End of training
Browse files- README.md +2 -1
- all_results.json +12 -0
- eval_results.json +7 -0
- train_results.json +8 -0
- trainer_state.json +3328 -0
- training_eval_loss.png +0 -0
- training_loss.png +0 -0
README.md
CHANGED
|
@@ -4,6 +4,7 @@ license: llama3.1
|
|
| 4 |
base_model: meta-llama/Meta-Llama-3.1-8B
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
|
|
|
| 7 |
- generated_from_trainer
|
| 8 |
model-index:
|
| 9 |
- name: llama3-1_8b_dcft-it
|
|
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
# llama3-1_8b_dcft-it
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B) on
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
- Loss: 0.7122
|
| 21 |
|
|
|
|
| 4 |
base_model: meta-llama/Meta-Llama-3.1-8B
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
| 7 |
+
- full
|
| 8 |
- generated_from_trainer
|
| 9 |
model-index:
|
| 10 |
- name: llama3-1_8b_dcft-it
|
|
|
|
| 16 |
|
| 17 |
# llama3-1_8b_dcft-it
|
| 18 |
|
| 19 |
+
This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B) on the mlfoundations-dev/DCLM-IT-Dataset dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
- Loss: 0.7122
|
| 22 |
|
all_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_loss": 0.7122252583503723,
|
| 4 |
+
"eval_runtime": 140.6773,
|
| 5 |
+
"eval_samples_per_second": 297.866,
|
| 6 |
+
"eval_steps_per_second": 1.166,
|
| 7 |
+
"total_flos": 2.2026686217575832e+20,
|
| 8 |
+
"train_loss": 0.6892199146274841,
|
| 9 |
+
"train_runtime": 30399.4314,
|
| 10 |
+
"train_samples_per_second": 78.569,
|
| 11 |
+
"train_steps_per_second": 0.153
|
| 12 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_loss": 0.7122252583503723,
|
| 4 |
+
"eval_runtime": 140.6773,
|
| 5 |
+
"eval_samples_per_second": 297.866,
|
| 6 |
+
"eval_steps_per_second": 1.166
|
| 7 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 2.2026686217575832e+20,
|
| 4 |
+
"train_loss": 0.6892199146274841,
|
| 5 |
+
"train_runtime": 30399.4314,
|
| 6 |
+
"train_samples_per_second": 78.569,
|
| 7 |
+
"train_steps_per_second": 0.153
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3328 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 4665,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.006430868167202572,
|
| 13 |
+
"grad_norm": 4.730106830596924,
|
| 14 |
+
"learning_rate": 5e-06,
|
| 15 |
+
"loss": 0.8965,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.012861736334405145,
|
| 20 |
+
"grad_norm": 2.0916528701782227,
|
| 21 |
+
"learning_rate": 5e-06,
|
| 22 |
+
"loss": 0.8501,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.01929260450160772,
|
| 27 |
+
"grad_norm": 0.8616329431533813,
|
| 28 |
+
"learning_rate": 5e-06,
|
| 29 |
+
"loss": 0.8318,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.02572347266881029,
|
| 34 |
+
"grad_norm": 0.7073356509208679,
|
| 35 |
+
"learning_rate": 5e-06,
|
| 36 |
+
"loss": 0.8193,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.03215434083601286,
|
| 41 |
+
"grad_norm": 1.3553227186203003,
|
| 42 |
+
"learning_rate": 5e-06,
|
| 43 |
+
"loss": 0.8094,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.03858520900321544,
|
| 48 |
+
"grad_norm": 0.6813921928405762,
|
| 49 |
+
"learning_rate": 5e-06,
|
| 50 |
+
"loss": 0.8042,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.04501607717041801,
|
| 55 |
+
"grad_norm": 1.011812448501587,
|
| 56 |
+
"learning_rate": 5e-06,
|
| 57 |
+
"loss": 0.7912,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.05144694533762058,
|
| 62 |
+
"grad_norm": 2.1421170234680176,
|
| 63 |
+
"learning_rate": 5e-06,
|
| 64 |
+
"loss": 0.7894,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.05787781350482315,
|
| 69 |
+
"grad_norm": 1.2183887958526611,
|
| 70 |
+
"learning_rate": 5e-06,
|
| 71 |
+
"loss": 0.7799,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.06430868167202572,
|
| 76 |
+
"grad_norm": 1.1371192932128906,
|
| 77 |
+
"learning_rate": 5e-06,
|
| 78 |
+
"loss": 0.7832,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.0707395498392283,
|
| 83 |
+
"grad_norm": 1.3382622003555298,
|
| 84 |
+
"learning_rate": 5e-06,
|
| 85 |
+
"loss": 0.7791,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.07717041800643087,
|
| 90 |
+
"grad_norm": 0.6512683629989624,
|
| 91 |
+
"learning_rate": 5e-06,
|
| 92 |
+
"loss": 0.7784,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.08360128617363344,
|
| 97 |
+
"grad_norm": 0.632128894329071,
|
| 98 |
+
"learning_rate": 5e-06,
|
| 99 |
+
"loss": 0.7665,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.09003215434083602,
|
| 104 |
+
"grad_norm": 0.4840221405029297,
|
| 105 |
+
"learning_rate": 5e-06,
|
| 106 |
+
"loss": 0.7643,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.09646302250803858,
|
| 111 |
+
"grad_norm": 0.5068333745002747,
|
| 112 |
+
"learning_rate": 5e-06,
|
| 113 |
+
"loss": 0.7772,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.10289389067524116,
|
| 118 |
+
"grad_norm": 0.5646687150001526,
|
| 119 |
+
"learning_rate": 5e-06,
|
| 120 |
+
"loss": 0.7679,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.10932475884244373,
|
| 125 |
+
"grad_norm": 0.9486843943595886,
|
| 126 |
+
"learning_rate": 5e-06,
|
| 127 |
+
"loss": 0.7724,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.1157556270096463,
|
| 132 |
+
"grad_norm": 0.5386438369750977,
|
| 133 |
+
"learning_rate": 5e-06,
|
| 134 |
+
"loss": 0.7615,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.12218649517684887,
|
| 139 |
+
"grad_norm": 0.7876360416412354,
|
| 140 |
+
"learning_rate": 5e-06,
|
| 141 |
+
"loss": 0.767,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.12861736334405144,
|
| 146 |
+
"grad_norm": 0.5463624596595764,
|
| 147 |
+
"learning_rate": 5e-06,
|
| 148 |
+
"loss": 0.7666,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.13504823151125403,
|
| 153 |
+
"grad_norm": 0.7764928340911865,
|
| 154 |
+
"learning_rate": 5e-06,
|
| 155 |
+
"loss": 0.7564,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.1414790996784566,
|
| 160 |
+
"grad_norm": 0.8027799725532532,
|
| 161 |
+
"learning_rate": 5e-06,
|
| 162 |
+
"loss": 0.7546,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.14790996784565916,
|
| 167 |
+
"grad_norm": 1.3793781995773315,
|
| 168 |
+
"learning_rate": 5e-06,
|
| 169 |
+
"loss": 0.7597,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.15434083601286175,
|
| 174 |
+
"grad_norm": 0.4461333155632019,
|
| 175 |
+
"learning_rate": 5e-06,
|
| 176 |
+
"loss": 0.7637,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.1607717041800643,
|
| 181 |
+
"grad_norm": 0.5069792866706848,
|
| 182 |
+
"learning_rate": 5e-06,
|
| 183 |
+
"loss": 0.7663,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.16720257234726688,
|
| 188 |
+
"grad_norm": 0.4338971972465515,
|
| 189 |
+
"learning_rate": 5e-06,
|
| 190 |
+
"loss": 0.7567,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.17363344051446947,
|
| 195 |
+
"grad_norm": 0.5476683974266052,
|
| 196 |
+
"learning_rate": 5e-06,
|
| 197 |
+
"loss": 0.7581,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.18006430868167203,
|
| 202 |
+
"grad_norm": 0.5481943488121033,
|
| 203 |
+
"learning_rate": 5e-06,
|
| 204 |
+
"loss": 0.7591,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.1864951768488746,
|
| 209 |
+
"grad_norm": 0.5772653818130493,
|
| 210 |
+
"learning_rate": 5e-06,
|
| 211 |
+
"loss": 0.754,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.19292604501607716,
|
| 216 |
+
"grad_norm": 0.455953449010849,
|
| 217 |
+
"learning_rate": 5e-06,
|
| 218 |
+
"loss": 0.7601,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.19935691318327975,
|
| 223 |
+
"grad_norm": 0.5032763481140137,
|
| 224 |
+
"learning_rate": 5e-06,
|
| 225 |
+
"loss": 0.7534,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.2057877813504823,
|
| 230 |
+
"grad_norm": 0.5598179697990417,
|
| 231 |
+
"learning_rate": 5e-06,
|
| 232 |
+
"loss": 0.7567,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.21221864951768488,
|
| 237 |
+
"grad_norm": 0.5307891368865967,
|
| 238 |
+
"learning_rate": 5e-06,
|
| 239 |
+
"loss": 0.7701,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.21864951768488747,
|
| 244 |
+
"grad_norm": 0.654017984867096,
|
| 245 |
+
"learning_rate": 5e-06,
|
| 246 |
+
"loss": 0.7565,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.22508038585209003,
|
| 251 |
+
"grad_norm": 0.49319934844970703,
|
| 252 |
+
"learning_rate": 5e-06,
|
| 253 |
+
"loss": 0.7423,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.2315112540192926,
|
| 258 |
+
"grad_norm": 0.5508062243461609,
|
| 259 |
+
"learning_rate": 5e-06,
|
| 260 |
+
"loss": 0.7569,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.2379421221864952,
|
| 265 |
+
"grad_norm": 0.5229071378707886,
|
| 266 |
+
"learning_rate": 5e-06,
|
| 267 |
+
"loss": 0.7557,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.24437299035369775,
|
| 272 |
+
"grad_norm": 0.5349363684654236,
|
| 273 |
+
"learning_rate": 5e-06,
|
| 274 |
+
"loss": 0.7569,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.2508038585209003,
|
| 279 |
+
"grad_norm": 0.4683885872364044,
|
| 280 |
+
"learning_rate": 5e-06,
|
| 281 |
+
"loss": 0.7496,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.2572347266881029,
|
| 286 |
+
"grad_norm": 0.4715147316455841,
|
| 287 |
+
"learning_rate": 5e-06,
|
| 288 |
+
"loss": 0.7528,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.26366559485530544,
|
| 293 |
+
"grad_norm": 0.42438462376594543,
|
| 294 |
+
"learning_rate": 5e-06,
|
| 295 |
+
"loss": 0.7612,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.27009646302250806,
|
| 300 |
+
"grad_norm": 0.7097042798995972,
|
| 301 |
+
"learning_rate": 5e-06,
|
| 302 |
+
"loss": 0.749,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.2765273311897106,
|
| 307 |
+
"grad_norm": 0.43447116017341614,
|
| 308 |
+
"learning_rate": 5e-06,
|
| 309 |
+
"loss": 0.7396,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.2829581993569132,
|
| 314 |
+
"grad_norm": 0.5114174485206604,
|
| 315 |
+
"learning_rate": 5e-06,
|
| 316 |
+
"loss": 0.7471,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.28938906752411575,
|
| 321 |
+
"grad_norm": 0.4831767678260803,
|
| 322 |
+
"learning_rate": 5e-06,
|
| 323 |
+
"loss": 0.7449,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.2958199356913183,
|
| 328 |
+
"grad_norm": 0.49875640869140625,
|
| 329 |
+
"learning_rate": 5e-06,
|
| 330 |
+
"loss": 0.7483,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.3022508038585209,
|
| 335 |
+
"grad_norm": 0.489754855632782,
|
| 336 |
+
"learning_rate": 5e-06,
|
| 337 |
+
"loss": 0.7451,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.3086816720257235,
|
| 342 |
+
"grad_norm": 0.5207638144493103,
|
| 343 |
+
"learning_rate": 5e-06,
|
| 344 |
+
"loss": 0.7463,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.31511254019292606,
|
| 349 |
+
"grad_norm": 0.5335565805435181,
|
| 350 |
+
"learning_rate": 5e-06,
|
| 351 |
+
"loss": 0.7403,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.3215434083601286,
|
| 356 |
+
"grad_norm": 0.5709009766578674,
|
| 357 |
+
"learning_rate": 5e-06,
|
| 358 |
+
"loss": 0.7425,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.3279742765273312,
|
| 363 |
+
"grad_norm": 0.45602336525917053,
|
| 364 |
+
"learning_rate": 5e-06,
|
| 365 |
+
"loss": 0.7506,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.33440514469453375,
|
| 370 |
+
"grad_norm": 0.49307122826576233,
|
| 371 |
+
"learning_rate": 5e-06,
|
| 372 |
+
"loss": 0.7404,
|
| 373 |
+
"step": 520
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.3408360128617363,
|
| 377 |
+
"grad_norm": 0.49949631094932556,
|
| 378 |
+
"learning_rate": 5e-06,
|
| 379 |
+
"loss": 0.7465,
|
| 380 |
+
"step": 530
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.34726688102893893,
|
| 384 |
+
"grad_norm": 0.43178921937942505,
|
| 385 |
+
"learning_rate": 5e-06,
|
| 386 |
+
"loss": 0.7531,
|
| 387 |
+
"step": 540
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.3536977491961415,
|
| 391 |
+
"grad_norm": 0.46599695086479187,
|
| 392 |
+
"learning_rate": 5e-06,
|
| 393 |
+
"loss": 0.7428,
|
| 394 |
+
"step": 550
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.36012861736334406,
|
| 398 |
+
"grad_norm": 0.4541228115558624,
|
| 399 |
+
"learning_rate": 5e-06,
|
| 400 |
+
"loss": 0.7427,
|
| 401 |
+
"step": 560
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.3665594855305466,
|
| 405 |
+
"grad_norm": 0.474062979221344,
|
| 406 |
+
"learning_rate": 5e-06,
|
| 407 |
+
"loss": 0.7408,
|
| 408 |
+
"step": 570
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.3729903536977492,
|
| 412 |
+
"grad_norm": 0.47080013155937195,
|
| 413 |
+
"learning_rate": 5e-06,
|
| 414 |
+
"loss": 0.7423,
|
| 415 |
+
"step": 580
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.37942122186495175,
|
| 419 |
+
"grad_norm": 0.46832209825515747,
|
| 420 |
+
"learning_rate": 5e-06,
|
| 421 |
+
"loss": 0.7383,
|
| 422 |
+
"step": 590
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.3858520900321543,
|
| 426 |
+
"grad_norm": 0.4157029986381531,
|
| 427 |
+
"learning_rate": 5e-06,
|
| 428 |
+
"loss": 0.7509,
|
| 429 |
+
"step": 600
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.39228295819935693,
|
| 433 |
+
"grad_norm": 0.512323260307312,
|
| 434 |
+
"learning_rate": 5e-06,
|
| 435 |
+
"loss": 0.7387,
|
| 436 |
+
"step": 610
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.3987138263665595,
|
| 440 |
+
"grad_norm": 0.4567117393016815,
|
| 441 |
+
"learning_rate": 5e-06,
|
| 442 |
+
"loss": 0.7379,
|
| 443 |
+
"step": 620
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.40514469453376206,
|
| 447 |
+
"grad_norm": 0.4581872522830963,
|
| 448 |
+
"learning_rate": 5e-06,
|
| 449 |
+
"loss": 0.7384,
|
| 450 |
+
"step": 630
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.4115755627009646,
|
| 454 |
+
"grad_norm": 0.43767768144607544,
|
| 455 |
+
"learning_rate": 5e-06,
|
| 456 |
+
"loss": 0.7349,
|
| 457 |
+
"step": 640
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.4180064308681672,
|
| 461 |
+
"grad_norm": 0.5342622399330139,
|
| 462 |
+
"learning_rate": 5e-06,
|
| 463 |
+
"loss": 0.7379,
|
| 464 |
+
"step": 650
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.42443729903536975,
|
| 468 |
+
"grad_norm": 0.561100423336029,
|
| 469 |
+
"learning_rate": 5e-06,
|
| 470 |
+
"loss": 0.7361,
|
| 471 |
+
"step": 660
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.43086816720257237,
|
| 475 |
+
"grad_norm": 0.4456179141998291,
|
| 476 |
+
"learning_rate": 5e-06,
|
| 477 |
+
"loss": 0.7373,
|
| 478 |
+
"step": 670
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.43729903536977494,
|
| 482 |
+
"grad_norm": 0.5494213104248047,
|
| 483 |
+
"learning_rate": 5e-06,
|
| 484 |
+
"loss": 0.7428,
|
| 485 |
+
"step": 680
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.4437299035369775,
|
| 489 |
+
"grad_norm": 0.6552912592887878,
|
| 490 |
+
"learning_rate": 5e-06,
|
| 491 |
+
"loss": 0.7397,
|
| 492 |
+
"step": 690
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.45016077170418006,
|
| 496 |
+
"grad_norm": 0.6181358098983765,
|
| 497 |
+
"learning_rate": 5e-06,
|
| 498 |
+
"loss": 0.7375,
|
| 499 |
+
"step": 700
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.4565916398713826,
|
| 503 |
+
"grad_norm": 0.45060378313064575,
|
| 504 |
+
"learning_rate": 5e-06,
|
| 505 |
+
"loss": 0.7352,
|
| 506 |
+
"step": 710
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.4630225080385852,
|
| 510 |
+
"grad_norm": 0.4362952411174774,
|
| 511 |
+
"learning_rate": 5e-06,
|
| 512 |
+
"loss": 0.7385,
|
| 513 |
+
"step": 720
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.4694533762057878,
|
| 517 |
+
"grad_norm": 0.4858558177947998,
|
| 518 |
+
"learning_rate": 5e-06,
|
| 519 |
+
"loss": 0.7364,
|
| 520 |
+
"step": 730
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.4758842443729904,
|
| 524 |
+
"grad_norm": 0.5299956202507019,
|
| 525 |
+
"learning_rate": 5e-06,
|
| 526 |
+
"loss": 0.7368,
|
| 527 |
+
"step": 740
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.48231511254019294,
|
| 531 |
+
"grad_norm": 0.6190028190612793,
|
| 532 |
+
"learning_rate": 5e-06,
|
| 533 |
+
"loss": 0.7286,
|
| 534 |
+
"step": 750
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.4887459807073955,
|
| 538 |
+
"grad_norm": 0.42335936427116394,
|
| 539 |
+
"learning_rate": 5e-06,
|
| 540 |
+
"loss": 0.7368,
|
| 541 |
+
"step": 760
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.49517684887459806,
|
| 545 |
+
"grad_norm": 0.5196449160575867,
|
| 546 |
+
"learning_rate": 5e-06,
|
| 547 |
+
"loss": 0.7432,
|
| 548 |
+
"step": 770
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.5016077170418006,
|
| 552 |
+
"grad_norm": 0.587582528591156,
|
| 553 |
+
"learning_rate": 5e-06,
|
| 554 |
+
"loss": 0.7405,
|
| 555 |
+
"step": 780
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.5080385852090032,
|
| 559 |
+
"grad_norm": 0.5125169157981873,
|
| 560 |
+
"learning_rate": 5e-06,
|
| 561 |
+
"loss": 0.7346,
|
| 562 |
+
"step": 790
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.5144694533762058,
|
| 566 |
+
"grad_norm": 0.42861226201057434,
|
| 567 |
+
"learning_rate": 5e-06,
|
| 568 |
+
"loss": 0.7343,
|
| 569 |
+
"step": 800
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.5209003215434084,
|
| 573 |
+
"grad_norm": 0.41551926732063293,
|
| 574 |
+
"learning_rate": 5e-06,
|
| 575 |
+
"loss": 0.7401,
|
| 576 |
+
"step": 810
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.5273311897106109,
|
| 580 |
+
"grad_norm": 0.511359453201294,
|
| 581 |
+
"learning_rate": 5e-06,
|
| 582 |
+
"loss": 0.7333,
|
| 583 |
+
"step": 820
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.5337620578778135,
|
| 587 |
+
"grad_norm": 0.4694862961769104,
|
| 588 |
+
"learning_rate": 5e-06,
|
| 589 |
+
"loss": 0.7405,
|
| 590 |
+
"step": 830
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.5401929260450161,
|
| 594 |
+
"grad_norm": 0.48257577419281006,
|
| 595 |
+
"learning_rate": 5e-06,
|
| 596 |
+
"loss": 0.735,
|
| 597 |
+
"step": 840
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.5466237942122186,
|
| 601 |
+
"grad_norm": 0.4408928155899048,
|
| 602 |
+
"learning_rate": 5e-06,
|
| 603 |
+
"loss": 0.7235,
|
| 604 |
+
"step": 850
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.5530546623794212,
|
| 608 |
+
"grad_norm": 0.5172277092933655,
|
| 609 |
+
"learning_rate": 5e-06,
|
| 610 |
+
"loss": 0.7398,
|
| 611 |
+
"step": 860
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.5594855305466238,
|
| 615 |
+
"grad_norm": 0.4982483983039856,
|
| 616 |
+
"learning_rate": 5e-06,
|
| 617 |
+
"loss": 0.7383,
|
| 618 |
+
"step": 870
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.5659163987138264,
|
| 622 |
+
"grad_norm": 0.4628170132637024,
|
| 623 |
+
"learning_rate": 5e-06,
|
| 624 |
+
"loss": 0.741,
|
| 625 |
+
"step": 880
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.572347266881029,
|
| 629 |
+
"grad_norm": 0.46205154061317444,
|
| 630 |
+
"learning_rate": 5e-06,
|
| 631 |
+
"loss": 0.7304,
|
| 632 |
+
"step": 890
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.5787781350482315,
|
| 636 |
+
"grad_norm": 0.4182255268096924,
|
| 637 |
+
"learning_rate": 5e-06,
|
| 638 |
+
"loss": 0.7351,
|
| 639 |
+
"step": 900
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.5852090032154341,
|
| 643 |
+
"grad_norm": 0.5075783729553223,
|
| 644 |
+
"learning_rate": 5e-06,
|
| 645 |
+
"loss": 0.732,
|
| 646 |
+
"step": 910
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.5916398713826366,
|
| 650 |
+
"grad_norm": 0.4244738817214966,
|
| 651 |
+
"learning_rate": 5e-06,
|
| 652 |
+
"loss": 0.7357,
|
| 653 |
+
"step": 920
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.5980707395498392,
|
| 657 |
+
"grad_norm": 0.473067969083786,
|
| 658 |
+
"learning_rate": 5e-06,
|
| 659 |
+
"loss": 0.739,
|
| 660 |
+
"step": 930
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.6045016077170418,
|
| 664 |
+
"grad_norm": 0.4673498272895813,
|
| 665 |
+
"learning_rate": 5e-06,
|
| 666 |
+
"loss": 0.7421,
|
| 667 |
+
"step": 940
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.6109324758842444,
|
| 671 |
+
"grad_norm": 0.41633403301239014,
|
| 672 |
+
"learning_rate": 5e-06,
|
| 673 |
+
"loss": 0.733,
|
| 674 |
+
"step": 950
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.617363344051447,
|
| 678 |
+
"grad_norm": 0.4132135212421417,
|
| 679 |
+
"learning_rate": 5e-06,
|
| 680 |
+
"loss": 0.7306,
|
| 681 |
+
"step": 960
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.6237942122186495,
|
| 685 |
+
"grad_norm": 0.48543354868888855,
|
| 686 |
+
"learning_rate": 5e-06,
|
| 687 |
+
"loss": 0.7267,
|
| 688 |
+
"step": 970
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.6302250803858521,
|
| 692 |
+
"grad_norm": 0.45808976888656616,
|
| 693 |
+
"learning_rate": 5e-06,
|
| 694 |
+
"loss": 0.7339,
|
| 695 |
+
"step": 980
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.6366559485530546,
|
| 699 |
+
"grad_norm": 0.46231311559677124,
|
| 700 |
+
"learning_rate": 5e-06,
|
| 701 |
+
"loss": 0.7321,
|
| 702 |
+
"step": 990
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.6430868167202572,
|
| 706 |
+
"grad_norm": 0.43952593207359314,
|
| 707 |
+
"learning_rate": 5e-06,
|
| 708 |
+
"loss": 0.7244,
|
| 709 |
+
"step": 1000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.6495176848874598,
|
| 713 |
+
"grad_norm": 0.4417444169521332,
|
| 714 |
+
"learning_rate": 5e-06,
|
| 715 |
+
"loss": 0.7339,
|
| 716 |
+
"step": 1010
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.6559485530546624,
|
| 720 |
+
"grad_norm": 0.460893839597702,
|
| 721 |
+
"learning_rate": 5e-06,
|
| 722 |
+
"loss": 0.7317,
|
| 723 |
+
"step": 1020
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.662379421221865,
|
| 727 |
+
"grad_norm": 0.4568830728530884,
|
| 728 |
+
"learning_rate": 5e-06,
|
| 729 |
+
"loss": 0.7294,
|
| 730 |
+
"step": 1030
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.6688102893890675,
|
| 734 |
+
"grad_norm": 0.4672453999519348,
|
| 735 |
+
"learning_rate": 5e-06,
|
| 736 |
+
"loss": 0.7306,
|
| 737 |
+
"step": 1040
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.6752411575562701,
|
| 741 |
+
"grad_norm": 0.677759051322937,
|
| 742 |
+
"learning_rate": 5e-06,
|
| 743 |
+
"loss": 0.7234,
|
| 744 |
+
"step": 1050
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.6816720257234726,
|
| 748 |
+
"grad_norm": 0.6139099597930908,
|
| 749 |
+
"learning_rate": 5e-06,
|
| 750 |
+
"loss": 0.7346,
|
| 751 |
+
"step": 1060
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.6881028938906752,
|
| 755 |
+
"grad_norm": 0.4375672936439514,
|
| 756 |
+
"learning_rate": 5e-06,
|
| 757 |
+
"loss": 0.7323,
|
| 758 |
+
"step": 1070
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.6945337620578779,
|
| 762 |
+
"grad_norm": 0.39761385321617126,
|
| 763 |
+
"learning_rate": 5e-06,
|
| 764 |
+
"loss": 0.7348,
|
| 765 |
+
"step": 1080
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.7009646302250804,
|
| 769 |
+
"grad_norm": 0.4165497422218323,
|
| 770 |
+
"learning_rate": 5e-06,
|
| 771 |
+
"loss": 0.7267,
|
| 772 |
+
"step": 1090
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.707395498392283,
|
| 776 |
+
"grad_norm": 0.4338400661945343,
|
| 777 |
+
"learning_rate": 5e-06,
|
| 778 |
+
"loss": 0.7305,
|
| 779 |
+
"step": 1100
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.7138263665594855,
|
| 783 |
+
"grad_norm": 0.43742889165878296,
|
| 784 |
+
"learning_rate": 5e-06,
|
| 785 |
+
"loss": 0.7311,
|
| 786 |
+
"step": 1110
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.7202572347266881,
|
| 790 |
+
"grad_norm": 0.45418253540992737,
|
| 791 |
+
"learning_rate": 5e-06,
|
| 792 |
+
"loss": 0.7277,
|
| 793 |
+
"step": 1120
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.7266881028938906,
|
| 797 |
+
"grad_norm": 0.42107295989990234,
|
| 798 |
+
"learning_rate": 5e-06,
|
| 799 |
+
"loss": 0.7319,
|
| 800 |
+
"step": 1130
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.7331189710610932,
|
| 804 |
+
"grad_norm": 0.40914231538772583,
|
| 805 |
+
"learning_rate": 5e-06,
|
| 806 |
+
"loss": 0.7251,
|
| 807 |
+
"step": 1140
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.7395498392282959,
|
| 811 |
+
"grad_norm": 0.42050373554229736,
|
| 812 |
+
"learning_rate": 5e-06,
|
| 813 |
+
"loss": 0.7403,
|
| 814 |
+
"step": 1150
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.7459807073954984,
|
| 818 |
+
"grad_norm": 0.4320564568042755,
|
| 819 |
+
"learning_rate": 5e-06,
|
| 820 |
+
"loss": 0.7319,
|
| 821 |
+
"step": 1160
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.752411575562701,
|
| 825 |
+
"grad_norm": 0.4363013803958893,
|
| 826 |
+
"learning_rate": 5e-06,
|
| 827 |
+
"loss": 0.7315,
|
| 828 |
+
"step": 1170
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.7588424437299035,
|
| 832 |
+
"grad_norm": 0.50197434425354,
|
| 833 |
+
"learning_rate": 5e-06,
|
| 834 |
+
"loss": 0.7287,
|
| 835 |
+
"step": 1180
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.7652733118971061,
|
| 839 |
+
"grad_norm": 0.5218095183372498,
|
| 840 |
+
"learning_rate": 5e-06,
|
| 841 |
+
"loss": 0.7294,
|
| 842 |
+
"step": 1190
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.7717041800643086,
|
| 846 |
+
"grad_norm": 0.4605177044868469,
|
| 847 |
+
"learning_rate": 5e-06,
|
| 848 |
+
"loss": 0.7218,
|
| 849 |
+
"step": 1200
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.7781350482315113,
|
| 853 |
+
"grad_norm": 0.4640117287635803,
|
| 854 |
+
"learning_rate": 5e-06,
|
| 855 |
+
"loss": 0.7297,
|
| 856 |
+
"step": 1210
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.7845659163987139,
|
| 860 |
+
"grad_norm": 0.43073731660842896,
|
| 861 |
+
"learning_rate": 5e-06,
|
| 862 |
+
"loss": 0.7299,
|
| 863 |
+
"step": 1220
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.7909967845659164,
|
| 867 |
+
"grad_norm": 0.43865692615509033,
|
| 868 |
+
"learning_rate": 5e-06,
|
| 869 |
+
"loss": 0.7299,
|
| 870 |
+
"step": 1230
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.797427652733119,
|
| 874 |
+
"grad_norm": 0.4125821590423584,
|
| 875 |
+
"learning_rate": 5e-06,
|
| 876 |
+
"loss": 0.738,
|
| 877 |
+
"step": 1240
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.8038585209003215,
|
| 881 |
+
"grad_norm": 0.41375529766082764,
|
| 882 |
+
"learning_rate": 5e-06,
|
| 883 |
+
"loss": 0.7324,
|
| 884 |
+
"step": 1250
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.8102893890675241,
|
| 888 |
+
"grad_norm": 0.43837442994117737,
|
| 889 |
+
"learning_rate": 5e-06,
|
| 890 |
+
"loss": 0.7219,
|
| 891 |
+
"step": 1260
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.8167202572347267,
|
| 895 |
+
"grad_norm": 0.45211347937583923,
|
| 896 |
+
"learning_rate": 5e-06,
|
| 897 |
+
"loss": 0.725,
|
| 898 |
+
"step": 1270
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.8231511254019293,
|
| 902 |
+
"grad_norm": 0.4578869640827179,
|
| 903 |
+
"learning_rate": 5e-06,
|
| 904 |
+
"loss": 0.7307,
|
| 905 |
+
"step": 1280
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.8295819935691319,
|
| 909 |
+
"grad_norm": 0.42134717106819153,
|
| 910 |
+
"learning_rate": 5e-06,
|
| 911 |
+
"loss": 0.7203,
|
| 912 |
+
"step": 1290
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.8360128617363344,
|
| 916 |
+
"grad_norm": 0.4580245912075043,
|
| 917 |
+
"learning_rate": 5e-06,
|
| 918 |
+
"loss": 0.7242,
|
| 919 |
+
"step": 1300
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.842443729903537,
|
| 923 |
+
"grad_norm": 0.4085647165775299,
|
| 924 |
+
"learning_rate": 5e-06,
|
| 925 |
+
"loss": 0.7182,
|
| 926 |
+
"step": 1310
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.8488745980707395,
|
| 930 |
+
"grad_norm": 0.5211206674575806,
|
| 931 |
+
"learning_rate": 5e-06,
|
| 932 |
+
"loss": 0.7259,
|
| 933 |
+
"step": 1320
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.8553054662379421,
|
| 937 |
+
"grad_norm": 0.5008974671363831,
|
| 938 |
+
"learning_rate": 5e-06,
|
| 939 |
+
"loss": 0.7164,
|
| 940 |
+
"step": 1330
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.8617363344051447,
|
| 944 |
+
"grad_norm": 0.4538878798484802,
|
| 945 |
+
"learning_rate": 5e-06,
|
| 946 |
+
"loss": 0.7198,
|
| 947 |
+
"step": 1340
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.8681672025723473,
|
| 951 |
+
"grad_norm": 0.407697468996048,
|
| 952 |
+
"learning_rate": 5e-06,
|
| 953 |
+
"loss": 0.7255,
|
| 954 |
+
"step": 1350
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.8745980707395499,
|
| 958 |
+
"grad_norm": 0.4785331189632416,
|
| 959 |
+
"learning_rate": 5e-06,
|
| 960 |
+
"loss": 0.7273,
|
| 961 |
+
"step": 1360
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.8810289389067524,
|
| 965 |
+
"grad_norm": 0.4148114323616028,
|
| 966 |
+
"learning_rate": 5e-06,
|
| 967 |
+
"loss": 0.7247,
|
| 968 |
+
"step": 1370
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.887459807073955,
|
| 972 |
+
"grad_norm": 0.38368263840675354,
|
| 973 |
+
"learning_rate": 5e-06,
|
| 974 |
+
"loss": 0.7268,
|
| 975 |
+
"step": 1380
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.8938906752411575,
|
| 979 |
+
"grad_norm": 0.40352755784988403,
|
| 980 |
+
"learning_rate": 5e-06,
|
| 981 |
+
"loss": 0.7285,
|
| 982 |
+
"step": 1390
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.9003215434083601,
|
| 986 |
+
"grad_norm": 0.4232112467288971,
|
| 987 |
+
"learning_rate": 5e-06,
|
| 988 |
+
"loss": 0.7249,
|
| 989 |
+
"step": 1400
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.9067524115755627,
|
| 993 |
+
"grad_norm": 0.4366598129272461,
|
| 994 |
+
"learning_rate": 5e-06,
|
| 995 |
+
"loss": 0.7249,
|
| 996 |
+
"step": 1410
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.9131832797427653,
|
| 1000 |
+
"grad_norm": 0.4052753150463104,
|
| 1001 |
+
"learning_rate": 5e-06,
|
| 1002 |
+
"loss": 0.7245,
|
| 1003 |
+
"step": 1420
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.9196141479099679,
|
| 1007 |
+
"grad_norm": 0.44511350989341736,
|
| 1008 |
+
"learning_rate": 5e-06,
|
| 1009 |
+
"loss": 0.7305,
|
| 1010 |
+
"step": 1430
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.9260450160771704,
|
| 1014 |
+
"grad_norm": 0.44077491760253906,
|
| 1015 |
+
"learning_rate": 5e-06,
|
| 1016 |
+
"loss": 0.7294,
|
| 1017 |
+
"step": 1440
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.932475884244373,
|
| 1021 |
+
"grad_norm": 0.48145055770874023,
|
| 1022 |
+
"learning_rate": 5e-06,
|
| 1023 |
+
"loss": 0.7215,
|
| 1024 |
+
"step": 1450
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.9389067524115756,
|
| 1028 |
+
"grad_norm": 0.42547109723091125,
|
| 1029 |
+
"learning_rate": 5e-06,
|
| 1030 |
+
"loss": 0.7292,
|
| 1031 |
+
"step": 1460
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.9453376205787781,
|
| 1035 |
+
"grad_norm": 0.4287051558494568,
|
| 1036 |
+
"learning_rate": 5e-06,
|
| 1037 |
+
"loss": 0.7155,
|
| 1038 |
+
"step": 1470
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.9517684887459807,
|
| 1042 |
+
"grad_norm": 0.4230569005012512,
|
| 1043 |
+
"learning_rate": 5e-06,
|
| 1044 |
+
"loss": 0.7198,
|
| 1045 |
+
"step": 1480
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.9581993569131833,
|
| 1049 |
+
"grad_norm": 0.4719720184803009,
|
| 1050 |
+
"learning_rate": 5e-06,
|
| 1051 |
+
"loss": 0.7297,
|
| 1052 |
+
"step": 1490
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.9646302250803859,
|
| 1056 |
+
"grad_norm": 0.43969443440437317,
|
| 1057 |
+
"learning_rate": 5e-06,
|
| 1058 |
+
"loss": 0.7287,
|
| 1059 |
+
"step": 1500
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.9710610932475884,
|
| 1063 |
+
"grad_norm": 0.4330804646015167,
|
| 1064 |
+
"learning_rate": 5e-06,
|
| 1065 |
+
"loss": 0.7209,
|
| 1066 |
+
"step": 1510
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.977491961414791,
|
| 1070 |
+
"grad_norm": 0.38477885723114014,
|
| 1071 |
+
"learning_rate": 5e-06,
|
| 1072 |
+
"loss": 0.7246,
|
| 1073 |
+
"step": 1520
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.9839228295819936,
|
| 1077 |
+
"grad_norm": 0.46622517704963684,
|
| 1078 |
+
"learning_rate": 5e-06,
|
| 1079 |
+
"loss": 0.7281,
|
| 1080 |
+
"step": 1530
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.9903536977491961,
|
| 1084 |
+
"grad_norm": 0.47588425874710083,
|
| 1085 |
+
"learning_rate": 5e-06,
|
| 1086 |
+
"loss": 0.7264,
|
| 1087 |
+
"step": 1540
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.9967845659163987,
|
| 1091 |
+
"grad_norm": 0.43833327293395996,
|
| 1092 |
+
"learning_rate": 5e-06,
|
| 1093 |
+
"loss": 0.7345,
|
| 1094 |
+
"step": 1550
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 1.0,
|
| 1098 |
+
"eval_loss": 0.7232183814048767,
|
| 1099 |
+
"eval_runtime": 140.4382,
|
| 1100 |
+
"eval_samples_per_second": 298.373,
|
| 1101 |
+
"eval_steps_per_second": 1.168,
|
| 1102 |
+
"step": 1555
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 1.0032154340836013,
|
| 1106 |
+
"grad_norm": 0.6193146705627441,
|
| 1107 |
+
"learning_rate": 5e-06,
|
| 1108 |
+
"loss": 0.7137,
|
| 1109 |
+
"step": 1560
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.0096463022508038,
|
| 1113 |
+
"grad_norm": 0.4350009560585022,
|
| 1114 |
+
"learning_rate": 5e-06,
|
| 1115 |
+
"loss": 0.682,
|
| 1116 |
+
"step": 1570
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 1.0160771704180065,
|
| 1120 |
+
"grad_norm": 0.44338950514793396,
|
| 1121 |
+
"learning_rate": 5e-06,
|
| 1122 |
+
"loss": 0.6971,
|
| 1123 |
+
"step": 1580
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 1.022508038585209,
|
| 1127 |
+
"grad_norm": 0.4447615444660187,
|
| 1128 |
+
"learning_rate": 5e-06,
|
| 1129 |
+
"loss": 0.687,
|
| 1130 |
+
"step": 1590
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 1.0289389067524115,
|
| 1134 |
+
"grad_norm": 0.4388461112976074,
|
| 1135 |
+
"learning_rate": 5e-06,
|
| 1136 |
+
"loss": 0.6841,
|
| 1137 |
+
"step": 1600
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 1.0353697749196142,
|
| 1141 |
+
"grad_norm": 0.523465096950531,
|
| 1142 |
+
"learning_rate": 5e-06,
|
| 1143 |
+
"loss": 0.6922,
|
| 1144 |
+
"step": 1610
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 1.0418006430868167,
|
| 1148 |
+
"grad_norm": 0.46927884221076965,
|
| 1149 |
+
"learning_rate": 5e-06,
|
| 1150 |
+
"loss": 0.678,
|
| 1151 |
+
"step": 1620
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 1.0482315112540193,
|
| 1155 |
+
"grad_norm": 0.47517913579940796,
|
| 1156 |
+
"learning_rate": 5e-06,
|
| 1157 |
+
"loss": 0.6735,
|
| 1158 |
+
"step": 1630
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 1.0546623794212218,
|
| 1162 |
+
"grad_norm": 0.42782580852508545,
|
| 1163 |
+
"learning_rate": 5e-06,
|
| 1164 |
+
"loss": 0.6835,
|
| 1165 |
+
"step": 1640
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 1.0610932475884245,
|
| 1169 |
+
"grad_norm": 0.4523870348930359,
|
| 1170 |
+
"learning_rate": 5e-06,
|
| 1171 |
+
"loss": 0.6902,
|
| 1172 |
+
"step": 1650
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 1.067524115755627,
|
| 1176 |
+
"grad_norm": 0.46784067153930664,
|
| 1177 |
+
"learning_rate": 5e-06,
|
| 1178 |
+
"loss": 0.6814,
|
| 1179 |
+
"step": 1660
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 1.0739549839228295,
|
| 1183 |
+
"grad_norm": 0.43110141158103943,
|
| 1184 |
+
"learning_rate": 5e-06,
|
| 1185 |
+
"loss": 0.6793,
|
| 1186 |
+
"step": 1670
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 1.0803858520900322,
|
| 1190 |
+
"grad_norm": 0.44774794578552246,
|
| 1191 |
+
"learning_rate": 5e-06,
|
| 1192 |
+
"loss": 0.6853,
|
| 1193 |
+
"step": 1680
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 1.0868167202572347,
|
| 1197 |
+
"grad_norm": 0.5212559103965759,
|
| 1198 |
+
"learning_rate": 5e-06,
|
| 1199 |
+
"loss": 0.6863,
|
| 1200 |
+
"step": 1690
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 1.0932475884244373,
|
| 1204 |
+
"grad_norm": 0.49354031682014465,
|
| 1205 |
+
"learning_rate": 5e-06,
|
| 1206 |
+
"loss": 0.68,
|
| 1207 |
+
"step": 1700
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.09967845659164,
|
| 1211 |
+
"grad_norm": 0.4270998239517212,
|
| 1212 |
+
"learning_rate": 5e-06,
|
| 1213 |
+
"loss": 0.6845,
|
| 1214 |
+
"step": 1710
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 1.1061093247588425,
|
| 1218 |
+
"grad_norm": 0.43898463249206543,
|
| 1219 |
+
"learning_rate": 5e-06,
|
| 1220 |
+
"loss": 0.683,
|
| 1221 |
+
"step": 1720
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 1.112540192926045,
|
| 1225 |
+
"grad_norm": 0.413997083902359,
|
| 1226 |
+
"learning_rate": 5e-06,
|
| 1227 |
+
"loss": 0.6755,
|
| 1228 |
+
"step": 1730
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 1.1189710610932475,
|
| 1232 |
+
"grad_norm": 0.4593328833580017,
|
| 1233 |
+
"learning_rate": 5e-06,
|
| 1234 |
+
"loss": 0.6888,
|
| 1235 |
+
"step": 1740
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 1.1254019292604502,
|
| 1239 |
+
"grad_norm": 0.47585126757621765,
|
| 1240 |
+
"learning_rate": 5e-06,
|
| 1241 |
+
"loss": 0.6865,
|
| 1242 |
+
"step": 1750
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 1.1318327974276527,
|
| 1246 |
+
"grad_norm": 0.4210260808467865,
|
| 1247 |
+
"learning_rate": 5e-06,
|
| 1248 |
+
"loss": 0.6833,
|
| 1249 |
+
"step": 1760
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.1382636655948553,
|
| 1253 |
+
"grad_norm": 0.44741642475128174,
|
| 1254 |
+
"learning_rate": 5e-06,
|
| 1255 |
+
"loss": 0.6849,
|
| 1256 |
+
"step": 1770
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 1.144694533762058,
|
| 1260 |
+
"grad_norm": 0.437013179063797,
|
| 1261 |
+
"learning_rate": 5e-06,
|
| 1262 |
+
"loss": 0.6876,
|
| 1263 |
+
"step": 1780
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.1511254019292605,
|
| 1267 |
+
"grad_norm": 0.45807886123657227,
|
| 1268 |
+
"learning_rate": 5e-06,
|
| 1269 |
+
"loss": 0.6858,
|
| 1270 |
+
"step": 1790
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 1.157556270096463,
|
| 1274 |
+
"grad_norm": 0.46383801102638245,
|
| 1275 |
+
"learning_rate": 5e-06,
|
| 1276 |
+
"loss": 0.6718,
|
| 1277 |
+
"step": 1800
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 1.1639871382636655,
|
| 1281 |
+
"grad_norm": 0.5218130946159363,
|
| 1282 |
+
"learning_rate": 5e-06,
|
| 1283 |
+
"loss": 0.6831,
|
| 1284 |
+
"step": 1810
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 1.1704180064308682,
|
| 1288 |
+
"grad_norm": 0.4491232931613922,
|
| 1289 |
+
"learning_rate": 5e-06,
|
| 1290 |
+
"loss": 0.6837,
|
| 1291 |
+
"step": 1820
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 1.1768488745980707,
|
| 1295 |
+
"grad_norm": 0.45633968710899353,
|
| 1296 |
+
"learning_rate": 5e-06,
|
| 1297 |
+
"loss": 0.6782,
|
| 1298 |
+
"step": 1830
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 1.1832797427652733,
|
| 1302 |
+
"grad_norm": 0.4493134915828705,
|
| 1303 |
+
"learning_rate": 5e-06,
|
| 1304 |
+
"loss": 0.685,
|
| 1305 |
+
"step": 1840
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 1.189710610932476,
|
| 1309 |
+
"grad_norm": 0.42465460300445557,
|
| 1310 |
+
"learning_rate": 5e-06,
|
| 1311 |
+
"loss": 0.6878,
|
| 1312 |
+
"step": 1850
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 1.1961414790996785,
|
| 1316 |
+
"grad_norm": 0.4019971787929535,
|
| 1317 |
+
"learning_rate": 5e-06,
|
| 1318 |
+
"loss": 0.677,
|
| 1319 |
+
"step": 1860
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 1.202572347266881,
|
| 1323 |
+
"grad_norm": 0.45171162486076355,
|
| 1324 |
+
"learning_rate": 5e-06,
|
| 1325 |
+
"loss": 0.686,
|
| 1326 |
+
"step": 1870
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 1.2090032154340835,
|
| 1330 |
+
"grad_norm": 0.43534427881240845,
|
| 1331 |
+
"learning_rate": 5e-06,
|
| 1332 |
+
"loss": 0.6816,
|
| 1333 |
+
"step": 1880
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 1.2154340836012862,
|
| 1337 |
+
"grad_norm": 0.4472478926181793,
|
| 1338 |
+
"learning_rate": 5e-06,
|
| 1339 |
+
"loss": 0.6839,
|
| 1340 |
+
"step": 1890
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 1.2218649517684887,
|
| 1344 |
+
"grad_norm": 0.41748490929603577,
|
| 1345 |
+
"learning_rate": 5e-06,
|
| 1346 |
+
"loss": 0.6913,
|
| 1347 |
+
"step": 1900
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 1.2282958199356913,
|
| 1351 |
+
"grad_norm": 0.4147058427333832,
|
| 1352 |
+
"learning_rate": 5e-06,
|
| 1353 |
+
"loss": 0.681,
|
| 1354 |
+
"step": 1910
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 1.234726688102894,
|
| 1358 |
+
"grad_norm": 0.41148611903190613,
|
| 1359 |
+
"learning_rate": 5e-06,
|
| 1360 |
+
"loss": 0.6819,
|
| 1361 |
+
"step": 1920
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 1.2411575562700965,
|
| 1365 |
+
"grad_norm": 0.4393286406993866,
|
| 1366 |
+
"learning_rate": 5e-06,
|
| 1367 |
+
"loss": 0.6903,
|
| 1368 |
+
"step": 1930
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 1.247588424437299,
|
| 1372 |
+
"grad_norm": 0.4369371235370636,
|
| 1373 |
+
"learning_rate": 5e-06,
|
| 1374 |
+
"loss": 0.6872,
|
| 1375 |
+
"step": 1940
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 1.2540192926045015,
|
| 1379 |
+
"grad_norm": 0.5068768858909607,
|
| 1380 |
+
"learning_rate": 5e-06,
|
| 1381 |
+
"loss": 0.6869,
|
| 1382 |
+
"step": 1950
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 1.2604501607717042,
|
| 1386 |
+
"grad_norm": 0.43698254227638245,
|
| 1387 |
+
"learning_rate": 5e-06,
|
| 1388 |
+
"loss": 0.6853,
|
| 1389 |
+
"step": 1960
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 1.2668810289389068,
|
| 1393 |
+
"grad_norm": 0.439619779586792,
|
| 1394 |
+
"learning_rate": 5e-06,
|
| 1395 |
+
"loss": 0.681,
|
| 1396 |
+
"step": 1970
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 1.2733118971061093,
|
| 1400 |
+
"grad_norm": 0.4120563566684723,
|
| 1401 |
+
"learning_rate": 5e-06,
|
| 1402 |
+
"loss": 0.6815,
|
| 1403 |
+
"step": 1980
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 1.279742765273312,
|
| 1407 |
+
"grad_norm": 0.39503830671310425,
|
| 1408 |
+
"learning_rate": 5e-06,
|
| 1409 |
+
"loss": 0.6789,
|
| 1410 |
+
"step": 1990
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 1.2861736334405145,
|
| 1414 |
+
"grad_norm": 0.46114471554756165,
|
| 1415 |
+
"learning_rate": 5e-06,
|
| 1416 |
+
"loss": 0.6858,
|
| 1417 |
+
"step": 2000
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.292604501607717,
|
| 1421 |
+
"grad_norm": 0.44008398056030273,
|
| 1422 |
+
"learning_rate": 5e-06,
|
| 1423 |
+
"loss": 0.6916,
|
| 1424 |
+
"step": 2010
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 1.2990353697749195,
|
| 1428 |
+
"grad_norm": 0.41201072931289673,
|
| 1429 |
+
"learning_rate": 5e-06,
|
| 1430 |
+
"loss": 0.6832,
|
| 1431 |
+
"step": 2020
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 1.3054662379421222,
|
| 1435 |
+
"grad_norm": 0.42483213543891907,
|
| 1436 |
+
"learning_rate": 5e-06,
|
| 1437 |
+
"loss": 0.6876,
|
| 1438 |
+
"step": 2030
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 1.3118971061093248,
|
| 1442 |
+
"grad_norm": 0.4412234127521515,
|
| 1443 |
+
"learning_rate": 5e-06,
|
| 1444 |
+
"loss": 0.6736,
|
| 1445 |
+
"step": 2040
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 1.3183279742765273,
|
| 1449 |
+
"grad_norm": 0.42483067512512207,
|
| 1450 |
+
"learning_rate": 5e-06,
|
| 1451 |
+
"loss": 0.6866,
|
| 1452 |
+
"step": 2050
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 1.32475884244373,
|
| 1456 |
+
"grad_norm": 0.412567675113678,
|
| 1457 |
+
"learning_rate": 5e-06,
|
| 1458 |
+
"loss": 0.6877,
|
| 1459 |
+
"step": 2060
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 1.3311897106109325,
|
| 1463 |
+
"grad_norm": 0.4586575925350189,
|
| 1464 |
+
"learning_rate": 5e-06,
|
| 1465 |
+
"loss": 0.6898,
|
| 1466 |
+
"step": 2070
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 1.337620578778135,
|
| 1470 |
+
"grad_norm": 0.41622015833854675,
|
| 1471 |
+
"learning_rate": 5e-06,
|
| 1472 |
+
"loss": 0.6836,
|
| 1473 |
+
"step": 2080
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 1.3440514469453375,
|
| 1477 |
+
"grad_norm": 0.4891818165779114,
|
| 1478 |
+
"learning_rate": 5e-06,
|
| 1479 |
+
"loss": 0.6825,
|
| 1480 |
+
"step": 2090
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 1.3504823151125402,
|
| 1484 |
+
"grad_norm": 0.42719897627830505,
|
| 1485 |
+
"learning_rate": 5e-06,
|
| 1486 |
+
"loss": 0.6915,
|
| 1487 |
+
"step": 2100
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 1.3569131832797428,
|
| 1491 |
+
"grad_norm": 0.4267047643661499,
|
| 1492 |
+
"learning_rate": 5e-06,
|
| 1493 |
+
"loss": 0.6829,
|
| 1494 |
+
"step": 2110
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 1.3633440514469453,
|
| 1498 |
+
"grad_norm": 0.48593926429748535,
|
| 1499 |
+
"learning_rate": 5e-06,
|
| 1500 |
+
"loss": 0.6865,
|
| 1501 |
+
"step": 2120
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.369774919614148,
|
| 1505 |
+
"grad_norm": 0.46337270736694336,
|
| 1506 |
+
"learning_rate": 5e-06,
|
| 1507 |
+
"loss": 0.6784,
|
| 1508 |
+
"step": 2130
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 1.3762057877813505,
|
| 1512 |
+
"grad_norm": 0.42621052265167236,
|
| 1513 |
+
"learning_rate": 5e-06,
|
| 1514 |
+
"loss": 0.6815,
|
| 1515 |
+
"step": 2140
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 1.382636655948553,
|
| 1519 |
+
"grad_norm": 0.4374221861362457,
|
| 1520 |
+
"learning_rate": 5e-06,
|
| 1521 |
+
"loss": 0.6823,
|
| 1522 |
+
"step": 2150
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 1.3890675241157555,
|
| 1526 |
+
"grad_norm": 0.49710190296173096,
|
| 1527 |
+
"learning_rate": 5e-06,
|
| 1528 |
+
"loss": 0.6788,
|
| 1529 |
+
"step": 2160
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 1.3954983922829582,
|
| 1533 |
+
"grad_norm": 0.43773987889289856,
|
| 1534 |
+
"learning_rate": 5e-06,
|
| 1535 |
+
"loss": 0.6897,
|
| 1536 |
+
"step": 2170
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 1.4019292604501608,
|
| 1540 |
+
"grad_norm": 0.420480877161026,
|
| 1541 |
+
"learning_rate": 5e-06,
|
| 1542 |
+
"loss": 0.6877,
|
| 1543 |
+
"step": 2180
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 1.4083601286173635,
|
| 1547 |
+
"grad_norm": 0.41698333621025085,
|
| 1548 |
+
"learning_rate": 5e-06,
|
| 1549 |
+
"loss": 0.6782,
|
| 1550 |
+
"step": 2190
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 1.414790996784566,
|
| 1554 |
+
"grad_norm": 0.45082971453666687,
|
| 1555 |
+
"learning_rate": 5e-06,
|
| 1556 |
+
"loss": 0.6884,
|
| 1557 |
+
"step": 2200
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 1.4212218649517685,
|
| 1561 |
+
"grad_norm": 0.44047510623931885,
|
| 1562 |
+
"learning_rate": 5e-06,
|
| 1563 |
+
"loss": 0.68,
|
| 1564 |
+
"step": 2210
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 1.427652733118971,
|
| 1568 |
+
"grad_norm": 0.43505367636680603,
|
| 1569 |
+
"learning_rate": 5e-06,
|
| 1570 |
+
"loss": 0.684,
|
| 1571 |
+
"step": 2220
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.4340836012861735,
|
| 1575 |
+
"grad_norm": 0.42423659563064575,
|
| 1576 |
+
"learning_rate": 5e-06,
|
| 1577 |
+
"loss": 0.6852,
|
| 1578 |
+
"step": 2230
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 1.4405144694533762,
|
| 1582 |
+
"grad_norm": 0.5248254537582397,
|
| 1583 |
+
"learning_rate": 5e-06,
|
| 1584 |
+
"loss": 0.6796,
|
| 1585 |
+
"step": 2240
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.4469453376205788,
|
| 1589 |
+
"grad_norm": 0.47595280408859253,
|
| 1590 |
+
"learning_rate": 5e-06,
|
| 1591 |
+
"loss": 0.6812,
|
| 1592 |
+
"step": 2250
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 1.4533762057877815,
|
| 1596 |
+
"grad_norm": 0.41155120730400085,
|
| 1597 |
+
"learning_rate": 5e-06,
|
| 1598 |
+
"loss": 0.6894,
|
| 1599 |
+
"step": 2260
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 1.459807073954984,
|
| 1603 |
+
"grad_norm": 0.4290079176425934,
|
| 1604 |
+
"learning_rate": 5e-06,
|
| 1605 |
+
"loss": 0.684,
|
| 1606 |
+
"step": 2270
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 1.4662379421221865,
|
| 1610 |
+
"grad_norm": 0.4066310226917267,
|
| 1611 |
+
"learning_rate": 5e-06,
|
| 1612 |
+
"loss": 0.6823,
|
| 1613 |
+
"step": 2280
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 1.472668810289389,
|
| 1617 |
+
"grad_norm": 0.4360986351966858,
|
| 1618 |
+
"learning_rate": 5e-06,
|
| 1619 |
+
"loss": 0.6886,
|
| 1620 |
+
"step": 2290
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 1.4790996784565915,
|
| 1624 |
+
"grad_norm": 0.3978036642074585,
|
| 1625 |
+
"learning_rate": 5e-06,
|
| 1626 |
+
"loss": 0.6849,
|
| 1627 |
+
"step": 2300
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 1.4855305466237942,
|
| 1631 |
+
"grad_norm": 0.41997137665748596,
|
| 1632 |
+
"learning_rate": 5e-06,
|
| 1633 |
+
"loss": 0.6834,
|
| 1634 |
+
"step": 2310
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 1.4919614147909968,
|
| 1638 |
+
"grad_norm": 0.4409499764442444,
|
| 1639 |
+
"learning_rate": 5e-06,
|
| 1640 |
+
"loss": 0.6873,
|
| 1641 |
+
"step": 2320
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 1.4983922829581995,
|
| 1645 |
+
"grad_norm": 0.44297707080841064,
|
| 1646 |
+
"learning_rate": 5e-06,
|
| 1647 |
+
"loss": 0.682,
|
| 1648 |
+
"step": 2330
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 1.504823151125402,
|
| 1652 |
+
"grad_norm": 0.4389215409755707,
|
| 1653 |
+
"learning_rate": 5e-06,
|
| 1654 |
+
"loss": 0.6892,
|
| 1655 |
+
"step": 2340
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 1.5112540192926045,
|
| 1659 |
+
"grad_norm": 0.44480717182159424,
|
| 1660 |
+
"learning_rate": 5e-06,
|
| 1661 |
+
"loss": 0.6884,
|
| 1662 |
+
"step": 2350
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 1.517684887459807,
|
| 1666 |
+
"grad_norm": 0.44355508685112,
|
| 1667 |
+
"learning_rate": 5e-06,
|
| 1668 |
+
"loss": 0.6876,
|
| 1669 |
+
"step": 2360
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 1.5241157556270095,
|
| 1673 |
+
"grad_norm": 0.46087756752967834,
|
| 1674 |
+
"learning_rate": 5e-06,
|
| 1675 |
+
"loss": 0.6828,
|
| 1676 |
+
"step": 2370
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 1.5305466237942122,
|
| 1680 |
+
"grad_norm": 0.4672592580318451,
|
| 1681 |
+
"learning_rate": 5e-06,
|
| 1682 |
+
"loss": 0.6871,
|
| 1683 |
+
"step": 2380
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 1.5369774919614148,
|
| 1687 |
+
"grad_norm": 0.3982642889022827,
|
| 1688 |
+
"learning_rate": 5e-06,
|
| 1689 |
+
"loss": 0.6819,
|
| 1690 |
+
"step": 2390
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 1.5434083601286175,
|
| 1694 |
+
"grad_norm": 0.40598487854003906,
|
| 1695 |
+
"learning_rate": 5e-06,
|
| 1696 |
+
"loss": 0.6838,
|
| 1697 |
+
"step": 2400
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 1.54983922829582,
|
| 1701 |
+
"grad_norm": 0.4345811605453491,
|
| 1702 |
+
"learning_rate": 5e-06,
|
| 1703 |
+
"loss": 0.6815,
|
| 1704 |
+
"step": 2410
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 1.5562700964630225,
|
| 1708 |
+
"grad_norm": 0.43568089604377747,
|
| 1709 |
+
"learning_rate": 5e-06,
|
| 1710 |
+
"loss": 0.6911,
|
| 1711 |
+
"step": 2420
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 1.562700964630225,
|
| 1715 |
+
"grad_norm": 0.43922844529151917,
|
| 1716 |
+
"learning_rate": 5e-06,
|
| 1717 |
+
"loss": 0.6834,
|
| 1718 |
+
"step": 2430
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 1.5691318327974275,
|
| 1722 |
+
"grad_norm": 0.4793626368045807,
|
| 1723 |
+
"learning_rate": 5e-06,
|
| 1724 |
+
"loss": 0.6833,
|
| 1725 |
+
"step": 2440
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 1.5755627009646302,
|
| 1729 |
+
"grad_norm": 0.45686307549476624,
|
| 1730 |
+
"learning_rate": 5e-06,
|
| 1731 |
+
"loss": 0.6864,
|
| 1732 |
+
"step": 2450
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 1.5819935691318328,
|
| 1736 |
+
"grad_norm": 0.41061604022979736,
|
| 1737 |
+
"learning_rate": 5e-06,
|
| 1738 |
+
"loss": 0.6768,
|
| 1739 |
+
"step": 2460
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 1.5884244372990355,
|
| 1743 |
+
"grad_norm": 0.45562943816185,
|
| 1744 |
+
"learning_rate": 5e-06,
|
| 1745 |
+
"loss": 0.6743,
|
| 1746 |
+
"step": 2470
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 1.594855305466238,
|
| 1750 |
+
"grad_norm": 0.4296809136867523,
|
| 1751 |
+
"learning_rate": 5e-06,
|
| 1752 |
+
"loss": 0.6802,
|
| 1753 |
+
"step": 2480
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 1.6012861736334405,
|
| 1757 |
+
"grad_norm": 0.4218539595603943,
|
| 1758 |
+
"learning_rate": 5e-06,
|
| 1759 |
+
"loss": 0.6857,
|
| 1760 |
+
"step": 2490
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 1.607717041800643,
|
| 1764 |
+
"grad_norm": 0.45991575717926025,
|
| 1765 |
+
"learning_rate": 5e-06,
|
| 1766 |
+
"loss": 0.684,
|
| 1767 |
+
"step": 2500
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 1.6141479099678455,
|
| 1771 |
+
"grad_norm": 0.43671083450317383,
|
| 1772 |
+
"learning_rate": 5e-06,
|
| 1773 |
+
"loss": 0.6872,
|
| 1774 |
+
"step": 2510
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 1.6205787781350482,
|
| 1778 |
+
"grad_norm": 0.4566083550453186,
|
| 1779 |
+
"learning_rate": 5e-06,
|
| 1780 |
+
"loss": 0.6808,
|
| 1781 |
+
"step": 2520
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 1.6270096463022508,
|
| 1785 |
+
"grad_norm": 0.4394720792770386,
|
| 1786 |
+
"learning_rate": 5e-06,
|
| 1787 |
+
"loss": 0.6841,
|
| 1788 |
+
"step": 2530
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 1.6334405144694535,
|
| 1792 |
+
"grad_norm": 0.4831625521183014,
|
| 1793 |
+
"learning_rate": 5e-06,
|
| 1794 |
+
"loss": 0.6846,
|
| 1795 |
+
"step": 2540
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 1.639871382636656,
|
| 1799 |
+
"grad_norm": 0.4193684458732605,
|
| 1800 |
+
"learning_rate": 5e-06,
|
| 1801 |
+
"loss": 0.6858,
|
| 1802 |
+
"step": 2550
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 1.6463022508038585,
|
| 1806 |
+
"grad_norm": 0.47811901569366455,
|
| 1807 |
+
"learning_rate": 5e-06,
|
| 1808 |
+
"loss": 0.6758,
|
| 1809 |
+
"step": 2560
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 1.652733118971061,
|
| 1813 |
+
"grad_norm": 0.4748084545135498,
|
| 1814 |
+
"learning_rate": 5e-06,
|
| 1815 |
+
"loss": 0.6861,
|
| 1816 |
+
"step": 2570
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 1.6591639871382635,
|
| 1820 |
+
"grad_norm": 0.42245808243751526,
|
| 1821 |
+
"learning_rate": 5e-06,
|
| 1822 |
+
"loss": 0.6745,
|
| 1823 |
+
"step": 2580
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 1.6655948553054662,
|
| 1827 |
+
"grad_norm": 0.4233896732330322,
|
| 1828 |
+
"learning_rate": 5e-06,
|
| 1829 |
+
"loss": 0.6862,
|
| 1830 |
+
"step": 2590
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 1.6720257234726688,
|
| 1834 |
+
"grad_norm": 0.4050581455230713,
|
| 1835 |
+
"learning_rate": 5e-06,
|
| 1836 |
+
"loss": 0.676,
|
| 1837 |
+
"step": 2600
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 1.6784565916398715,
|
| 1841 |
+
"grad_norm": 0.41761529445648193,
|
| 1842 |
+
"learning_rate": 5e-06,
|
| 1843 |
+
"loss": 0.6802,
|
| 1844 |
+
"step": 2610
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 1.684887459807074,
|
| 1848 |
+
"grad_norm": 0.4060108959674835,
|
| 1849 |
+
"learning_rate": 5e-06,
|
| 1850 |
+
"loss": 0.684,
|
| 1851 |
+
"step": 2620
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 1.6913183279742765,
|
| 1855 |
+
"grad_norm": 0.40532317757606506,
|
| 1856 |
+
"learning_rate": 5e-06,
|
| 1857 |
+
"loss": 0.68,
|
| 1858 |
+
"step": 2630
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 1.697749196141479,
|
| 1862 |
+
"grad_norm": 0.47850868105888367,
|
| 1863 |
+
"learning_rate": 5e-06,
|
| 1864 |
+
"loss": 0.6879,
|
| 1865 |
+
"step": 2640
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 1.7041800643086815,
|
| 1869 |
+
"grad_norm": 0.4595736861228943,
|
| 1870 |
+
"learning_rate": 5e-06,
|
| 1871 |
+
"loss": 0.6861,
|
| 1872 |
+
"step": 2650
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 1.7106109324758842,
|
| 1876 |
+
"grad_norm": 0.3999944031238556,
|
| 1877 |
+
"learning_rate": 5e-06,
|
| 1878 |
+
"loss": 0.6811,
|
| 1879 |
+
"step": 2660
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 1.717041800643087,
|
| 1883 |
+
"grad_norm": 0.4152866303920746,
|
| 1884 |
+
"learning_rate": 5e-06,
|
| 1885 |
+
"loss": 0.676,
|
| 1886 |
+
"step": 2670
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 1.7234726688102895,
|
| 1890 |
+
"grad_norm": 0.5088204741477966,
|
| 1891 |
+
"learning_rate": 5e-06,
|
| 1892 |
+
"loss": 0.6868,
|
| 1893 |
+
"step": 2680
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 1.729903536977492,
|
| 1897 |
+
"grad_norm": 0.4383102059364319,
|
| 1898 |
+
"learning_rate": 5e-06,
|
| 1899 |
+
"loss": 0.6818,
|
| 1900 |
+
"step": 2690
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 1.7363344051446945,
|
| 1904 |
+
"grad_norm": 0.4219019114971161,
|
| 1905 |
+
"learning_rate": 5e-06,
|
| 1906 |
+
"loss": 0.6767,
|
| 1907 |
+
"step": 2700
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 1.742765273311897,
|
| 1911 |
+
"grad_norm": 0.4663357138633728,
|
| 1912 |
+
"learning_rate": 5e-06,
|
| 1913 |
+
"loss": 0.6783,
|
| 1914 |
+
"step": 2710
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 1.7491961414790995,
|
| 1918 |
+
"grad_norm": 0.43959125876426697,
|
| 1919 |
+
"learning_rate": 5e-06,
|
| 1920 |
+
"loss": 0.6824,
|
| 1921 |
+
"step": 2720
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 1.7556270096463023,
|
| 1925 |
+
"grad_norm": 0.43940484523773193,
|
| 1926 |
+
"learning_rate": 5e-06,
|
| 1927 |
+
"loss": 0.6769,
|
| 1928 |
+
"step": 2730
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 1.762057877813505,
|
| 1932 |
+
"grad_norm": 0.4071805477142334,
|
| 1933 |
+
"learning_rate": 5e-06,
|
| 1934 |
+
"loss": 0.6893,
|
| 1935 |
+
"step": 2740
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 1.7684887459807075,
|
| 1939 |
+
"grad_norm": 0.40912047028541565,
|
| 1940 |
+
"learning_rate": 5e-06,
|
| 1941 |
+
"loss": 0.6877,
|
| 1942 |
+
"step": 2750
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 1.77491961414791,
|
| 1946 |
+
"grad_norm": 0.4339078366756439,
|
| 1947 |
+
"learning_rate": 5e-06,
|
| 1948 |
+
"loss": 0.6788,
|
| 1949 |
+
"step": 2760
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 1.7813504823151125,
|
| 1953 |
+
"grad_norm": 0.4326707720756531,
|
| 1954 |
+
"learning_rate": 5e-06,
|
| 1955 |
+
"loss": 0.6831,
|
| 1956 |
+
"step": 2770
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 1.787781350482315,
|
| 1960 |
+
"grad_norm": 0.4307939410209656,
|
| 1961 |
+
"learning_rate": 5e-06,
|
| 1962 |
+
"loss": 0.6878,
|
| 1963 |
+
"step": 2780
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 1.7942122186495175,
|
| 1967 |
+
"grad_norm": 0.3918740749359131,
|
| 1968 |
+
"learning_rate": 5e-06,
|
| 1969 |
+
"loss": 0.6912,
|
| 1970 |
+
"step": 2790
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 1.8006430868167203,
|
| 1974 |
+
"grad_norm": 0.40045326948165894,
|
| 1975 |
+
"learning_rate": 5e-06,
|
| 1976 |
+
"loss": 0.6852,
|
| 1977 |
+
"step": 2800
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 1.807073954983923,
|
| 1981 |
+
"grad_norm": 0.40693676471710205,
|
| 1982 |
+
"learning_rate": 5e-06,
|
| 1983 |
+
"loss": 0.6842,
|
| 1984 |
+
"step": 2810
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 1.8135048231511255,
|
| 1988 |
+
"grad_norm": 0.3957662284374237,
|
| 1989 |
+
"learning_rate": 5e-06,
|
| 1990 |
+
"loss": 0.6777,
|
| 1991 |
+
"step": 2820
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 1.819935691318328,
|
| 1995 |
+
"grad_norm": 0.4147331714630127,
|
| 1996 |
+
"learning_rate": 5e-06,
|
| 1997 |
+
"loss": 0.679,
|
| 1998 |
+
"step": 2830
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 1.8263665594855305,
|
| 2002 |
+
"grad_norm": 0.421351820230484,
|
| 2003 |
+
"learning_rate": 5e-06,
|
| 2004 |
+
"loss": 0.6872,
|
| 2005 |
+
"step": 2840
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 1.832797427652733,
|
| 2009 |
+
"grad_norm": 0.4355145990848541,
|
| 2010 |
+
"learning_rate": 5e-06,
|
| 2011 |
+
"loss": 0.6847,
|
| 2012 |
+
"step": 2850
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 1.8392282958199357,
|
| 2016 |
+
"grad_norm": 0.41382715106010437,
|
| 2017 |
+
"learning_rate": 5e-06,
|
| 2018 |
+
"loss": 0.6777,
|
| 2019 |
+
"step": 2860
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 1.8456591639871383,
|
| 2023 |
+
"grad_norm": 0.4280456006526947,
|
| 2024 |
+
"learning_rate": 5e-06,
|
| 2025 |
+
"loss": 0.6781,
|
| 2026 |
+
"step": 2870
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 1.852090032154341,
|
| 2030 |
+
"grad_norm": 0.49057596921920776,
|
| 2031 |
+
"learning_rate": 5e-06,
|
| 2032 |
+
"loss": 0.6844,
|
| 2033 |
+
"step": 2880
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 1.8585209003215435,
|
| 2037 |
+
"grad_norm": 0.41203203797340393,
|
| 2038 |
+
"learning_rate": 5e-06,
|
| 2039 |
+
"loss": 0.6858,
|
| 2040 |
+
"step": 2890
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 1.864951768488746,
|
| 2044 |
+
"grad_norm": 0.4137505888938904,
|
| 2045 |
+
"learning_rate": 5e-06,
|
| 2046 |
+
"loss": 0.6808,
|
| 2047 |
+
"step": 2900
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 1.8713826366559485,
|
| 2051 |
+
"grad_norm": 0.46201327443122864,
|
| 2052 |
+
"learning_rate": 5e-06,
|
| 2053 |
+
"loss": 0.6798,
|
| 2054 |
+
"step": 2910
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 1.877813504823151,
|
| 2058 |
+
"grad_norm": 0.4486853778362274,
|
| 2059 |
+
"learning_rate": 5e-06,
|
| 2060 |
+
"loss": 0.6862,
|
| 2061 |
+
"step": 2920
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 1.8842443729903537,
|
| 2065 |
+
"grad_norm": 0.4082019329071045,
|
| 2066 |
+
"learning_rate": 5e-06,
|
| 2067 |
+
"loss": 0.6843,
|
| 2068 |
+
"step": 2930
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 1.8906752411575563,
|
| 2072 |
+
"grad_norm": 0.48371437191963196,
|
| 2073 |
+
"learning_rate": 5e-06,
|
| 2074 |
+
"loss": 0.6781,
|
| 2075 |
+
"step": 2940
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 1.897106109324759,
|
| 2079 |
+
"grad_norm": 0.4703909456729889,
|
| 2080 |
+
"learning_rate": 5e-06,
|
| 2081 |
+
"loss": 0.6843,
|
| 2082 |
+
"step": 2950
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 1.9035369774919615,
|
| 2086 |
+
"grad_norm": 0.40100324153900146,
|
| 2087 |
+
"learning_rate": 5e-06,
|
| 2088 |
+
"loss": 0.6838,
|
| 2089 |
+
"step": 2960
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"epoch": 1.909967845659164,
|
| 2093 |
+
"grad_norm": 0.4890632629394531,
|
| 2094 |
+
"learning_rate": 5e-06,
|
| 2095 |
+
"loss": 0.6771,
|
| 2096 |
+
"step": 2970
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 1.9163987138263665,
|
| 2100 |
+
"grad_norm": 0.4435786008834839,
|
| 2101 |
+
"learning_rate": 5e-06,
|
| 2102 |
+
"loss": 0.6812,
|
| 2103 |
+
"step": 2980
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 1.922829581993569,
|
| 2107 |
+
"grad_norm": 0.43092137575149536,
|
| 2108 |
+
"learning_rate": 5e-06,
|
| 2109 |
+
"loss": 0.6723,
|
| 2110 |
+
"step": 2990
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 1.9292604501607717,
|
| 2114 |
+
"grad_norm": 0.45887503027915955,
|
| 2115 |
+
"learning_rate": 5e-06,
|
| 2116 |
+
"loss": 0.6811,
|
| 2117 |
+
"step": 3000
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 1.9356913183279743,
|
| 2121 |
+
"grad_norm": 0.4193710386753082,
|
| 2122 |
+
"learning_rate": 5e-06,
|
| 2123 |
+
"loss": 0.6801,
|
| 2124 |
+
"step": 3010
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 1.942122186495177,
|
| 2128 |
+
"grad_norm": 0.45064231753349304,
|
| 2129 |
+
"learning_rate": 5e-06,
|
| 2130 |
+
"loss": 0.6718,
|
| 2131 |
+
"step": 3020
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 1.9485530546623795,
|
| 2135 |
+
"grad_norm": 0.42754465341567993,
|
| 2136 |
+
"learning_rate": 5e-06,
|
| 2137 |
+
"loss": 0.6736,
|
| 2138 |
+
"step": 3030
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 1.954983922829582,
|
| 2142 |
+
"grad_norm": 0.41854625940322876,
|
| 2143 |
+
"learning_rate": 5e-06,
|
| 2144 |
+
"loss": 0.6818,
|
| 2145 |
+
"step": 3040
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 1.9614147909967845,
|
| 2149 |
+
"grad_norm": 0.4096625745296478,
|
| 2150 |
+
"learning_rate": 5e-06,
|
| 2151 |
+
"loss": 0.6787,
|
| 2152 |
+
"step": 3050
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 1.967845659163987,
|
| 2156 |
+
"grad_norm": 0.41215813159942627,
|
| 2157 |
+
"learning_rate": 5e-06,
|
| 2158 |
+
"loss": 0.6752,
|
| 2159 |
+
"step": 3060
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 1.9742765273311897,
|
| 2163 |
+
"grad_norm": 0.47351738810539246,
|
| 2164 |
+
"learning_rate": 5e-06,
|
| 2165 |
+
"loss": 0.6835,
|
| 2166 |
+
"step": 3070
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 1.9807073954983923,
|
| 2170 |
+
"grad_norm": 0.4130474925041199,
|
| 2171 |
+
"learning_rate": 5e-06,
|
| 2172 |
+
"loss": 0.6803,
|
| 2173 |
+
"step": 3080
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 1.987138263665595,
|
| 2177 |
+
"grad_norm": 0.41312044858932495,
|
| 2178 |
+
"learning_rate": 5e-06,
|
| 2179 |
+
"loss": 0.6729,
|
| 2180 |
+
"step": 3090
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 1.9935691318327975,
|
| 2184 |
+
"grad_norm": 0.40591350197792053,
|
| 2185 |
+
"learning_rate": 5e-06,
|
| 2186 |
+
"loss": 0.6722,
|
| 2187 |
+
"step": 3100
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 2.0,
|
| 2191 |
+
"grad_norm": 0.40826907753944397,
|
| 2192 |
+
"learning_rate": 5e-06,
|
| 2193 |
+
"loss": 0.686,
|
| 2194 |
+
"step": 3110
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 2.0,
|
| 2198 |
+
"eval_loss": 0.7126035690307617,
|
| 2199 |
+
"eval_runtime": 139.886,
|
| 2200 |
+
"eval_samples_per_second": 299.551,
|
| 2201 |
+
"eval_steps_per_second": 1.172,
|
| 2202 |
+
"step": 3110
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 2.0064308681672025,
|
| 2206 |
+
"grad_norm": 0.561937153339386,
|
| 2207 |
+
"learning_rate": 5e-06,
|
| 2208 |
+
"loss": 0.6404,
|
| 2209 |
+
"step": 3120
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 2.012861736334405,
|
| 2213 |
+
"grad_norm": 0.5182865858078003,
|
| 2214 |
+
"learning_rate": 5e-06,
|
| 2215 |
+
"loss": 0.64,
|
| 2216 |
+
"step": 3130
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 2.0192926045016075,
|
| 2220 |
+
"grad_norm": 0.5018469095230103,
|
| 2221 |
+
"learning_rate": 5e-06,
|
| 2222 |
+
"loss": 0.6359,
|
| 2223 |
+
"step": 3140
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 2.0257234726688105,
|
| 2227 |
+
"grad_norm": 0.4944702982902527,
|
| 2228 |
+
"learning_rate": 5e-06,
|
| 2229 |
+
"loss": 0.6381,
|
| 2230 |
+
"step": 3150
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 2.032154340836013,
|
| 2234 |
+
"grad_norm": 0.4453904330730438,
|
| 2235 |
+
"learning_rate": 5e-06,
|
| 2236 |
+
"loss": 0.6372,
|
| 2237 |
+
"step": 3160
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 2.0385852090032155,
|
| 2241 |
+
"grad_norm": 0.43754857778549194,
|
| 2242 |
+
"learning_rate": 5e-06,
|
| 2243 |
+
"loss": 0.6394,
|
| 2244 |
+
"step": 3170
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 2.045016077170418,
|
| 2248 |
+
"grad_norm": 0.43891504406929016,
|
| 2249 |
+
"learning_rate": 5e-06,
|
| 2250 |
+
"loss": 0.6429,
|
| 2251 |
+
"step": 3180
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 2.0514469453376205,
|
| 2255 |
+
"grad_norm": 0.42905229330062866,
|
| 2256 |
+
"learning_rate": 5e-06,
|
| 2257 |
+
"loss": 0.6352,
|
| 2258 |
+
"step": 3190
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 2.057877813504823,
|
| 2262 |
+
"grad_norm": 0.4976702928543091,
|
| 2263 |
+
"learning_rate": 5e-06,
|
| 2264 |
+
"loss": 0.6394,
|
| 2265 |
+
"step": 3200
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 2.0643086816720255,
|
| 2269 |
+
"grad_norm": 0.4487721621990204,
|
| 2270 |
+
"learning_rate": 5e-06,
|
| 2271 |
+
"loss": 0.6367,
|
| 2272 |
+
"step": 3210
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 2.0707395498392285,
|
| 2276 |
+
"grad_norm": 0.48409563302993774,
|
| 2277 |
+
"learning_rate": 5e-06,
|
| 2278 |
+
"loss": 0.6364,
|
| 2279 |
+
"step": 3220
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 2.077170418006431,
|
| 2283 |
+
"grad_norm": 0.4372307360172272,
|
| 2284 |
+
"learning_rate": 5e-06,
|
| 2285 |
+
"loss": 0.6441,
|
| 2286 |
+
"step": 3230
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 2.0836012861736335,
|
| 2290 |
+
"grad_norm": 0.46748873591423035,
|
| 2291 |
+
"learning_rate": 5e-06,
|
| 2292 |
+
"loss": 0.6364,
|
| 2293 |
+
"step": 3240
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 2.090032154340836,
|
| 2297 |
+
"grad_norm": 0.46158337593078613,
|
| 2298 |
+
"learning_rate": 5e-06,
|
| 2299 |
+
"loss": 0.6334,
|
| 2300 |
+
"step": 3250
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 2.0964630225080385,
|
| 2304 |
+
"grad_norm": 0.43315762281417847,
|
| 2305 |
+
"learning_rate": 5e-06,
|
| 2306 |
+
"loss": 0.6395,
|
| 2307 |
+
"step": 3260
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 2.102893890675241,
|
| 2311 |
+
"grad_norm": 0.4499836564064026,
|
| 2312 |
+
"learning_rate": 5e-06,
|
| 2313 |
+
"loss": 0.6444,
|
| 2314 |
+
"step": 3270
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 2.1093247588424435,
|
| 2318 |
+
"grad_norm": 0.45043113827705383,
|
| 2319 |
+
"learning_rate": 5e-06,
|
| 2320 |
+
"loss": 0.6384,
|
| 2321 |
+
"step": 3280
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 2.1157556270096465,
|
| 2325 |
+
"grad_norm": 0.4223954677581787,
|
| 2326 |
+
"learning_rate": 5e-06,
|
| 2327 |
+
"loss": 0.638,
|
| 2328 |
+
"step": 3290
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 2.122186495176849,
|
| 2332 |
+
"grad_norm": 0.45464426279067993,
|
| 2333 |
+
"learning_rate": 5e-06,
|
| 2334 |
+
"loss": 0.6466,
|
| 2335 |
+
"step": 3300
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 2.1286173633440515,
|
| 2339 |
+
"grad_norm": 0.4449687898159027,
|
| 2340 |
+
"learning_rate": 5e-06,
|
| 2341 |
+
"loss": 0.6297,
|
| 2342 |
+
"step": 3310
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 2.135048231511254,
|
| 2346 |
+
"grad_norm": 0.4286421537399292,
|
| 2347 |
+
"learning_rate": 5e-06,
|
| 2348 |
+
"loss": 0.6408,
|
| 2349 |
+
"step": 3320
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 2.1414790996784565,
|
| 2353 |
+
"grad_norm": 0.4531044065952301,
|
| 2354 |
+
"learning_rate": 5e-06,
|
| 2355 |
+
"loss": 0.6416,
|
| 2356 |
+
"step": 3330
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 2.147909967845659,
|
| 2360 |
+
"grad_norm": 0.44140374660491943,
|
| 2361 |
+
"learning_rate": 5e-06,
|
| 2362 |
+
"loss": 0.636,
|
| 2363 |
+
"step": 3340
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 2.154340836012862,
|
| 2367 |
+
"grad_norm": 0.4317072927951813,
|
| 2368 |
+
"learning_rate": 5e-06,
|
| 2369 |
+
"loss": 0.6395,
|
| 2370 |
+
"step": 3350
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 2.1607717041800645,
|
| 2374 |
+
"grad_norm": 0.47132372856140137,
|
| 2375 |
+
"learning_rate": 5e-06,
|
| 2376 |
+
"loss": 0.6233,
|
| 2377 |
+
"step": 3360
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 2.167202572347267,
|
| 2381 |
+
"grad_norm": 0.4424266815185547,
|
| 2382 |
+
"learning_rate": 5e-06,
|
| 2383 |
+
"loss": 0.6417,
|
| 2384 |
+
"step": 3370
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 2.1736334405144695,
|
| 2388 |
+
"grad_norm": 0.42858994007110596,
|
| 2389 |
+
"learning_rate": 5e-06,
|
| 2390 |
+
"loss": 0.6418,
|
| 2391 |
+
"step": 3380
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 2.180064308681672,
|
| 2395 |
+
"grad_norm": 0.4775191843509674,
|
| 2396 |
+
"learning_rate": 5e-06,
|
| 2397 |
+
"loss": 0.6369,
|
| 2398 |
+
"step": 3390
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 2.1864951768488745,
|
| 2402 |
+
"grad_norm": 0.46210339665412903,
|
| 2403 |
+
"learning_rate": 5e-06,
|
| 2404 |
+
"loss": 0.6427,
|
| 2405 |
+
"step": 3400
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 2.192926045016077,
|
| 2409 |
+
"grad_norm": 0.46885842084884644,
|
| 2410 |
+
"learning_rate": 5e-06,
|
| 2411 |
+
"loss": 0.6456,
|
| 2412 |
+
"step": 3410
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 2.19935691318328,
|
| 2416 |
+
"grad_norm": 0.4760117828845978,
|
| 2417 |
+
"learning_rate": 5e-06,
|
| 2418 |
+
"loss": 0.6435,
|
| 2419 |
+
"step": 3420
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 2.2057877813504825,
|
| 2423 |
+
"grad_norm": 0.5036518573760986,
|
| 2424 |
+
"learning_rate": 5e-06,
|
| 2425 |
+
"loss": 0.6383,
|
| 2426 |
+
"step": 3430
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 2.212218649517685,
|
| 2430 |
+
"grad_norm": 0.4869937002658844,
|
| 2431 |
+
"learning_rate": 5e-06,
|
| 2432 |
+
"loss": 0.634,
|
| 2433 |
+
"step": 3440
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 2.2186495176848875,
|
| 2437 |
+
"grad_norm": 0.4235919117927551,
|
| 2438 |
+
"learning_rate": 5e-06,
|
| 2439 |
+
"loss": 0.6383,
|
| 2440 |
+
"step": 3450
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 2.22508038585209,
|
| 2444 |
+
"grad_norm": 0.4805637001991272,
|
| 2445 |
+
"learning_rate": 5e-06,
|
| 2446 |
+
"loss": 0.6413,
|
| 2447 |
+
"step": 3460
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 2.2315112540192925,
|
| 2451 |
+
"grad_norm": 0.4369070827960968,
|
| 2452 |
+
"learning_rate": 5e-06,
|
| 2453 |
+
"loss": 0.6348,
|
| 2454 |
+
"step": 3470
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 2.237942122186495,
|
| 2458 |
+
"grad_norm": 0.43729808926582336,
|
| 2459 |
+
"learning_rate": 5e-06,
|
| 2460 |
+
"loss": 0.6489,
|
| 2461 |
+
"step": 3480
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 2.244372990353698,
|
| 2465 |
+
"grad_norm": 0.514873206615448,
|
| 2466 |
+
"learning_rate": 5e-06,
|
| 2467 |
+
"loss": 0.6451,
|
| 2468 |
+
"step": 3490
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 2.2508038585209005,
|
| 2472 |
+
"grad_norm": 0.4235300123691559,
|
| 2473 |
+
"learning_rate": 5e-06,
|
| 2474 |
+
"loss": 0.6355,
|
| 2475 |
+
"step": 3500
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 2.257234726688103,
|
| 2479 |
+
"grad_norm": 0.4553775191307068,
|
| 2480 |
+
"learning_rate": 5e-06,
|
| 2481 |
+
"loss": 0.6374,
|
| 2482 |
+
"step": 3510
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 2.2636655948553055,
|
| 2486 |
+
"grad_norm": 0.444014310836792,
|
| 2487 |
+
"learning_rate": 5e-06,
|
| 2488 |
+
"loss": 0.6504,
|
| 2489 |
+
"step": 3520
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 2.270096463022508,
|
| 2493 |
+
"grad_norm": 0.4582684636116028,
|
| 2494 |
+
"learning_rate": 5e-06,
|
| 2495 |
+
"loss": 0.6338,
|
| 2496 |
+
"step": 3530
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 2.2765273311897105,
|
| 2500 |
+
"grad_norm": 0.5020269751548767,
|
| 2501 |
+
"learning_rate": 5e-06,
|
| 2502 |
+
"loss": 0.634,
|
| 2503 |
+
"step": 3540
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 2.282958199356913,
|
| 2507 |
+
"grad_norm": 0.44650107622146606,
|
| 2508 |
+
"learning_rate": 5e-06,
|
| 2509 |
+
"loss": 0.6382,
|
| 2510 |
+
"step": 3550
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 2.289389067524116,
|
| 2514 |
+
"grad_norm": 0.46265995502471924,
|
| 2515 |
+
"learning_rate": 5e-06,
|
| 2516 |
+
"loss": 0.6351,
|
| 2517 |
+
"step": 3560
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 2.2958199356913185,
|
| 2521 |
+
"grad_norm": 0.44239240884780884,
|
| 2522 |
+
"learning_rate": 5e-06,
|
| 2523 |
+
"loss": 0.6406,
|
| 2524 |
+
"step": 3570
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 2.302250803858521,
|
| 2528 |
+
"grad_norm": 0.49851298332214355,
|
| 2529 |
+
"learning_rate": 5e-06,
|
| 2530 |
+
"loss": 0.6373,
|
| 2531 |
+
"step": 3580
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 2.3086816720257235,
|
| 2535 |
+
"grad_norm": 0.44977161288261414,
|
| 2536 |
+
"learning_rate": 5e-06,
|
| 2537 |
+
"loss": 0.637,
|
| 2538 |
+
"step": 3590
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 2.315112540192926,
|
| 2542 |
+
"grad_norm": 0.47006750106811523,
|
| 2543 |
+
"learning_rate": 5e-06,
|
| 2544 |
+
"loss": 0.6429,
|
| 2545 |
+
"step": 3600
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 2.3215434083601285,
|
| 2549 |
+
"grad_norm": 0.43386831879615784,
|
| 2550 |
+
"learning_rate": 5e-06,
|
| 2551 |
+
"loss": 0.6411,
|
| 2552 |
+
"step": 3610
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 2.327974276527331,
|
| 2556 |
+
"grad_norm": 0.4304748475551605,
|
| 2557 |
+
"learning_rate": 5e-06,
|
| 2558 |
+
"loss": 0.6404,
|
| 2559 |
+
"step": 3620
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 2.334405144694534,
|
| 2563 |
+
"grad_norm": 0.4528113901615143,
|
| 2564 |
+
"learning_rate": 5e-06,
|
| 2565 |
+
"loss": 0.6409,
|
| 2566 |
+
"step": 3630
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 2.3408360128617365,
|
| 2570 |
+
"grad_norm": 0.43319591879844666,
|
| 2571 |
+
"learning_rate": 5e-06,
|
| 2572 |
+
"loss": 0.6397,
|
| 2573 |
+
"step": 3640
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 2.347266881028939,
|
| 2577 |
+
"grad_norm": 0.45046094059944153,
|
| 2578 |
+
"learning_rate": 5e-06,
|
| 2579 |
+
"loss": 0.6437,
|
| 2580 |
+
"step": 3650
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 2.3536977491961415,
|
| 2584 |
+
"grad_norm": 0.4646282494068146,
|
| 2585 |
+
"learning_rate": 5e-06,
|
| 2586 |
+
"loss": 0.6473,
|
| 2587 |
+
"step": 3660
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 2.360128617363344,
|
| 2591 |
+
"grad_norm": 0.4470469355583191,
|
| 2592 |
+
"learning_rate": 5e-06,
|
| 2593 |
+
"loss": 0.6337,
|
| 2594 |
+
"step": 3670
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 2.3665594855305465,
|
| 2598 |
+
"grad_norm": 0.45889395475387573,
|
| 2599 |
+
"learning_rate": 5e-06,
|
| 2600 |
+
"loss": 0.6393,
|
| 2601 |
+
"step": 3680
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 2.372990353697749,
|
| 2605 |
+
"grad_norm": 0.41356027126312256,
|
| 2606 |
+
"learning_rate": 5e-06,
|
| 2607 |
+
"loss": 0.6354,
|
| 2608 |
+
"step": 3690
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 2.379421221864952,
|
| 2612 |
+
"grad_norm": 0.4529729187488556,
|
| 2613 |
+
"learning_rate": 5e-06,
|
| 2614 |
+
"loss": 0.6403,
|
| 2615 |
+
"step": 3700
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 2.3858520900321545,
|
| 2619 |
+
"grad_norm": 0.4174403250217438,
|
| 2620 |
+
"learning_rate": 5e-06,
|
| 2621 |
+
"loss": 0.6377,
|
| 2622 |
+
"step": 3710
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 2.392282958199357,
|
| 2626 |
+
"grad_norm": 0.4575382173061371,
|
| 2627 |
+
"learning_rate": 5e-06,
|
| 2628 |
+
"loss": 0.6391,
|
| 2629 |
+
"step": 3720
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 2.3987138263665595,
|
| 2633 |
+
"grad_norm": 0.4538198709487915,
|
| 2634 |
+
"learning_rate": 5e-06,
|
| 2635 |
+
"loss": 0.6415,
|
| 2636 |
+
"step": 3730
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 2.405144694533762,
|
| 2640 |
+
"grad_norm": 0.4399804472923279,
|
| 2641 |
+
"learning_rate": 5e-06,
|
| 2642 |
+
"loss": 0.6348,
|
| 2643 |
+
"step": 3740
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 2.4115755627009645,
|
| 2647 |
+
"grad_norm": 0.4382459819316864,
|
| 2648 |
+
"learning_rate": 5e-06,
|
| 2649 |
+
"loss": 0.6415,
|
| 2650 |
+
"step": 3750
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 2.418006430868167,
|
| 2654 |
+
"grad_norm": 0.42854952812194824,
|
| 2655 |
+
"learning_rate": 5e-06,
|
| 2656 |
+
"loss": 0.6377,
|
| 2657 |
+
"step": 3760
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 2.42443729903537,
|
| 2661 |
+
"grad_norm": 0.4322606921195984,
|
| 2662 |
+
"learning_rate": 5e-06,
|
| 2663 |
+
"loss": 0.6435,
|
| 2664 |
+
"step": 3770
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 2.4308681672025725,
|
| 2668 |
+
"grad_norm": 0.45412078499794006,
|
| 2669 |
+
"learning_rate": 5e-06,
|
| 2670 |
+
"loss": 0.6401,
|
| 2671 |
+
"step": 3780
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 2.437299035369775,
|
| 2675 |
+
"grad_norm": 0.4531524181365967,
|
| 2676 |
+
"learning_rate": 5e-06,
|
| 2677 |
+
"loss": 0.6454,
|
| 2678 |
+
"step": 3790
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 2.4437299035369775,
|
| 2682 |
+
"grad_norm": 0.44421976804733276,
|
| 2683 |
+
"learning_rate": 5e-06,
|
| 2684 |
+
"loss": 0.6484,
|
| 2685 |
+
"step": 3800
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 2.45016077170418,
|
| 2689 |
+
"grad_norm": 0.4720653295516968,
|
| 2690 |
+
"learning_rate": 5e-06,
|
| 2691 |
+
"loss": 0.6431,
|
| 2692 |
+
"step": 3810
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 2.4565916398713825,
|
| 2696 |
+
"grad_norm": 0.4742264747619629,
|
| 2697 |
+
"learning_rate": 5e-06,
|
| 2698 |
+
"loss": 0.6371,
|
| 2699 |
+
"step": 3820
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 2.463022508038585,
|
| 2703 |
+
"grad_norm": 0.43964919447898865,
|
| 2704 |
+
"learning_rate": 5e-06,
|
| 2705 |
+
"loss": 0.6398,
|
| 2706 |
+
"step": 3830
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 2.469453376205788,
|
| 2710 |
+
"grad_norm": 0.43848881125450134,
|
| 2711 |
+
"learning_rate": 5e-06,
|
| 2712 |
+
"loss": 0.6419,
|
| 2713 |
+
"step": 3840
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 2.4758842443729905,
|
| 2717 |
+
"grad_norm": 0.4461026191711426,
|
| 2718 |
+
"learning_rate": 5e-06,
|
| 2719 |
+
"loss": 0.6379,
|
| 2720 |
+
"step": 3850
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 2.482315112540193,
|
| 2724 |
+
"grad_norm": 0.4415052831172943,
|
| 2725 |
+
"learning_rate": 5e-06,
|
| 2726 |
+
"loss": 0.6439,
|
| 2727 |
+
"step": 3860
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 2.4887459807073955,
|
| 2731 |
+
"grad_norm": 0.4191666543483734,
|
| 2732 |
+
"learning_rate": 5e-06,
|
| 2733 |
+
"loss": 0.6446,
|
| 2734 |
+
"step": 3870
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 2.495176848874598,
|
| 2738 |
+
"grad_norm": 0.43842655420303345,
|
| 2739 |
+
"learning_rate": 5e-06,
|
| 2740 |
+
"loss": 0.6439,
|
| 2741 |
+
"step": 3880
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 2.5016077170418005,
|
| 2745 |
+
"grad_norm": 0.4282081127166748,
|
| 2746 |
+
"learning_rate": 5e-06,
|
| 2747 |
+
"loss": 0.6479,
|
| 2748 |
+
"step": 3890
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 2.508038585209003,
|
| 2752 |
+
"grad_norm": 0.4626299738883972,
|
| 2753 |
+
"learning_rate": 5e-06,
|
| 2754 |
+
"loss": 0.6448,
|
| 2755 |
+
"step": 3900
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 2.514469453376206,
|
| 2759 |
+
"grad_norm": 0.44208213686943054,
|
| 2760 |
+
"learning_rate": 5e-06,
|
| 2761 |
+
"loss": 0.6439,
|
| 2762 |
+
"step": 3910
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 2.5209003215434085,
|
| 2766 |
+
"grad_norm": 0.43402865529060364,
|
| 2767 |
+
"learning_rate": 5e-06,
|
| 2768 |
+
"loss": 0.6369,
|
| 2769 |
+
"step": 3920
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 2.527331189710611,
|
| 2773 |
+
"grad_norm": 0.475366473197937,
|
| 2774 |
+
"learning_rate": 5e-06,
|
| 2775 |
+
"loss": 0.6446,
|
| 2776 |
+
"step": 3930
|
| 2777 |
+
},
|
| 2778 |
+
{
|
| 2779 |
+
"epoch": 2.5337620578778135,
|
| 2780 |
+
"grad_norm": 0.4425564408302307,
|
| 2781 |
+
"learning_rate": 5e-06,
|
| 2782 |
+
"loss": 0.6384,
|
| 2783 |
+
"step": 3940
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 2.540192926045016,
|
| 2787 |
+
"grad_norm": 0.43618571758270264,
|
| 2788 |
+
"learning_rate": 5e-06,
|
| 2789 |
+
"loss": 0.6408,
|
| 2790 |
+
"step": 3950
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 2.5466237942122185,
|
| 2794 |
+
"grad_norm": 0.49698591232299805,
|
| 2795 |
+
"learning_rate": 5e-06,
|
| 2796 |
+
"loss": 0.6303,
|
| 2797 |
+
"step": 3960
|
| 2798 |
+
},
|
| 2799 |
+
{
|
| 2800 |
+
"epoch": 2.553054662379421,
|
| 2801 |
+
"grad_norm": 0.5115300416946411,
|
| 2802 |
+
"learning_rate": 5e-06,
|
| 2803 |
+
"loss": 0.651,
|
| 2804 |
+
"step": 3970
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 2.559485530546624,
|
| 2808 |
+
"grad_norm": 0.6572542190551758,
|
| 2809 |
+
"learning_rate": 5e-06,
|
| 2810 |
+
"loss": 0.643,
|
| 2811 |
+
"step": 3980
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 2.5659163987138265,
|
| 2815 |
+
"grad_norm": 0.42369475960731506,
|
| 2816 |
+
"learning_rate": 5e-06,
|
| 2817 |
+
"loss": 0.6388,
|
| 2818 |
+
"step": 3990
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 2.572347266881029,
|
| 2822 |
+
"grad_norm": 0.44530463218688965,
|
| 2823 |
+
"learning_rate": 5e-06,
|
| 2824 |
+
"loss": 0.6418,
|
| 2825 |
+
"step": 4000
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 2.5787781350482315,
|
| 2829 |
+
"grad_norm": 0.49780553579330444,
|
| 2830 |
+
"learning_rate": 5e-06,
|
| 2831 |
+
"loss": 0.6334,
|
| 2832 |
+
"step": 4010
|
| 2833 |
+
},
|
| 2834 |
+
{
|
| 2835 |
+
"epoch": 2.585209003215434,
|
| 2836 |
+
"grad_norm": 0.45106691122055054,
|
| 2837 |
+
"learning_rate": 5e-06,
|
| 2838 |
+
"loss": 0.6398,
|
| 2839 |
+
"step": 4020
|
| 2840 |
+
},
|
| 2841 |
+
{
|
| 2842 |
+
"epoch": 2.5916398713826365,
|
| 2843 |
+
"grad_norm": 0.43149319291114807,
|
| 2844 |
+
"learning_rate": 5e-06,
|
| 2845 |
+
"loss": 0.6392,
|
| 2846 |
+
"step": 4030
|
| 2847 |
+
},
|
| 2848 |
+
{
|
| 2849 |
+
"epoch": 2.598070739549839,
|
| 2850 |
+
"grad_norm": 0.44727641344070435,
|
| 2851 |
+
"learning_rate": 5e-06,
|
| 2852 |
+
"loss": 0.643,
|
| 2853 |
+
"step": 4040
|
| 2854 |
+
},
|
| 2855 |
+
{
|
| 2856 |
+
"epoch": 2.604501607717042,
|
| 2857 |
+
"grad_norm": 0.46136513352394104,
|
| 2858 |
+
"learning_rate": 5e-06,
|
| 2859 |
+
"loss": 0.6407,
|
| 2860 |
+
"step": 4050
|
| 2861 |
+
},
|
| 2862 |
+
{
|
| 2863 |
+
"epoch": 2.6109324758842445,
|
| 2864 |
+
"grad_norm": 0.43373093008995056,
|
| 2865 |
+
"learning_rate": 5e-06,
|
| 2866 |
+
"loss": 0.635,
|
| 2867 |
+
"step": 4060
|
| 2868 |
+
},
|
| 2869 |
+
{
|
| 2870 |
+
"epoch": 2.617363344051447,
|
| 2871 |
+
"grad_norm": 0.42928022146224976,
|
| 2872 |
+
"learning_rate": 5e-06,
|
| 2873 |
+
"loss": 0.6364,
|
| 2874 |
+
"step": 4070
|
| 2875 |
+
},
|
| 2876 |
+
{
|
| 2877 |
+
"epoch": 2.6237942122186495,
|
| 2878 |
+
"grad_norm": 0.4322867691516876,
|
| 2879 |
+
"learning_rate": 5e-06,
|
| 2880 |
+
"loss": 0.6411,
|
| 2881 |
+
"step": 4080
|
| 2882 |
+
},
|
| 2883 |
+
{
|
| 2884 |
+
"epoch": 2.630225080385852,
|
| 2885 |
+
"grad_norm": 0.4528902471065521,
|
| 2886 |
+
"learning_rate": 5e-06,
|
| 2887 |
+
"loss": 0.6382,
|
| 2888 |
+
"step": 4090
|
| 2889 |
+
},
|
| 2890 |
+
{
|
| 2891 |
+
"epoch": 2.6366559485530545,
|
| 2892 |
+
"grad_norm": 0.45928648114204407,
|
| 2893 |
+
"learning_rate": 5e-06,
|
| 2894 |
+
"loss": 0.6429,
|
| 2895 |
+
"step": 4100
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 2.643086816720257,
|
| 2899 |
+
"grad_norm": 0.4491860270500183,
|
| 2900 |
+
"learning_rate": 5e-06,
|
| 2901 |
+
"loss": 0.6378,
|
| 2902 |
+
"step": 4110
|
| 2903 |
+
},
|
| 2904 |
+
{
|
| 2905 |
+
"epoch": 2.64951768488746,
|
| 2906 |
+
"grad_norm": 0.44847381114959717,
|
| 2907 |
+
"learning_rate": 5e-06,
|
| 2908 |
+
"loss": 0.6444,
|
| 2909 |
+
"step": 4120
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 2.6559485530546625,
|
| 2913 |
+
"grad_norm": 0.42893797159194946,
|
| 2914 |
+
"learning_rate": 5e-06,
|
| 2915 |
+
"loss": 0.6424,
|
| 2916 |
+
"step": 4130
|
| 2917 |
+
},
|
| 2918 |
+
{
|
| 2919 |
+
"epoch": 2.662379421221865,
|
| 2920 |
+
"grad_norm": 0.4790765345096588,
|
| 2921 |
+
"learning_rate": 5e-06,
|
| 2922 |
+
"loss": 0.6402,
|
| 2923 |
+
"step": 4140
|
| 2924 |
+
},
|
| 2925 |
+
{
|
| 2926 |
+
"epoch": 2.6688102893890675,
|
| 2927 |
+
"grad_norm": 0.4363596737384796,
|
| 2928 |
+
"learning_rate": 5e-06,
|
| 2929 |
+
"loss": 0.6402,
|
| 2930 |
+
"step": 4150
|
| 2931 |
+
},
|
| 2932 |
+
{
|
| 2933 |
+
"epoch": 2.67524115755627,
|
| 2934 |
+
"grad_norm": 0.4559105634689331,
|
| 2935 |
+
"learning_rate": 5e-06,
|
| 2936 |
+
"loss": 0.6475,
|
| 2937 |
+
"step": 4160
|
| 2938 |
+
},
|
| 2939 |
+
{
|
| 2940 |
+
"epoch": 2.6816720257234725,
|
| 2941 |
+
"grad_norm": 0.4630337059497833,
|
| 2942 |
+
"learning_rate": 5e-06,
|
| 2943 |
+
"loss": 0.6443,
|
| 2944 |
+
"step": 4170
|
| 2945 |
+
},
|
| 2946 |
+
{
|
| 2947 |
+
"epoch": 2.688102893890675,
|
| 2948 |
+
"grad_norm": 0.4628975987434387,
|
| 2949 |
+
"learning_rate": 5e-06,
|
| 2950 |
+
"loss": 0.6425,
|
| 2951 |
+
"step": 4180
|
| 2952 |
+
},
|
| 2953 |
+
{
|
| 2954 |
+
"epoch": 2.694533762057878,
|
| 2955 |
+
"grad_norm": 0.507043719291687,
|
| 2956 |
+
"learning_rate": 5e-06,
|
| 2957 |
+
"loss": 0.6385,
|
| 2958 |
+
"step": 4190
|
| 2959 |
+
},
|
| 2960 |
+
{
|
| 2961 |
+
"epoch": 2.7009646302250805,
|
| 2962 |
+
"grad_norm": 0.4438357353210449,
|
| 2963 |
+
"learning_rate": 5e-06,
|
| 2964 |
+
"loss": 0.6415,
|
| 2965 |
+
"step": 4200
|
| 2966 |
+
},
|
| 2967 |
+
{
|
| 2968 |
+
"epoch": 2.707395498392283,
|
| 2969 |
+
"grad_norm": 0.4772716462612152,
|
| 2970 |
+
"learning_rate": 5e-06,
|
| 2971 |
+
"loss": 0.6392,
|
| 2972 |
+
"step": 4210
|
| 2973 |
+
},
|
| 2974 |
+
{
|
| 2975 |
+
"epoch": 2.7138263665594855,
|
| 2976 |
+
"grad_norm": 0.4417891502380371,
|
| 2977 |
+
"learning_rate": 5e-06,
|
| 2978 |
+
"loss": 0.6365,
|
| 2979 |
+
"step": 4220
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 2.720257234726688,
|
| 2983 |
+
"grad_norm": 0.44056421518325806,
|
| 2984 |
+
"learning_rate": 5e-06,
|
| 2985 |
+
"loss": 0.6365,
|
| 2986 |
+
"step": 4230
|
| 2987 |
+
},
|
| 2988 |
+
{
|
| 2989 |
+
"epoch": 2.7266881028938905,
|
| 2990 |
+
"grad_norm": 0.4614117741584778,
|
| 2991 |
+
"learning_rate": 5e-06,
|
| 2992 |
+
"loss": 0.6425,
|
| 2993 |
+
"step": 4240
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 2.733118971061093,
|
| 2997 |
+
"grad_norm": 0.5050578117370605,
|
| 2998 |
+
"learning_rate": 5e-06,
|
| 2999 |
+
"loss": 0.6471,
|
| 3000 |
+
"step": 4250
|
| 3001 |
+
},
|
| 3002 |
+
{
|
| 3003 |
+
"epoch": 2.739549839228296,
|
| 3004 |
+
"grad_norm": 0.4652516543865204,
|
| 3005 |
+
"learning_rate": 5e-06,
|
| 3006 |
+
"loss": 0.6356,
|
| 3007 |
+
"step": 4260
|
| 3008 |
+
},
|
| 3009 |
+
{
|
| 3010 |
+
"epoch": 2.7459807073954985,
|
| 3011 |
+
"grad_norm": 0.44051194190979004,
|
| 3012 |
+
"learning_rate": 5e-06,
|
| 3013 |
+
"loss": 0.647,
|
| 3014 |
+
"step": 4270
|
| 3015 |
+
},
|
| 3016 |
+
{
|
| 3017 |
+
"epoch": 2.752411575562701,
|
| 3018 |
+
"grad_norm": 0.45722419023513794,
|
| 3019 |
+
"learning_rate": 5e-06,
|
| 3020 |
+
"loss": 0.64,
|
| 3021 |
+
"step": 4280
|
| 3022 |
+
},
|
| 3023 |
+
{
|
| 3024 |
+
"epoch": 2.7588424437299035,
|
| 3025 |
+
"grad_norm": 0.4401618242263794,
|
| 3026 |
+
"learning_rate": 5e-06,
|
| 3027 |
+
"loss": 0.6439,
|
| 3028 |
+
"step": 4290
|
| 3029 |
+
},
|
| 3030 |
+
{
|
| 3031 |
+
"epoch": 2.765273311897106,
|
| 3032 |
+
"grad_norm": 0.4500727653503418,
|
| 3033 |
+
"learning_rate": 5e-06,
|
| 3034 |
+
"loss": 0.6355,
|
| 3035 |
+
"step": 4300
|
| 3036 |
+
},
|
| 3037 |
+
{
|
| 3038 |
+
"epoch": 2.7717041800643085,
|
| 3039 |
+
"grad_norm": 0.44464367628097534,
|
| 3040 |
+
"learning_rate": 5e-06,
|
| 3041 |
+
"loss": 0.641,
|
| 3042 |
+
"step": 4310
|
| 3043 |
+
},
|
| 3044 |
+
{
|
| 3045 |
+
"epoch": 2.778135048231511,
|
| 3046 |
+
"grad_norm": 0.433173805475235,
|
| 3047 |
+
"learning_rate": 5e-06,
|
| 3048 |
+
"loss": 0.6479,
|
| 3049 |
+
"step": 4320
|
| 3050 |
+
},
|
| 3051 |
+
{
|
| 3052 |
+
"epoch": 2.784565916398714,
|
| 3053 |
+
"grad_norm": 0.44234412908554077,
|
| 3054 |
+
"learning_rate": 5e-06,
|
| 3055 |
+
"loss": 0.6474,
|
| 3056 |
+
"step": 4330
|
| 3057 |
+
},
|
| 3058 |
+
{
|
| 3059 |
+
"epoch": 2.7909967845659165,
|
| 3060 |
+
"grad_norm": 0.4479934275150299,
|
| 3061 |
+
"learning_rate": 5e-06,
|
| 3062 |
+
"loss": 0.6428,
|
| 3063 |
+
"step": 4340
|
| 3064 |
+
},
|
| 3065 |
+
{
|
| 3066 |
+
"epoch": 2.797427652733119,
|
| 3067 |
+
"grad_norm": 0.44619840383529663,
|
| 3068 |
+
"learning_rate": 5e-06,
|
| 3069 |
+
"loss": 0.6441,
|
| 3070 |
+
"step": 4350
|
| 3071 |
+
},
|
| 3072 |
+
{
|
| 3073 |
+
"epoch": 2.8038585209003215,
|
| 3074 |
+
"grad_norm": 0.4315251111984253,
|
| 3075 |
+
"learning_rate": 5e-06,
|
| 3076 |
+
"loss": 0.6477,
|
| 3077 |
+
"step": 4360
|
| 3078 |
+
},
|
| 3079 |
+
{
|
| 3080 |
+
"epoch": 2.810289389067524,
|
| 3081 |
+
"grad_norm": 0.43586409091949463,
|
| 3082 |
+
"learning_rate": 5e-06,
|
| 3083 |
+
"loss": 0.6504,
|
| 3084 |
+
"step": 4370
|
| 3085 |
+
},
|
| 3086 |
+
{
|
| 3087 |
+
"epoch": 2.816720257234727,
|
| 3088 |
+
"grad_norm": 0.4625302851200104,
|
| 3089 |
+
"learning_rate": 5e-06,
|
| 3090 |
+
"loss": 0.6523,
|
| 3091 |
+
"step": 4380
|
| 3092 |
+
},
|
| 3093 |
+
{
|
| 3094 |
+
"epoch": 2.823151125401929,
|
| 3095 |
+
"grad_norm": 0.46072930097579956,
|
| 3096 |
+
"learning_rate": 5e-06,
|
| 3097 |
+
"loss": 0.6423,
|
| 3098 |
+
"step": 4390
|
| 3099 |
+
},
|
| 3100 |
+
{
|
| 3101 |
+
"epoch": 2.829581993569132,
|
| 3102 |
+
"grad_norm": 0.4475337266921997,
|
| 3103 |
+
"learning_rate": 5e-06,
|
| 3104 |
+
"loss": 0.6483,
|
| 3105 |
+
"step": 4400
|
| 3106 |
+
},
|
| 3107 |
+
{
|
| 3108 |
+
"epoch": 2.8360128617363345,
|
| 3109 |
+
"grad_norm": 0.43071475625038147,
|
| 3110 |
+
"learning_rate": 5e-06,
|
| 3111 |
+
"loss": 0.6426,
|
| 3112 |
+
"step": 4410
|
| 3113 |
+
},
|
| 3114 |
+
{
|
| 3115 |
+
"epoch": 2.842443729903537,
|
| 3116 |
+
"grad_norm": 0.4402061700820923,
|
| 3117 |
+
"learning_rate": 5e-06,
|
| 3118 |
+
"loss": 0.6383,
|
| 3119 |
+
"step": 4420
|
| 3120 |
+
},
|
| 3121 |
+
{
|
| 3122 |
+
"epoch": 2.8488745980707395,
|
| 3123 |
+
"grad_norm": 0.47659188508987427,
|
| 3124 |
+
"learning_rate": 5e-06,
|
| 3125 |
+
"loss": 0.6378,
|
| 3126 |
+
"step": 4430
|
| 3127 |
+
},
|
| 3128 |
+
{
|
| 3129 |
+
"epoch": 2.855305466237942,
|
| 3130 |
+
"grad_norm": 0.44324415922164917,
|
| 3131 |
+
"learning_rate": 5e-06,
|
| 3132 |
+
"loss": 0.6445,
|
| 3133 |
+
"step": 4440
|
| 3134 |
+
},
|
| 3135 |
+
{
|
| 3136 |
+
"epoch": 2.861736334405145,
|
| 3137 |
+
"grad_norm": 0.5019335746765137,
|
| 3138 |
+
"learning_rate": 5e-06,
|
| 3139 |
+
"loss": 0.6429,
|
| 3140 |
+
"step": 4450
|
| 3141 |
+
},
|
| 3142 |
+
{
|
| 3143 |
+
"epoch": 2.868167202572347,
|
| 3144 |
+
"grad_norm": 0.47943827509880066,
|
| 3145 |
+
"learning_rate": 5e-06,
|
| 3146 |
+
"loss": 0.6403,
|
| 3147 |
+
"step": 4460
|
| 3148 |
+
},
|
| 3149 |
+
{
|
| 3150 |
+
"epoch": 2.87459807073955,
|
| 3151 |
+
"grad_norm": 0.4461328387260437,
|
| 3152 |
+
"learning_rate": 5e-06,
|
| 3153 |
+
"loss": 0.6423,
|
| 3154 |
+
"step": 4470
|
| 3155 |
+
},
|
| 3156 |
+
{
|
| 3157 |
+
"epoch": 2.8810289389067525,
|
| 3158 |
+
"grad_norm": 0.4263225495815277,
|
| 3159 |
+
"learning_rate": 5e-06,
|
| 3160 |
+
"loss": 0.6379,
|
| 3161 |
+
"step": 4480
|
| 3162 |
+
},
|
| 3163 |
+
{
|
| 3164 |
+
"epoch": 2.887459807073955,
|
| 3165 |
+
"grad_norm": 0.45077577233314514,
|
| 3166 |
+
"learning_rate": 5e-06,
|
| 3167 |
+
"loss": 0.6419,
|
| 3168 |
+
"step": 4490
|
| 3169 |
+
},
|
| 3170 |
+
{
|
| 3171 |
+
"epoch": 2.8938906752411575,
|
| 3172 |
+
"grad_norm": 0.4050474762916565,
|
| 3173 |
+
"learning_rate": 5e-06,
|
| 3174 |
+
"loss": 0.6361,
|
| 3175 |
+
"step": 4500
|
| 3176 |
+
},
|
| 3177 |
+
{
|
| 3178 |
+
"epoch": 2.90032154340836,
|
| 3179 |
+
"grad_norm": 0.4340299069881439,
|
| 3180 |
+
"learning_rate": 5e-06,
|
| 3181 |
+
"loss": 0.6386,
|
| 3182 |
+
"step": 4510
|
| 3183 |
+
},
|
| 3184 |
+
{
|
| 3185 |
+
"epoch": 2.906752411575563,
|
| 3186 |
+
"grad_norm": 0.42679664492607117,
|
| 3187 |
+
"learning_rate": 5e-06,
|
| 3188 |
+
"loss": 0.6423,
|
| 3189 |
+
"step": 4520
|
| 3190 |
+
},
|
| 3191 |
+
{
|
| 3192 |
+
"epoch": 2.913183279742765,
|
| 3193 |
+
"grad_norm": 0.4645717144012451,
|
| 3194 |
+
"learning_rate": 5e-06,
|
| 3195 |
+
"loss": 0.6422,
|
| 3196 |
+
"step": 4530
|
| 3197 |
+
},
|
| 3198 |
+
{
|
| 3199 |
+
"epoch": 2.919614147909968,
|
| 3200 |
+
"grad_norm": 0.471055805683136,
|
| 3201 |
+
"learning_rate": 5e-06,
|
| 3202 |
+
"loss": 0.6333,
|
| 3203 |
+
"step": 4540
|
| 3204 |
+
},
|
| 3205 |
+
{
|
| 3206 |
+
"epoch": 2.9260450160771705,
|
| 3207 |
+
"grad_norm": 0.5037959814071655,
|
| 3208 |
+
"learning_rate": 5e-06,
|
| 3209 |
+
"loss": 0.6419,
|
| 3210 |
+
"step": 4550
|
| 3211 |
+
},
|
| 3212 |
+
{
|
| 3213 |
+
"epoch": 2.932475884244373,
|
| 3214 |
+
"grad_norm": 0.4540598690509796,
|
| 3215 |
+
"learning_rate": 5e-06,
|
| 3216 |
+
"loss": 0.6393,
|
| 3217 |
+
"step": 4560
|
| 3218 |
+
},
|
| 3219 |
+
{
|
| 3220 |
+
"epoch": 2.9389067524115755,
|
| 3221 |
+
"grad_norm": 0.44270047545433044,
|
| 3222 |
+
"learning_rate": 5e-06,
|
| 3223 |
+
"loss": 0.6367,
|
| 3224 |
+
"step": 4570
|
| 3225 |
+
},
|
| 3226 |
+
{
|
| 3227 |
+
"epoch": 2.945337620578778,
|
| 3228 |
+
"grad_norm": 0.43946129083633423,
|
| 3229 |
+
"learning_rate": 5e-06,
|
| 3230 |
+
"loss": 0.6402,
|
| 3231 |
+
"step": 4580
|
| 3232 |
+
},
|
| 3233 |
+
{
|
| 3234 |
+
"epoch": 2.951768488745981,
|
| 3235 |
+
"grad_norm": 0.46501582860946655,
|
| 3236 |
+
"learning_rate": 5e-06,
|
| 3237 |
+
"loss": 0.6372,
|
| 3238 |
+
"step": 4590
|
| 3239 |
+
},
|
| 3240 |
+
{
|
| 3241 |
+
"epoch": 2.958199356913183,
|
| 3242 |
+
"grad_norm": 0.4638145864009857,
|
| 3243 |
+
"learning_rate": 5e-06,
|
| 3244 |
+
"loss": 0.6479,
|
| 3245 |
+
"step": 4600
|
| 3246 |
+
},
|
| 3247 |
+
{
|
| 3248 |
+
"epoch": 2.964630225080386,
|
| 3249 |
+
"grad_norm": 0.4363510310649872,
|
| 3250 |
+
"learning_rate": 5e-06,
|
| 3251 |
+
"loss": 0.6391,
|
| 3252 |
+
"step": 4610
|
| 3253 |
+
},
|
| 3254 |
+
{
|
| 3255 |
+
"epoch": 2.9710610932475885,
|
| 3256 |
+
"grad_norm": 0.42436662316322327,
|
| 3257 |
+
"learning_rate": 5e-06,
|
| 3258 |
+
"loss": 0.643,
|
| 3259 |
+
"step": 4620
|
| 3260 |
+
},
|
| 3261 |
+
{
|
| 3262 |
+
"epoch": 2.977491961414791,
|
| 3263 |
+
"grad_norm": 0.4569937288761139,
|
| 3264 |
+
"learning_rate": 5e-06,
|
| 3265 |
+
"loss": 0.6369,
|
| 3266 |
+
"step": 4630
|
| 3267 |
+
},
|
| 3268 |
+
{
|
| 3269 |
+
"epoch": 2.9839228295819935,
|
| 3270 |
+
"grad_norm": 0.4475838243961334,
|
| 3271 |
+
"learning_rate": 5e-06,
|
| 3272 |
+
"loss": 0.6402,
|
| 3273 |
+
"step": 4640
|
| 3274 |
+
},
|
| 3275 |
+
{
|
| 3276 |
+
"epoch": 2.990353697749196,
|
| 3277 |
+
"grad_norm": 0.47496703267097473,
|
| 3278 |
+
"learning_rate": 5e-06,
|
| 3279 |
+
"loss": 0.6315,
|
| 3280 |
+
"step": 4650
|
| 3281 |
+
},
|
| 3282 |
+
{
|
| 3283 |
+
"epoch": 2.996784565916399,
|
| 3284 |
+
"grad_norm": 0.4460706114768982,
|
| 3285 |
+
"learning_rate": 5e-06,
|
| 3286 |
+
"loss": 0.6421,
|
| 3287 |
+
"step": 4660
|
| 3288 |
+
},
|
| 3289 |
+
{
|
| 3290 |
+
"epoch": 3.0,
|
| 3291 |
+
"eval_loss": 0.7122252583503723,
|
| 3292 |
+
"eval_runtime": 140.1434,
|
| 3293 |
+
"eval_samples_per_second": 299.001,
|
| 3294 |
+
"eval_steps_per_second": 1.17,
|
| 3295 |
+
"step": 4665
|
| 3296 |
+
},
|
| 3297 |
+
{
|
| 3298 |
+
"epoch": 3.0,
|
| 3299 |
+
"step": 4665,
|
| 3300 |
+
"total_flos": 2.2026686217575832e+20,
|
| 3301 |
+
"train_loss": 0.6892199146274841,
|
| 3302 |
+
"train_runtime": 30399.4314,
|
| 3303 |
+
"train_samples_per_second": 78.569,
|
| 3304 |
+
"train_steps_per_second": 0.153
|
| 3305 |
+
}
|
| 3306 |
+
],
|
| 3307 |
+
"logging_steps": 10,
|
| 3308 |
+
"max_steps": 4665,
|
| 3309 |
+
"num_input_tokens_seen": 0,
|
| 3310 |
+
"num_train_epochs": 3,
|
| 3311 |
+
"save_steps": 500,
|
| 3312 |
+
"stateful_callbacks": {
|
| 3313 |
+
"TrainerControl": {
|
| 3314 |
+
"args": {
|
| 3315 |
+
"should_epoch_stop": false,
|
| 3316 |
+
"should_evaluate": false,
|
| 3317 |
+
"should_log": false,
|
| 3318 |
+
"should_save": true,
|
| 3319 |
+
"should_training_stop": true
|
| 3320 |
+
},
|
| 3321 |
+
"attributes": {}
|
| 3322 |
+
}
|
| 3323 |
+
},
|
| 3324 |
+
"total_flos": 2.2026686217575832e+20,
|
| 3325 |
+
"train_batch_size": 16,
|
| 3326 |
+
"trial_name": null,
|
| 3327 |
+
"trial_params": null
|
| 3328 |
+
}
|
training_eval_loss.png
ADDED
|
training_loss.png
ADDED
|