636 |
+
},
|
| 3637 |
+
{
|
| 3638 |
+
"epoch": 2.995670995670996,
|
| 3639 |
+
"grad_norm": 0.08184675524305934,
|
| 3640 |
+
"learning_rate": 0.0,
|
| 3641 |
+
"loss": 0.3398,
|
| 3642 |
+
"step": 519
|
| 3643 |
+
},
|
| 3644 |
+
{
|
| 3645 |
+
"epoch": 2.995670995670996,
|
| 3646 |
+
"step": 519,
|
| 3647 |
+
"total_flos": 1.0686069936733815e+19,
|
| 3648 |
+
"train_loss": 0.4367897748258073,
|
| 3649 |
+
"train_runtime": 27788.7927,
|
| 3650 |
+
"train_samples_per_second": 9.571,
|
| 3651 |
+
"train_steps_per_second": 0.019
|
| 3652 |
+
}
|
| 3653 |
+
],
|
| 3654 |
+
"logging_steps": 1.0,
|
| 3655 |
+
"max_steps": 519,
|
| 3656 |
+
"num_input_tokens_seen": 0,
|
| 3657 |
+
"num_train_epochs": 3,
|
| 3658 |
+
"save_steps": 500,
|
| 3659 |
+
"stateful_callbacks": {
|
| 3660 |
+
"TrainerControl": {
|
| 3661 |
+
"args": {
|
| 3662 |
+
"should_epoch_stop": false,
|
| 3663 |
+
"should_evaluate": false,
|
| 3664 |
+
"should_log": false,
|
| 3665 |
+
"should_save": true,
|
| 3666 |
+
"should_training_stop": true
|
| 3667 |
+
},
|
| 3668 |
+
"attributes": {}
|
| 3669 |
+
}
|
| 3670 |
+
},
|
| 3671 |
+
"total_flos": 1.0686069936733815e+19,
|
| 3672 |
+
"train_batch_size": 1,
|
| 3673 |
+
"trial_name": null,
|
| 3674 |
+
"trial_params": null
|
| 3675 |
+
}
|
training_loss.png
ADDED
|