End of training
Browse files- README.md +2 -1
- all_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +3717 -0
- training_loss.png +0 -0
README.md
CHANGED
|
@@ -4,6 +4,7 @@ license: apache-2.0
|
|
| 4 |
base_model: Qwen/Qwen2.5-7B-Instruct
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
|
|
|
| 7 |
- generated_from_trainer
|
| 8 |
model-index:
|
| 9 |
- name: openthoughts3_science
|
|
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
# openthoughts3_science
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct) on
|
| 19 |
|
| 20 |
## Model description
|
| 21 |
|
|
|
|
| 4 |
base_model: Qwen/Qwen2.5-7B-Instruct
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
| 7 |
+
- full
|
| 8 |
- generated_from_trainer
|
| 9 |
model-index:
|
| 10 |
- name: openthoughts3_science
|
|
|
|
| 16 |
|
| 17 |
# openthoughts3_science
|
| 18 |
|
| 19 |
+
This model is a fine-tuned version of [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct) on the mlfoundations-dev/openthoughts3_science dataset.
|
| 20 |
|
| 21 |
## Model description
|
| 22 |
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"total_flos": 1.350981278339013e+19,
|
| 4 |
+
"train_loss": 1.0422365560985747,
|
| 5 |
+
"train_runtime": 52983.8768,
|
| 6 |
+
"train_samples_per_second": 5.072,
|
| 7 |
+
"train_steps_per_second": 0.01
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"total_flos": 1.350981278339013e+19,
|
| 4 |
+
"train_loss": 1.0422365560985747,
|
| 5 |
+
"train_runtime": 52983.8768,
|
| 6 |
+
"train_samples_per_second": 5.072,
|
| 7 |
+
"train_steps_per_second": 0.01
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3717 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 525,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.009523809523809525,
|
| 13 |
+
"grad_norm": 7.880720841975985,
|
| 14 |
+
"learning_rate": 1.509433962264151e-06,
|
| 15 |
+
"loss": 1.5676,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.01904761904761905,
|
| 20 |
+
"grad_norm": 7.860108844770417,
|
| 21 |
+
"learning_rate": 3.018867924528302e-06,
|
| 22 |
+
"loss": 1.5671,
|
| 23 |
+
"step": 2
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.02857142857142857,
|
| 27 |
+
"grad_norm": 7.7425094912660395,
|
| 28 |
+
"learning_rate": 4.528301886792453e-06,
|
| 29 |
+
"loss": 1.5511,
|
| 30 |
+
"step": 3
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.0380952380952381,
|
| 34 |
+
"grad_norm": 5.876257335486737,
|
| 35 |
+
"learning_rate": 6.037735849056604e-06,
|
| 36 |
+
"loss": 1.5083,
|
| 37 |
+
"step": 4
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.047619047619047616,
|
| 41 |
+
"grad_norm": 2.92144085922626,
|
| 42 |
+
"learning_rate": 7.5471698113207555e-06,
|
| 43 |
+
"loss": 1.4418,
|
| 44 |
+
"step": 5
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.05714285714285714,
|
| 48 |
+
"grad_norm": 2.432882075707435,
|
| 49 |
+
"learning_rate": 9.056603773584907e-06,
|
| 50 |
+
"loss": 1.4431,
|
| 51 |
+
"step": 6
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.06666666666666667,
|
| 55 |
+
"grad_norm": 6.42341738028925,
|
| 56 |
+
"learning_rate": 1.0566037735849058e-05,
|
| 57 |
+
"loss": 1.4534,
|
| 58 |
+
"step": 7
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.0761904761904762,
|
| 62 |
+
"grad_norm": 6.844420009717874,
|
| 63 |
+
"learning_rate": 1.2075471698113209e-05,
|
| 64 |
+
"loss": 1.4503,
|
| 65 |
+
"step": 8
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.08571428571428572,
|
| 69 |
+
"grad_norm": 8.72224792197926,
|
| 70 |
+
"learning_rate": 1.3584905660377358e-05,
|
| 71 |
+
"loss": 1.4895,
|
| 72 |
+
"step": 9
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.09523809523809523,
|
| 76 |
+
"grad_norm": 6.221915658880092,
|
| 77 |
+
"learning_rate": 1.5094339622641511e-05,
|
| 78 |
+
"loss": 1.4558,
|
| 79 |
+
"step": 10
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.10476190476190476,
|
| 83 |
+
"grad_norm": 4.297680995629592,
|
| 84 |
+
"learning_rate": 1.6603773584905664e-05,
|
| 85 |
+
"loss": 1.4005,
|
| 86 |
+
"step": 11
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.11428571428571428,
|
| 90 |
+
"grad_norm": 3.3637103263399886,
|
| 91 |
+
"learning_rate": 1.8113207547169813e-05,
|
| 92 |
+
"loss": 1.353,
|
| 93 |
+
"step": 12
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.12380952380952381,
|
| 97 |
+
"grad_norm": 2.102279140328028,
|
| 98 |
+
"learning_rate": 1.9622641509433963e-05,
|
| 99 |
+
"loss": 1.3221,
|
| 100 |
+
"step": 13
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.13333333333333333,
|
| 104 |
+
"grad_norm": 2.5449653919034234,
|
| 105 |
+
"learning_rate": 2.1132075471698115e-05,
|
| 106 |
+
"loss": 1.3029,
|
| 107 |
+
"step": 14
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.14285714285714285,
|
| 111 |
+
"grad_norm": 1.6782911213475984,
|
| 112 |
+
"learning_rate": 2.2641509433962265e-05,
|
| 113 |
+
"loss": 1.304,
|
| 114 |
+
"step": 15
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.1523809523809524,
|
| 118 |
+
"grad_norm": 1.7888774786584605,
|
| 119 |
+
"learning_rate": 2.4150943396226418e-05,
|
| 120 |
+
"loss": 1.2618,
|
| 121 |
+
"step": 16
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.1619047619047619,
|
| 125 |
+
"grad_norm": 1.5997868512124438,
|
| 126 |
+
"learning_rate": 2.5660377358490567e-05,
|
| 127 |
+
"loss": 1.2583,
|
| 128 |
+
"step": 17
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.17142857142857143,
|
| 132 |
+
"grad_norm": 1.2377999305796528,
|
| 133 |
+
"learning_rate": 2.7169811320754716e-05,
|
| 134 |
+
"loss": 1.2703,
|
| 135 |
+
"step": 18
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.18095238095238095,
|
| 139 |
+
"grad_norm": 1.486080611758641,
|
| 140 |
+
"learning_rate": 2.867924528301887e-05,
|
| 141 |
+
"loss": 1.2392,
|
| 142 |
+
"step": 19
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.19047619047619047,
|
| 146 |
+
"grad_norm": 1.27108266101122,
|
| 147 |
+
"learning_rate": 3.0188679245283022e-05,
|
| 148 |
+
"loss": 1.2377,
|
| 149 |
+
"step": 20
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.2,
|
| 153 |
+
"grad_norm": 0.9831199960325674,
|
| 154 |
+
"learning_rate": 3.169811320754717e-05,
|
| 155 |
+
"loss": 1.2123,
|
| 156 |
+
"step": 21
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.20952380952380953,
|
| 160 |
+
"grad_norm": 1.7763756530914847,
|
| 161 |
+
"learning_rate": 3.320754716981133e-05,
|
| 162 |
+
"loss": 1.2336,
|
| 163 |
+
"step": 22
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.21904761904761905,
|
| 167 |
+
"grad_norm": 1.3105067418162915,
|
| 168 |
+
"learning_rate": 3.471698113207548e-05,
|
| 169 |
+
"loss": 1.2172,
|
| 170 |
+
"step": 23
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.22857142857142856,
|
| 174 |
+
"grad_norm": 1.639321729355214,
|
| 175 |
+
"learning_rate": 3.6226415094339626e-05,
|
| 176 |
+
"loss": 1.2173,
|
| 177 |
+
"step": 24
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.23809523809523808,
|
| 181 |
+
"grad_norm": 1.476554104561696,
|
| 182 |
+
"learning_rate": 3.7735849056603776e-05,
|
| 183 |
+
"loss": 1.201,
|
| 184 |
+
"step": 25
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.24761904761904763,
|
| 188 |
+
"grad_norm": 1.530897069889547,
|
| 189 |
+
"learning_rate": 3.9245283018867925e-05,
|
| 190 |
+
"loss": 1.2083,
|
| 191 |
+
"step": 26
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.2571428571428571,
|
| 195 |
+
"grad_norm": 1.5322183506473377,
|
| 196 |
+
"learning_rate": 4.075471698113208e-05,
|
| 197 |
+
"loss": 1.2032,
|
| 198 |
+
"step": 27
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.26666666666666666,
|
| 202 |
+
"grad_norm": 1.3706277501006499,
|
| 203 |
+
"learning_rate": 4.226415094339623e-05,
|
| 204 |
+
"loss": 1.1896,
|
| 205 |
+
"step": 28
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.2761904761904762,
|
| 209 |
+
"grad_norm": 1.675751835429746,
|
| 210 |
+
"learning_rate": 4.377358490566038e-05,
|
| 211 |
+
"loss": 1.1955,
|
| 212 |
+
"step": 29
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.2857142857142857,
|
| 216 |
+
"grad_norm": 1.2446280009546904,
|
| 217 |
+
"learning_rate": 4.528301886792453e-05,
|
| 218 |
+
"loss": 1.1714,
|
| 219 |
+
"step": 30
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.29523809523809524,
|
| 223 |
+
"grad_norm": 1.232539792154481,
|
| 224 |
+
"learning_rate": 4.679245283018868e-05,
|
| 225 |
+
"loss": 1.1935,
|
| 226 |
+
"step": 31
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.3047619047619048,
|
| 230 |
+
"grad_norm": 1.1894442744080997,
|
| 231 |
+
"learning_rate": 4.8301886792452835e-05,
|
| 232 |
+
"loss": 1.1823,
|
| 233 |
+
"step": 32
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.3142857142857143,
|
| 237 |
+
"grad_norm": 1.6698359989666955,
|
| 238 |
+
"learning_rate": 4.9811320754716985e-05,
|
| 239 |
+
"loss": 1.1844,
|
| 240 |
+
"step": 33
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.3238095238095238,
|
| 244 |
+
"grad_norm": 1.1584686212441244,
|
| 245 |
+
"learning_rate": 5.1320754716981134e-05,
|
| 246 |
+
"loss": 1.168,
|
| 247 |
+
"step": 34
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.3333333333333333,
|
| 251 |
+
"grad_norm": 1.7374801678617988,
|
| 252 |
+
"learning_rate": 5.283018867924528e-05,
|
| 253 |
+
"loss": 1.175,
|
| 254 |
+
"step": 35
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.34285714285714286,
|
| 258 |
+
"grad_norm": 1.3774261987594567,
|
| 259 |
+
"learning_rate": 5.433962264150943e-05,
|
| 260 |
+
"loss": 1.1885,
|
| 261 |
+
"step": 36
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.3523809523809524,
|
| 265 |
+
"grad_norm": 1.144616111806608,
|
| 266 |
+
"learning_rate": 5.584905660377359e-05,
|
| 267 |
+
"loss": 1.1609,
|
| 268 |
+
"step": 37
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.3619047619047619,
|
| 272 |
+
"grad_norm": 2.044834450527845,
|
| 273 |
+
"learning_rate": 5.735849056603774e-05,
|
| 274 |
+
"loss": 1.1627,
|
| 275 |
+
"step": 38
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.37142857142857144,
|
| 279 |
+
"grad_norm": 1.8059910262399932,
|
| 280 |
+
"learning_rate": 5.886792452830189e-05,
|
| 281 |
+
"loss": 1.1599,
|
| 282 |
+
"step": 39
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.38095238095238093,
|
| 286 |
+
"grad_norm": 2.170117071827184,
|
| 287 |
+
"learning_rate": 6.0377358490566044e-05,
|
| 288 |
+
"loss": 1.1639,
|
| 289 |
+
"step": 40
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.3904761904761905,
|
| 293 |
+
"grad_norm": 1.3892171371581836,
|
| 294 |
+
"learning_rate": 6.18867924528302e-05,
|
| 295 |
+
"loss": 1.1672,
|
| 296 |
+
"step": 41
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.4,
|
| 300 |
+
"grad_norm": 3.12613489185021,
|
| 301 |
+
"learning_rate": 6.339622641509434e-05,
|
| 302 |
+
"loss": 1.1621,
|
| 303 |
+
"step": 42
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.4095238095238095,
|
| 307 |
+
"grad_norm": 1.9337153601065566,
|
| 308 |
+
"learning_rate": 6.49056603773585e-05,
|
| 309 |
+
"loss": 1.1624,
|
| 310 |
+
"step": 43
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.41904761904761906,
|
| 314 |
+
"grad_norm": 2.8416059583439055,
|
| 315 |
+
"learning_rate": 6.641509433962266e-05,
|
| 316 |
+
"loss": 1.1505,
|
| 317 |
+
"step": 44
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.42857142857142855,
|
| 321 |
+
"grad_norm": 1.9476637352918562,
|
| 322 |
+
"learning_rate": 6.79245283018868e-05,
|
| 323 |
+
"loss": 1.1661,
|
| 324 |
+
"step": 45
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.4380952380952381,
|
| 328 |
+
"grad_norm": 2.276534088165202,
|
| 329 |
+
"learning_rate": 6.943396226415095e-05,
|
| 330 |
+
"loss": 1.1519,
|
| 331 |
+
"step": 46
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.44761904761904764,
|
| 335 |
+
"grad_norm": 1.7741206434308783,
|
| 336 |
+
"learning_rate": 7.09433962264151e-05,
|
| 337 |
+
"loss": 1.1397,
|
| 338 |
+
"step": 47
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.45714285714285713,
|
| 342 |
+
"grad_norm": 1.2590011451410716,
|
| 343 |
+
"learning_rate": 7.245283018867925e-05,
|
| 344 |
+
"loss": 1.1496,
|
| 345 |
+
"step": 48
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.4666666666666667,
|
| 349 |
+
"grad_norm": 2.2235594751942145,
|
| 350 |
+
"learning_rate": 7.396226415094341e-05,
|
| 351 |
+
"loss": 1.1646,
|
| 352 |
+
"step": 49
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.47619047619047616,
|
| 356 |
+
"grad_norm": 1.3848690153622902,
|
| 357 |
+
"learning_rate": 7.547169811320755e-05,
|
| 358 |
+
"loss": 1.1446,
|
| 359 |
+
"step": 50
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.4857142857142857,
|
| 363 |
+
"grad_norm": 2.2564098595161375,
|
| 364 |
+
"learning_rate": 7.698113207547171e-05,
|
| 365 |
+
"loss": 1.1521,
|
| 366 |
+
"step": 51
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.49523809523809526,
|
| 370 |
+
"grad_norm": 2.2598603091896234,
|
| 371 |
+
"learning_rate": 7.849056603773585e-05,
|
| 372 |
+
"loss": 1.1501,
|
| 373 |
+
"step": 52
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.5047619047619047,
|
| 377 |
+
"grad_norm": 1.9630215060124938,
|
| 378 |
+
"learning_rate": 8e-05,
|
| 379 |
+
"loss": 1.1558,
|
| 380 |
+
"step": 53
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.5142857142857142,
|
| 384 |
+
"grad_norm": 2.449644231642783,
|
| 385 |
+
"learning_rate": 7.99991139787449e-05,
|
| 386 |
+
"loss": 1.1478,
|
| 387 |
+
"step": 54
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.5238095238095238,
|
| 391 |
+
"grad_norm": 1.8285447577052554,
|
| 392 |
+
"learning_rate": 7.999645595423128e-05,
|
| 393 |
+
"loss": 1.1522,
|
| 394 |
+
"step": 55
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.5333333333333333,
|
| 398 |
+
"grad_norm": 1.6446782549594272,
|
| 399 |
+
"learning_rate": 7.999202604421244e-05,
|
| 400 |
+
"loss": 1.1488,
|
| 401 |
+
"step": 56
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.5428571428571428,
|
| 405 |
+
"grad_norm": 1.6532082997276394,
|
| 406 |
+
"learning_rate": 7.998582444493811e-05,
|
| 407 |
+
"loss": 1.135,
|
| 408 |
+
"step": 57
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.5523809523809524,
|
| 412 |
+
"grad_norm": 1.3217448563312193,
|
| 413 |
+
"learning_rate": 7.997785143114573e-05,
|
| 414 |
+
"loss": 1.1381,
|
| 415 |
+
"step": 58
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.5619047619047619,
|
| 419 |
+
"grad_norm": 1.5193922578374879,
|
| 420 |
+
"learning_rate": 7.996810735604828e-05,
|
| 421 |
+
"loss": 1.1488,
|
| 422 |
+
"step": 59
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.5714285714285714,
|
| 426 |
+
"grad_norm": 2.032121014977749,
|
| 427 |
+
"learning_rate": 7.995659265131865e-05,
|
| 428 |
+
"loss": 1.1486,
|
| 429 |
+
"step": 60
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.580952380952381,
|
| 433 |
+
"grad_norm": 1.1859854441122255,
|
| 434 |
+
"learning_rate": 7.994330782707048e-05,
|
| 435 |
+
"loss": 1.1463,
|
| 436 |
+
"step": 61
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.5904761904761905,
|
| 440 |
+
"grad_norm": 1.8225832846306127,
|
| 441 |
+
"learning_rate": 7.992825347183563e-05,
|
| 442 |
+
"loss": 1.1422,
|
| 443 |
+
"step": 62
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.6,
|
| 447 |
+
"grad_norm": 3.1767099132769046,
|
| 448 |
+
"learning_rate": 7.991143025253801e-05,
|
| 449 |
+
"loss": 1.1281,
|
| 450 |
+
"step": 63
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.6095238095238096,
|
| 454 |
+
"grad_norm": 1.7321438313588222,
|
| 455 |
+
"learning_rate": 7.989283891446413e-05,
|
| 456 |
+
"loss": 1.1509,
|
| 457 |
+
"step": 64
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.6190476190476191,
|
| 461 |
+
"grad_norm": 3.7251686905910355,
|
| 462 |
+
"learning_rate": 7.987248028123001e-05,
|
| 463 |
+
"loss": 1.161,
|
| 464 |
+
"step": 65
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.6285714285714286,
|
| 468 |
+
"grad_norm": 2.9492630393156514,
|
| 469 |
+
"learning_rate": 7.985035525474475e-05,
|
| 470 |
+
"loss": 1.1583,
|
| 471 |
+
"step": 66
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.638095238095238,
|
| 475 |
+
"grad_norm": 2.9426097334554897,
|
| 476 |
+
"learning_rate": 7.982646481517054e-05,
|
| 477 |
+
"loss": 1.1442,
|
| 478 |
+
"step": 67
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.6476190476190476,
|
| 482 |
+
"grad_norm": 2.469295775843566,
|
| 483 |
+
"learning_rate": 7.980081002087923e-05,
|
| 484 |
+
"loss": 1.1401,
|
| 485 |
+
"step": 68
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.6571428571428571,
|
| 489 |
+
"grad_norm": 2.652729866210755,
|
| 490 |
+
"learning_rate": 7.97733920084055e-05,
|
| 491 |
+
"loss": 1.1488,
|
| 492 |
+
"step": 69
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.6666666666666666,
|
| 496 |
+
"grad_norm": 2.2477356726812125,
|
| 497 |
+
"learning_rate": 7.97442119923964e-05,
|
| 498 |
+
"loss": 1.1241,
|
| 499 |
+
"step": 70
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.6761904761904762,
|
| 503 |
+
"grad_norm": 2.071700619970133,
|
| 504 |
+
"learning_rate": 7.971327126555767e-05,
|
| 505 |
+
"loss": 1.1383,
|
| 506 |
+
"step": 71
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.6857142857142857,
|
| 510 |
+
"grad_norm": 1.9746783226121274,
|
| 511 |
+
"learning_rate": 7.968057119859639e-05,
|
| 512 |
+
"loss": 1.132,
|
| 513 |
+
"step": 72
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.6952380952380952,
|
| 517 |
+
"grad_norm": 1.5945211845534675,
|
| 518 |
+
"learning_rate": 7.96461132401603e-05,
|
| 519 |
+
"loss": 1.1326,
|
| 520 |
+
"step": 73
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.7047619047619048,
|
| 524 |
+
"grad_norm": 1.9842082262603224,
|
| 525 |
+
"learning_rate": 7.960989891677354e-05,
|
| 526 |
+
"loss": 1.131,
|
| 527 |
+
"step": 74
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.7142857142857143,
|
| 531 |
+
"grad_norm": 1.0744704158928022,
|
| 532 |
+
"learning_rate": 7.957192983276915e-05,
|
| 533 |
+
"loss": 1.1286,
|
| 534 |
+
"step": 75
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.7238095238095238,
|
| 538 |
+
"grad_norm": 2.4648933927530954,
|
| 539 |
+
"learning_rate": 7.953220767021789e-05,
|
| 540 |
+
"loss": 1.1334,
|
| 541 |
+
"step": 76
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.7333333333333333,
|
| 545 |
+
"grad_norm": 1.6229327724824014,
|
| 546 |
+
"learning_rate": 7.949073418885378e-05,
|
| 547 |
+
"loss": 1.1268,
|
| 548 |
+
"step": 77
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.7428571428571429,
|
| 552 |
+
"grad_norm": 2.070091284758308,
|
| 553 |
+
"learning_rate": 7.944751122599613e-05,
|
| 554 |
+
"loss": 1.13,
|
| 555 |
+
"step": 78
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.7523809523809524,
|
| 559 |
+
"grad_norm": 1.5442744548822007,
|
| 560 |
+
"learning_rate": 7.940254069646813e-05,
|
| 561 |
+
"loss": 1.1219,
|
| 562 |
+
"step": 79
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.7619047619047619,
|
| 566 |
+
"grad_norm": 1.737018571216328,
|
| 567 |
+
"learning_rate": 7.935582459251202e-05,
|
| 568 |
+
"loss": 1.1219,
|
| 569 |
+
"step": 80
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.7714285714285715,
|
| 573 |
+
"grad_norm": 1.5613075146451556,
|
| 574 |
+
"learning_rate": 7.930736498370085e-05,
|
| 575 |
+
"loss": 1.1096,
|
| 576 |
+
"step": 81
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.780952380952381,
|
| 580 |
+
"grad_norm": 1.0920845324913702,
|
| 581 |
+
"learning_rate": 7.925716401684678e-05,
|
| 582 |
+
"loss": 1.119,
|
| 583 |
+
"step": 82
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.7904761904761904,
|
| 587 |
+
"grad_norm": 1.9034104210285714,
|
| 588 |
+
"learning_rate": 7.920522391590604e-05,
|
| 589 |
+
"loss": 1.1236,
|
| 590 |
+
"step": 83
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.8,
|
| 594 |
+
"grad_norm": 1.345035723453052,
|
| 595 |
+
"learning_rate": 7.915154698188027e-05,
|
| 596 |
+
"loss": 1.1324,
|
| 597 |
+
"step": 84
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.8095238095238095,
|
| 601 |
+
"grad_norm": 1.2875694611881727,
|
| 602 |
+
"learning_rate": 7.909613559271467e-05,
|
| 603 |
+
"loss": 1.1136,
|
| 604 |
+
"step": 85
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.819047619047619,
|
| 608 |
+
"grad_norm": 1.0224645041826803,
|
| 609 |
+
"learning_rate": 7.90389922031927e-05,
|
| 610 |
+
"loss": 1.1226,
|
| 611 |
+
"step": 86
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.8285714285714286,
|
| 615 |
+
"grad_norm": 1.3257820885017062,
|
| 616 |
+
"learning_rate": 7.898011934482725e-05,
|
| 617 |
+
"loss": 1.1214,
|
| 618 |
+
"step": 87
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.8380952380952381,
|
| 622 |
+
"grad_norm": 1.839693710231441,
|
| 623 |
+
"learning_rate": 7.89195196257485e-05,
|
| 624 |
+
"loss": 1.1406,
|
| 625 |
+
"step": 88
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.8476190476190476,
|
| 629 |
+
"grad_norm": 1.2987193541355626,
|
| 630 |
+
"learning_rate": 7.88571957305884e-05,
|
| 631 |
+
"loss": 1.1231,
|
| 632 |
+
"step": 89
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.8571428571428571,
|
| 636 |
+
"grad_norm": 1.4098694982637217,
|
| 637 |
+
"learning_rate": 7.879315042036176e-05,
|
| 638 |
+
"loss": 1.1386,
|
| 639 |
+
"step": 90
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.8666666666666667,
|
| 643 |
+
"grad_norm": 1.6007046498489645,
|
| 644 |
+
"learning_rate": 7.872738653234387e-05,
|
| 645 |
+
"loss": 1.1108,
|
| 646 |
+
"step": 91
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.8761904761904762,
|
| 650 |
+
"grad_norm": 1.0544725776613069,
|
| 651 |
+
"learning_rate": 7.865990697994488e-05,
|
| 652 |
+
"loss": 1.1285,
|
| 653 |
+
"step": 92
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.8857142857142857,
|
| 657 |
+
"grad_norm": 1.0388252500166617,
|
| 658 |
+
"learning_rate": 7.859071475258065e-05,
|
| 659 |
+
"loss": 1.1265,
|
| 660 |
+
"step": 93
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.8952380952380953,
|
| 664 |
+
"grad_norm": 1.3249551090301468,
|
| 665 |
+
"learning_rate": 7.85198129155404e-05,
|
| 666 |
+
"loss": 1.1229,
|
| 667 |
+
"step": 94
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.9047619047619048,
|
| 671 |
+
"grad_norm": 1.2506636003424314,
|
| 672 |
+
"learning_rate": 7.844720460985086e-05,
|
| 673 |
+
"loss": 1.1284,
|
| 674 |
+
"step": 95
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.9142857142857143,
|
| 678 |
+
"grad_norm": 1.1709974083254924,
|
| 679 |
+
"learning_rate": 7.837289305213715e-05,
|
| 680 |
+
"loss": 1.1221,
|
| 681 |
+
"step": 96
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.9238095238095239,
|
| 685 |
+
"grad_norm": 1.5298739597522653,
|
| 686 |
+
"learning_rate": 7.829688153448022e-05,
|
| 687 |
+
"loss": 1.122,
|
| 688 |
+
"step": 97
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.9333333333333333,
|
| 692 |
+
"grad_norm": 1.0400221166560895,
|
| 693 |
+
"learning_rate": 7.821917342427112e-05,
|
| 694 |
+
"loss": 1.1125,
|
| 695 |
+
"step": 98
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.9428571428571428,
|
| 699 |
+
"grad_norm": 1.1449775808812492,
|
| 700 |
+
"learning_rate": 7.81397721640617e-05,
|
| 701 |
+
"loss": 1.1098,
|
| 702 |
+
"step": 99
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.9523809523809523,
|
| 706 |
+
"grad_norm": 1.09547909217965,
|
| 707 |
+
"learning_rate": 7.805868127141217e-05,
|
| 708 |
+
"loss": 1.1097,
|
| 709 |
+
"step": 100
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.9619047619047619,
|
| 713 |
+
"grad_norm": 1.0999337584802662,
|
| 714 |
+
"learning_rate": 7.797590433873526e-05,
|
| 715 |
+
"loss": 1.1146,
|
| 716 |
+
"step": 101
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.9714285714285714,
|
| 720 |
+
"grad_norm": 1.5977618876155488,
|
| 721 |
+
"learning_rate": 7.789144503313704e-05,
|
| 722 |
+
"loss": 1.1375,
|
| 723 |
+
"step": 102
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.9809523809523809,
|
| 727 |
+
"grad_norm": 0.8781433045555708,
|
| 728 |
+
"learning_rate": 7.780530709625455e-05,
|
| 729 |
+
"loss": 1.1191,
|
| 730 |
+
"step": 103
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.9904761904761905,
|
| 734 |
+
"grad_norm": 1.2752214725002813,
|
| 735 |
+
"learning_rate": 7.771749434408989e-05,
|
| 736 |
+
"loss": 1.1173,
|
| 737 |
+
"step": 104
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 1.0,
|
| 741 |
+
"grad_norm": 1.0844449972573356,
|
| 742 |
+
"learning_rate": 7.762801066684133e-05,
|
| 743 |
+
"loss": 1.1137,
|
| 744 |
+
"step": 105
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 1.0095238095238095,
|
| 748 |
+
"grad_norm": 1.082120963718663,
|
| 749 |
+
"learning_rate": 7.753686002873087e-05,
|
| 750 |
+
"loss": 1.0737,
|
| 751 |
+
"step": 106
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 1.019047619047619,
|
| 755 |
+
"grad_norm": 1.1974759867644165,
|
| 756 |
+
"learning_rate": 7.744404646782866e-05,
|
| 757 |
+
"loss": 1.0849,
|
| 758 |
+
"step": 107
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 1.0285714285714285,
|
| 762 |
+
"grad_norm": 1.6630439720081593,
|
| 763 |
+
"learning_rate": 7.734957409587404e-05,
|
| 764 |
+
"loss": 1.0861,
|
| 765 |
+
"step": 108
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 1.0380952380952382,
|
| 769 |
+
"grad_norm": 0.8074012816288818,
|
| 770 |
+
"learning_rate": 7.725344709809355e-05,
|
| 771 |
+
"loss": 1.0785,
|
| 772 |
+
"step": 109
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 1.0476190476190477,
|
| 776 |
+
"grad_norm": 1.6368997832238315,
|
| 777 |
+
"learning_rate": 7.715566973301529e-05,
|
| 778 |
+
"loss": 1.0963,
|
| 779 |
+
"step": 110
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 1.0571428571428572,
|
| 783 |
+
"grad_norm": 1.297969917454931,
|
| 784 |
+
"learning_rate": 7.70562463322805e-05,
|
| 785 |
+
"loss": 1.0728,
|
| 786 |
+
"step": 111
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 1.0666666666666667,
|
| 790 |
+
"grad_norm": 1.4036049691933277,
|
| 791 |
+
"learning_rate": 7.695518130045147e-05,
|
| 792 |
+
"loss": 1.0806,
|
| 793 |
+
"step": 112
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 1.0761904761904761,
|
| 797 |
+
"grad_norm": 0.9013266727291808,
|
| 798 |
+
"learning_rate": 7.685247911481652e-05,
|
| 799 |
+
"loss": 1.0981,
|
| 800 |
+
"step": 113
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 1.0857142857142856,
|
| 804 |
+
"grad_norm": 1.0917190640539323,
|
| 805 |
+
"learning_rate": 7.674814432519163e-05,
|
| 806 |
+
"loss": 1.0734,
|
| 807 |
+
"step": 114
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 1.0952380952380953,
|
| 811 |
+
"grad_norm": 1.0218269030800684,
|
| 812 |
+
"learning_rate": 7.664218155371884e-05,
|
| 813 |
+
"loss": 1.0658,
|
| 814 |
+
"step": 115
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 1.1047619047619048,
|
| 818 |
+
"grad_norm": 0.9481079847045792,
|
| 819 |
+
"learning_rate": 7.653459549466157e-05,
|
| 820 |
+
"loss": 1.086,
|
| 821 |
+
"step": 116
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 1.1142857142857143,
|
| 825 |
+
"grad_norm": 1.1795366957840878,
|
| 826 |
+
"learning_rate": 7.642539091419654e-05,
|
| 827 |
+
"loss": 1.1062,
|
| 828 |
+
"step": 117
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 1.1238095238095238,
|
| 832 |
+
"grad_norm": 1.2913118926572955,
|
| 833 |
+
"learning_rate": 7.631457265020274e-05,
|
| 834 |
+
"loss": 1.075,
|
| 835 |
+
"step": 118
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 1.1333333333333333,
|
| 839 |
+
"grad_norm": 0.8478614709821788,
|
| 840 |
+
"learning_rate": 7.620214561204704e-05,
|
| 841 |
+
"loss": 1.0683,
|
| 842 |
+
"step": 119
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 1.1428571428571428,
|
| 846 |
+
"grad_norm": 1.0862987356843379,
|
| 847 |
+
"learning_rate": 7.608811478036671e-05,
|
| 848 |
+
"loss": 1.0846,
|
| 849 |
+
"step": 120
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 1.1523809523809523,
|
| 853 |
+
"grad_norm": 1.090222081314668,
|
| 854 |
+
"learning_rate": 7.597248520684878e-05,
|
| 855 |
+
"loss": 1.0905,
|
| 856 |
+
"step": 121
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 1.161904761904762,
|
| 860 |
+
"grad_norm": 0.7532978409995891,
|
| 861 |
+
"learning_rate": 7.585526201400623e-05,
|
| 862 |
+
"loss": 1.0791,
|
| 863 |
+
"step": 122
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 1.1714285714285715,
|
| 867 |
+
"grad_norm": 1.1901023584369215,
|
| 868 |
+
"learning_rate": 7.57364503949511e-05,
|
| 869 |
+
"loss": 1.0869,
|
| 870 |
+
"step": 123
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 1.180952380952381,
|
| 874 |
+
"grad_norm": 1.2655635735771573,
|
| 875 |
+
"learning_rate": 7.56160556131644e-05,
|
| 876 |
+
"loss": 1.0688,
|
| 877 |
+
"step": 124
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 1.1904761904761905,
|
| 881 |
+
"grad_norm": 0.8393572943728383,
|
| 882 |
+
"learning_rate": 7.549408300226287e-05,
|
| 883 |
+
"loss": 1.0798,
|
| 884 |
+
"step": 125
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 1.2,
|
| 888 |
+
"grad_norm": 1.0763767921818734,
|
| 889 |
+
"learning_rate": 7.537053796576282e-05,
|
| 890 |
+
"loss": 1.0755,
|
| 891 |
+
"step": 126
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 1.2095238095238094,
|
| 895 |
+
"grad_norm": 0.810181087854412,
|
| 896 |
+
"learning_rate": 7.524542597684066e-05,
|
| 897 |
+
"loss": 1.0722,
|
| 898 |
+
"step": 127
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 1.2190476190476192,
|
| 902 |
+
"grad_norm": 0.7248802686574851,
|
| 903 |
+
"learning_rate": 7.51187525780905e-05,
|
| 904 |
+
"loss": 1.0718,
|
| 905 |
+
"step": 128
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 1.2285714285714286,
|
| 909 |
+
"grad_norm": 1.3361324797449061,
|
| 910 |
+
"learning_rate": 7.499052338127845e-05,
|
| 911 |
+
"loss": 1.0778,
|
| 912 |
+
"step": 129
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 1.2380952380952381,
|
| 916 |
+
"grad_norm": 0.8378010550345113,
|
| 917 |
+
"learning_rate": 7.486074406709429e-05,
|
| 918 |
+
"loss": 1.068,
|
| 919 |
+
"step": 130
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 1.2476190476190476,
|
| 923 |
+
"grad_norm": 0.703515432967172,
|
| 924 |
+
"learning_rate": 7.47294203848995e-05,
|
| 925 |
+
"loss": 1.0832,
|
| 926 |
+
"step": 131
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 1.2571428571428571,
|
| 930 |
+
"grad_norm": 0.5700766848892007,
|
| 931 |
+
"learning_rate": 7.459655815247278e-05,
|
| 932 |
+
"loss": 1.0702,
|
| 933 |
+
"step": 132
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 1.2666666666666666,
|
| 937 |
+
"grad_norm": 0.7201933038596476,
|
| 938 |
+
"learning_rate": 7.446216325575225e-05,
|
| 939 |
+
"loss": 1.0751,
|
| 940 |
+
"step": 133
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 1.276190476190476,
|
| 944 |
+
"grad_norm": 0.9463247169207133,
|
| 945 |
+
"learning_rate": 7.432624164857465e-05,
|
| 946 |
+
"loss": 1.0597,
|
| 947 |
+
"step": 134
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 1.2857142857142856,
|
| 951 |
+
"grad_norm": 1.0411988380085484,
|
| 952 |
+
"learning_rate": 7.418879935241162e-05,
|
| 953 |
+
"loss": 1.0694,
|
| 954 |
+
"step": 135
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 1.2952380952380953,
|
| 958 |
+
"grad_norm": 1.0701186830070861,
|
| 959 |
+
"learning_rate": 7.404984245610296e-05,
|
| 960 |
+
"loss": 1.0706,
|
| 961 |
+
"step": 136
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 1.3047619047619048,
|
| 965 |
+
"grad_norm": 1.1554699877331098,
|
| 966 |
+
"learning_rate": 7.390937711558683e-05,
|
| 967 |
+
"loss": 1.0619,
|
| 968 |
+
"step": 137
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 1.3142857142857143,
|
| 972 |
+
"grad_norm": 1.1704855402151975,
|
| 973 |
+
"learning_rate": 7.376740955362715e-05,
|
| 974 |
+
"loss": 1.0535,
|
| 975 |
+
"step": 138
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 1.3238095238095238,
|
| 979 |
+
"grad_norm": 1.1741068413982254,
|
| 980 |
+
"learning_rate": 7.362394605953773e-05,
|
| 981 |
+
"loss": 1.0775,
|
| 982 |
+
"step": 139
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 1.3333333333333333,
|
| 986 |
+
"grad_norm": 1.2007489108920013,
|
| 987 |
+
"learning_rate": 7.347899298890386e-05,
|
| 988 |
+
"loss": 1.0726,
|
| 989 |
+
"step": 140
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 1.342857142857143,
|
| 993 |
+
"grad_norm": 0.9964002243232726,
|
| 994 |
+
"learning_rate": 7.33325567633006e-05,
|
| 995 |
+
"loss": 1.0602,
|
| 996 |
+
"step": 141
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 1.3523809523809525,
|
| 1000 |
+
"grad_norm": 0.8355502476587309,
|
| 1001 |
+
"learning_rate": 7.31846438700084e-05,
|
| 1002 |
+
"loss": 1.0774,
|
| 1003 |
+
"step": 142
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 1.361904761904762,
|
| 1007 |
+
"grad_norm": 0.7029298880866707,
|
| 1008 |
+
"learning_rate": 7.303526086172558e-05,
|
| 1009 |
+
"loss": 1.0846,
|
| 1010 |
+
"step": 143
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 1.3714285714285714,
|
| 1014 |
+
"grad_norm": 0.9269565547406854,
|
| 1015 |
+
"learning_rate": 7.288441435627821e-05,
|
| 1016 |
+
"loss": 1.0627,
|
| 1017 |
+
"step": 144
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 1.380952380952381,
|
| 1021 |
+
"grad_norm": 1.0950185791329867,
|
| 1022 |
+
"learning_rate": 7.273211103632676e-05,
|
| 1023 |
+
"loss": 1.0841,
|
| 1024 |
+
"step": 145
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 1.3904761904761904,
|
| 1028 |
+
"grad_norm": 1.0029086769564732,
|
| 1029 |
+
"learning_rate": 7.25783576490702e-05,
|
| 1030 |
+
"loss": 1.0559,
|
| 1031 |
+
"step": 146
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 1.4,
|
| 1035 |
+
"grad_norm": 1.0116599010130591,
|
| 1036 |
+
"learning_rate": 7.242316100594696e-05,
|
| 1037 |
+
"loss": 1.063,
|
| 1038 |
+
"step": 147
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 1.4095238095238094,
|
| 1042 |
+
"grad_norm": 0.8363130430044641,
|
| 1043 |
+
"learning_rate": 7.226652798233327e-05,
|
| 1044 |
+
"loss": 1.0601,
|
| 1045 |
+
"step": 148
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 1.4190476190476191,
|
| 1049 |
+
"grad_norm": 0.7749911266241076,
|
| 1050 |
+
"learning_rate": 7.210846551723855e-05,
|
| 1051 |
+
"loss": 1.0655,
|
| 1052 |
+
"step": 149
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 1.4285714285714286,
|
| 1056 |
+
"grad_norm": 0.8621768850747066,
|
| 1057 |
+
"learning_rate": 7.194898061299798e-05,
|
| 1058 |
+
"loss": 1.0687,
|
| 1059 |
+
"step": 150
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 1.438095238095238,
|
| 1063 |
+
"grad_norm": 1.0632400314141561,
|
| 1064 |
+
"learning_rate": 7.17880803349623e-05,
|
| 1065 |
+
"loss": 1.0594,
|
| 1066 |
+
"step": 151
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 1.4476190476190476,
|
| 1070 |
+
"grad_norm": 1.384452095634303,
|
| 1071 |
+
"learning_rate": 7.162577181118485e-05,
|
| 1072 |
+
"loss": 1.0747,
|
| 1073 |
+
"step": 152
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 1.457142857142857,
|
| 1077 |
+
"grad_norm": 0.6834451700354741,
|
| 1078 |
+
"learning_rate": 7.146206223210571e-05,
|
| 1079 |
+
"loss": 1.0722,
|
| 1080 |
+
"step": 153
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 1.4666666666666668,
|
| 1084 |
+
"grad_norm": 0.6642752224995787,
|
| 1085 |
+
"learning_rate": 7.129695885023321e-05,
|
| 1086 |
+
"loss": 1.0605,
|
| 1087 |
+
"step": 154
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 1.4761904761904763,
|
| 1091 |
+
"grad_norm": 1.2463789389464905,
|
| 1092 |
+
"learning_rate": 7.113046897982265e-05,
|
| 1093 |
+
"loss": 1.0678,
|
| 1094 |
+
"step": 155
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 1.4857142857142858,
|
| 1098 |
+
"grad_norm": 1.1799339849059132,
|
| 1099 |
+
"learning_rate": 7.09625999965522e-05,
|
| 1100 |
+
"loss": 1.0805,
|
| 1101 |
+
"step": 156
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 1.4952380952380953,
|
| 1105 |
+
"grad_norm": 0.5844684489642771,
|
| 1106 |
+
"learning_rate": 7.079335933719625e-05,
|
| 1107 |
+
"loss": 1.0627,
|
| 1108 |
+
"step": 157
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 1.5047619047619047,
|
| 1112 |
+
"grad_norm": 0.6983703223969,
|
| 1113 |
+
"learning_rate": 7.062275449929587e-05,
|
| 1114 |
+
"loss": 1.0685,
|
| 1115 |
+
"step": 158
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 1.5142857142857142,
|
| 1119 |
+
"grad_norm": 1.0157049520030346,
|
| 1120 |
+
"learning_rate": 7.045079304082667e-05,
|
| 1121 |
+
"loss": 1.057,
|
| 1122 |
+
"step": 159
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 1.5238095238095237,
|
| 1126 |
+
"grad_norm": 1.21522775185842,
|
| 1127 |
+
"learning_rate": 7.027748257986403e-05,
|
| 1128 |
+
"loss": 1.0497,
|
| 1129 |
+
"step": 160
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 1.5333333333333332,
|
| 1133 |
+
"grad_norm": 0.670913402144578,
|
| 1134 |
+
"learning_rate": 7.010283079424553e-05,
|
| 1135 |
+
"loss": 1.0714,
|
| 1136 |
+
"step": 161
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 1.5428571428571427,
|
| 1140 |
+
"grad_norm": 0.5725919893988906,
|
| 1141 |
+
"learning_rate": 6.992684542123094e-05,
|
| 1142 |
+
"loss": 1.0651,
|
| 1143 |
+
"step": 162
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 1.5523809523809524,
|
| 1147 |
+
"grad_norm": 0.5816949338841475,
|
| 1148 |
+
"learning_rate": 6.974953425715926e-05,
|
| 1149 |
+
"loss": 1.0625,
|
| 1150 |
+
"step": 163
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 1.561904761904762,
|
| 1154 |
+
"grad_norm": 0.621239138517028,
|
| 1155 |
+
"learning_rate": 6.957090515710353e-05,
|
| 1156 |
+
"loss": 1.06,
|
| 1157 |
+
"step": 164
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 1.5714285714285714,
|
| 1161 |
+
"grad_norm": 0.8232854555952457,
|
| 1162 |
+
"learning_rate": 6.939096603452269e-05,
|
| 1163 |
+
"loss": 1.0607,
|
| 1164 |
+
"step": 165
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 1.580952380952381,
|
| 1168 |
+
"grad_norm": 0.7980171860515516,
|
| 1169 |
+
"learning_rate": 6.920972486091113e-05,
|
| 1170 |
+
"loss": 1.069,
|
| 1171 |
+
"step": 166
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 1.5904761904761906,
|
| 1175 |
+
"grad_norm": 0.8469950727776042,
|
| 1176 |
+
"learning_rate": 6.902718966544545e-05,
|
| 1177 |
+
"loss": 1.0555,
|
| 1178 |
+
"step": 167
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 1.6,
|
| 1182 |
+
"grad_norm": 1.0917122939746875,
|
| 1183 |
+
"learning_rate": 6.884336853462879e-05,
|
| 1184 |
+
"loss": 1.073,
|
| 1185 |
+
"step": 168
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 1.6095238095238096,
|
| 1189 |
+
"grad_norm": 0.92965912537962,
|
| 1190 |
+
"learning_rate": 6.865826961193261e-05,
|
| 1191 |
+
"loss": 1.0662,
|
| 1192 |
+
"step": 169
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 1.619047619047619,
|
| 1196 |
+
"grad_norm": 0.7708978762494281,
|
| 1197 |
+
"learning_rate": 6.84719010974359e-05,
|
| 1198 |
+
"loss": 1.0524,
|
| 1199 |
+
"step": 170
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 1.6285714285714286,
|
| 1203 |
+
"grad_norm": 0.6756433297794424,
|
| 1204 |
+
"learning_rate": 6.828427124746191e-05,
|
| 1205 |
+
"loss": 1.0692,
|
| 1206 |
+
"step": 171
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 1.638095238095238,
|
| 1210 |
+
"grad_norm": 0.6747029343891224,
|
| 1211 |
+
"learning_rate": 6.80953883742124e-05,
|
| 1212 |
+
"loss": 1.0607,
|
| 1213 |
+
"step": 172
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 1.6476190476190475,
|
| 1217 |
+
"grad_norm": 0.6109245622680719,
|
| 1218 |
+
"learning_rate": 6.790526084539939e-05,
|
| 1219 |
+
"loss": 1.0578,
|
| 1220 |
+
"step": 173
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 1.657142857142857,
|
| 1224 |
+
"grad_norm": 0.7524343677177509,
|
| 1225 |
+
"learning_rate": 6.771389708387448e-05,
|
| 1226 |
+
"loss": 1.0666,
|
| 1227 |
+
"step": 174
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 1.6666666666666665,
|
| 1231 |
+
"grad_norm": 1.1491906497079498,
|
| 1232 |
+
"learning_rate": 6.752130556725567e-05,
|
| 1233 |
+
"loss": 1.0694,
|
| 1234 |
+
"step": 175
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 1.6761904761904762,
|
| 1238 |
+
"grad_norm": 0.8088967378835389,
|
| 1239 |
+
"learning_rate": 6.73274948275518e-05,
|
| 1240 |
+
"loss": 1.0632,
|
| 1241 |
+
"step": 176
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 1.6857142857142857,
|
| 1245 |
+
"grad_norm": 0.4254639452827212,
|
| 1246 |
+
"learning_rate": 6.713247345078465e-05,
|
| 1247 |
+
"loss": 1.0543,
|
| 1248 |
+
"step": 177
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 1.6952380952380952,
|
| 1252 |
+
"grad_norm": 0.5687155529327818,
|
| 1253 |
+
"learning_rate": 6.693625007660845e-05,
|
| 1254 |
+
"loss": 1.0575,
|
| 1255 |
+
"step": 178
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 1.704761904761905,
|
| 1259 |
+
"grad_norm": 0.6667850106047994,
|
| 1260 |
+
"learning_rate": 6.673883339792723e-05,
|
| 1261 |
+
"loss": 1.055,
|
| 1262 |
+
"step": 179
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 1.7142857142857144,
|
| 1266 |
+
"grad_norm": 0.612985384932019,
|
| 1267 |
+
"learning_rate": 6.654023216050963e-05,
|
| 1268 |
+
"loss": 1.071,
|
| 1269 |
+
"step": 180
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 1.723809523809524,
|
| 1273 |
+
"grad_norm": 0.5640012274326677,
|
| 1274 |
+
"learning_rate": 6.634045516260156e-05,
|
| 1275 |
+
"loss": 1.051,
|
| 1276 |
+
"step": 181
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 1.7333333333333334,
|
| 1280 |
+
"grad_norm": 0.499428294662914,
|
| 1281 |
+
"learning_rate": 6.613951125453632e-05,
|
| 1282 |
+
"loss": 1.0768,
|
| 1283 |
+
"step": 182
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 1.7428571428571429,
|
| 1287 |
+
"grad_norm": 0.42771581794966373,
|
| 1288 |
+
"learning_rate": 6.593740933834262e-05,
|
| 1289 |
+
"loss": 1.0566,
|
| 1290 |
+
"step": 183
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 1.7523809523809524,
|
| 1294 |
+
"grad_norm": 0.8182121479285371,
|
| 1295 |
+
"learning_rate": 6.573415836735011e-05,
|
| 1296 |
+
"loss": 1.073,
|
| 1297 |
+
"step": 184
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 1.7619047619047619,
|
| 1301 |
+
"grad_norm": 0.45335952004384406,
|
| 1302 |
+
"learning_rate": 6.552976734579281e-05,
|
| 1303 |
+
"loss": 1.0602,
|
| 1304 |
+
"step": 185
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 1.7714285714285714,
|
| 1308 |
+
"grad_norm": 0.47019401062933164,
|
| 1309 |
+
"learning_rate": 6.53242453284102e-05,
|
| 1310 |
+
"loss": 1.0592,
|
| 1311 |
+
"step": 186
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 1.7809523809523808,
|
| 1315 |
+
"grad_norm": 0.4216198182174328,
|
| 1316 |
+
"learning_rate": 6.511760142004608e-05,
|
| 1317 |
+
"loss": 1.0607,
|
| 1318 |
+
"step": 187
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 1.7904761904761903,
|
| 1322 |
+
"grad_norm": 0.4390063285487437,
|
| 1323 |
+
"learning_rate": 6.49098447752452e-05,
|
| 1324 |
+
"loss": 1.0685,
|
| 1325 |
+
"step": 188
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 1.8,
|
| 1329 |
+
"grad_norm": 0.44096942987470333,
|
| 1330 |
+
"learning_rate": 6.470098459784768e-05,
|
| 1331 |
+
"loss": 1.0566,
|
| 1332 |
+
"step": 189
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 1.8095238095238095,
|
| 1336 |
+
"grad_norm": 0.3569409404531036,
|
| 1337 |
+
"learning_rate": 6.449103014058139e-05,
|
| 1338 |
+
"loss": 1.0588,
|
| 1339 |
+
"step": 190
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 1.819047619047619,
|
| 1343 |
+
"grad_norm": 0.47305630507582075,
|
| 1344 |
+
"learning_rate": 6.427999070465191e-05,
|
| 1345 |
+
"loss": 1.0758,
|
| 1346 |
+
"step": 191
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 1.8285714285714287,
|
| 1350 |
+
"grad_norm": 0.34682771443136134,
|
| 1351 |
+
"learning_rate": 6.406787563933053e-05,
|
| 1352 |
+
"loss": 1.0561,
|
| 1353 |
+
"step": 192
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 1.8380952380952382,
|
| 1357 |
+
"grad_norm": 0.3784058265103603,
|
| 1358 |
+
"learning_rate": 6.385469434154006e-05,
|
| 1359 |
+
"loss": 1.0806,
|
| 1360 |
+
"step": 193
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 1.8476190476190477,
|
| 1364 |
+
"grad_norm": 0.45193884219920866,
|
| 1365 |
+
"learning_rate": 6.364045625543856e-05,
|
| 1366 |
+
"loss": 1.0601,
|
| 1367 |
+
"step": 194
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 1.8571428571428572,
|
| 1371 |
+
"grad_norm": 0.5301378494163913,
|
| 1372 |
+
"learning_rate": 6.342517087200094e-05,
|
| 1373 |
+
"loss": 1.0586,
|
| 1374 |
+
"step": 195
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 1.8666666666666667,
|
| 1378 |
+
"grad_norm": 0.563802211092738,
|
| 1379 |
+
"learning_rate": 6.320884772859845e-05,
|
| 1380 |
+
"loss": 1.0571,
|
| 1381 |
+
"step": 196
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 1.8761904761904762,
|
| 1385 |
+
"grad_norm": 0.5201827508182623,
|
| 1386 |
+
"learning_rate": 6.29914964085763e-05,
|
| 1387 |
+
"loss": 1.0571,
|
| 1388 |
+
"step": 197
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 1.8857142857142857,
|
| 1392 |
+
"grad_norm": 0.4236836095612993,
|
| 1393 |
+
"learning_rate": 6.277312654082886e-05,
|
| 1394 |
+
"loss": 1.0606,
|
| 1395 |
+
"step": 198
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 1.8952380952380952,
|
| 1399 |
+
"grad_norm": 0.38720748352240386,
|
| 1400 |
+
"learning_rate": 6.255374779937344e-05,
|
| 1401 |
+
"loss": 1.0493,
|
| 1402 |
+
"step": 199
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 1.9047619047619047,
|
| 1406 |
+
"grad_norm": 0.42659615830981307,
|
| 1407 |
+
"learning_rate": 6.23333699029214e-05,
|
| 1408 |
+
"loss": 1.0619,
|
| 1409 |
+
"step": 200
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 1.9142857142857141,
|
| 1413 |
+
"grad_norm": 0.6061421220455705,
|
| 1414 |
+
"learning_rate": 6.211200261444774e-05,
|
| 1415 |
+
"loss": 1.0541,
|
| 1416 |
+
"step": 201
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 1.9238095238095239,
|
| 1420 |
+
"grad_norm": 0.797781403199275,
|
| 1421 |
+
"learning_rate": 6.188965574075863e-05,
|
| 1422 |
+
"loss": 1.0559,
|
| 1423 |
+
"step": 202
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 1.9333333333333333,
|
| 1427 |
+
"grad_norm": 0.8282451159457771,
|
| 1428 |
+
"learning_rate": 6.166633913205684e-05,
|
| 1429 |
+
"loss": 1.0545,
|
| 1430 |
+
"step": 203
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 1.9428571428571428,
|
| 1434 |
+
"grad_norm": 0.9009492897590493,
|
| 1435 |
+
"learning_rate": 6.144206268150549e-05,
|
| 1436 |
+
"loss": 1.0586,
|
| 1437 |
+
"step": 204
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 1.9523809523809523,
|
| 1441 |
+
"grad_norm": 1.0020870163569926,
|
| 1442 |
+
"learning_rate": 6.12168363247897e-05,
|
| 1443 |
+
"loss": 1.0457,
|
| 1444 |
+
"step": 205
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 1.961904761904762,
|
| 1448 |
+
"grad_norm": 1.1260023046138727,
|
| 1449 |
+
"learning_rate": 6.0990670039676416e-05,
|
| 1450 |
+
"loss": 1.054,
|
| 1451 |
+
"step": 206
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 1.9714285714285715,
|
| 1455 |
+
"grad_norm": 0.7820493480578619,
|
| 1456 |
+
"learning_rate": 6.0763573845572434e-05,
|
| 1457 |
+
"loss": 1.0674,
|
| 1458 |
+
"step": 207
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 1.980952380952381,
|
| 1462 |
+
"grad_norm": 0.47269444067540217,
|
| 1463 |
+
"learning_rate": 6.053555780308049e-05,
|
| 1464 |
+
"loss": 1.0469,
|
| 1465 |
+
"step": 208
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 1.9904761904761905,
|
| 1469 |
+
"grad_norm": 0.30952011756676745,
|
| 1470 |
+
"learning_rate": 6.03066320135536e-05,
|
| 1471 |
+
"loss": 1.0457,
|
| 1472 |
+
"step": 209
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 2.0,
|
| 1476 |
+
"grad_norm": 0.5189001553144519,
|
| 1477 |
+
"learning_rate": 6.0076806618647545e-05,
|
| 1478 |
+
"loss": 1.0549,
|
| 1479 |
+
"step": 210
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 2.0095238095238095,
|
| 1483 |
+
"grad_norm": 0.7439078466919193,
|
| 1484 |
+
"learning_rate": 5.984609179987155e-05,
|
| 1485 |
+
"loss": 1.0304,
|
| 1486 |
+
"step": 211
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 2.019047619047619,
|
| 1490 |
+
"grad_norm": 0.8444989260384058,
|
| 1491 |
+
"learning_rate": 5.961449777813727e-05,
|
| 1492 |
+
"loss": 1.0483,
|
| 1493 |
+
"step": 212
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 2.0285714285714285,
|
| 1497 |
+
"grad_norm": 0.85105512066596,
|
| 1498 |
+
"learning_rate": 5.9382034813306014e-05,
|
| 1499 |
+
"loss": 1.0192,
|
| 1500 |
+
"step": 213
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 2.038095238095238,
|
| 1504 |
+
"grad_norm": 0.8717088777508358,
|
| 1505 |
+
"learning_rate": 5.914871320373417e-05,
|
| 1506 |
+
"loss": 1.0262,
|
| 1507 |
+
"step": 214
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 2.0476190476190474,
|
| 1511 |
+
"grad_norm": 0.8473077460098071,
|
| 1512 |
+
"learning_rate": 5.891454328581702e-05,
|
| 1513 |
+
"loss": 1.0354,
|
| 1514 |
+
"step": 215
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 2.057142857142857,
|
| 1518 |
+
"grad_norm": 0.7459552028603442,
|
| 1519 |
+
"learning_rate": 5.8679535433530756e-05,
|
| 1520 |
+
"loss": 1.0151,
|
| 1521 |
+
"step": 216
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 2.066666666666667,
|
| 1525 |
+
"grad_norm": 0.6103690018974712,
|
| 1526 |
+
"learning_rate": 5.844370005797304e-05,
|
| 1527 |
+
"loss": 1.0169,
|
| 1528 |
+
"step": 217
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 2.0761904761904764,
|
| 1532 |
+
"grad_norm": 0.4922703155453895,
|
| 1533 |
+
"learning_rate": 5.820704760690161e-05,
|
| 1534 |
+
"loss": 1.0189,
|
| 1535 |
+
"step": 218
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 2.085714285714286,
|
| 1539 |
+
"grad_norm": 0.5537849831993843,
|
| 1540 |
+
"learning_rate": 5.796958856427155e-05,
|
| 1541 |
+
"loss": 1.0177,
|
| 1542 |
+
"step": 219
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 2.0952380952380953,
|
| 1546 |
+
"grad_norm": 0.5404302742305147,
|
| 1547 |
+
"learning_rate": 5.7731333449770833e-05,
|
| 1548 |
+
"loss": 1.0095,
|
| 1549 |
+
"step": 220
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 2.104761904761905,
|
| 1553 |
+
"grad_norm": 0.4534851743612291,
|
| 1554 |
+
"learning_rate": 5.7492292818354224e-05,
|
| 1555 |
+
"loss": 1.0187,
|
| 1556 |
+
"step": 221
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 2.1142857142857143,
|
| 1560 |
+
"grad_norm": 0.3065835944952437,
|
| 1561 |
+
"learning_rate": 5.725247725977573e-05,
|
| 1562 |
+
"loss": 1.0034,
|
| 1563 |
+
"step": 222
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 2.123809523809524,
|
| 1567 |
+
"grad_norm": 0.4746500125323341,
|
| 1568 |
+
"learning_rate": 5.7011897398119486e-05,
|
| 1569 |
+
"loss": 1.0156,
|
| 1570 |
+
"step": 223
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 2.1333333333333333,
|
| 1574 |
+
"grad_norm": 0.5364964068164462,
|
| 1575 |
+
"learning_rate": 5.6770563891329e-05,
|
| 1576 |
+
"loss": 1.023,
|
| 1577 |
+
"step": 224
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 2.142857142857143,
|
| 1581 |
+
"grad_norm": 0.39435975272411355,
|
| 1582 |
+
"learning_rate": 5.652848743073513e-05,
|
| 1583 |
+
"loss": 1.0161,
|
| 1584 |
+
"step": 225
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 2.1523809523809523,
|
| 1588 |
+
"grad_norm": 0.2917840449905482,
|
| 1589 |
+
"learning_rate": 5.628567874058235e-05,
|
| 1590 |
+
"loss": 1.0176,
|
| 1591 |
+
"step": 226
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 2.1619047619047618,
|
| 1595 |
+
"grad_norm": 0.4309558938275066,
|
| 1596 |
+
"learning_rate": 5.6042148577553665e-05,
|
| 1597 |
+
"loss": 1.0189,
|
| 1598 |
+
"step": 227
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 2.1714285714285713,
|
| 1602 |
+
"grad_norm": 0.4049822234252623,
|
| 1603 |
+
"learning_rate": 5.5797907730294123e-05,
|
| 1604 |
+
"loss": 1.0079,
|
| 1605 |
+
"step": 228
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 2.1809523809523808,
|
| 1609 |
+
"grad_norm": 0.2826706608436985,
|
| 1610 |
+
"learning_rate": 5.555296701893284e-05,
|
| 1611 |
+
"loss": 1.0161,
|
| 1612 |
+
"step": 229
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 2.1904761904761907,
|
| 1616 |
+
"grad_norm": 0.4502126269593606,
|
| 1617 |
+
"learning_rate": 5.5307337294603595e-05,
|
| 1618 |
+
"loss": 1.0196,
|
| 1619 |
+
"step": 230
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 2.2,
|
| 1623 |
+
"grad_norm": 0.41114563221017636,
|
| 1624 |
+
"learning_rate": 5.506102943896426e-05,
|
| 1625 |
+
"loss": 1.016,
|
| 1626 |
+
"step": 231
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 2.2095238095238097,
|
| 1630 |
+
"grad_norm": 0.313503216833648,
|
| 1631 |
+
"learning_rate": 5.481405436371459e-05,
|
| 1632 |
+
"loss": 1.0089,
|
| 1633 |
+
"step": 232
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 2.219047619047619,
|
| 1637 |
+
"grad_norm": 0.36100525990234567,
|
| 1638 |
+
"learning_rate": 5.45664230101129e-05,
|
| 1639 |
+
"loss": 1.0031,
|
| 1640 |
+
"step": 233
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 2.2285714285714286,
|
| 1644 |
+
"grad_norm": 0.3757564707600635,
|
| 1645 |
+
"learning_rate": 5.431814634849131e-05,
|
| 1646 |
+
"loss": 1.0144,
|
| 1647 |
+
"step": 234
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 2.238095238095238,
|
| 1651 |
+
"grad_norm": 0.3578831855885671,
|
| 1652 |
+
"learning_rate": 5.40692353777698e-05,
|
| 1653 |
+
"loss": 1.0291,
|
| 1654 |
+
"step": 235
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 2.2476190476190476,
|
| 1658 |
+
"grad_norm": 0.279381430603173,
|
| 1659 |
+
"learning_rate": 5.38197011249689e-05,
|
| 1660 |
+
"loss": 1.0038,
|
| 1661 |
+
"step": 236
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 2.257142857142857,
|
| 1665 |
+
"grad_norm": 0.31128636254168296,
|
| 1666 |
+
"learning_rate": 5.356955464472121e-05,
|
| 1667 |
+
"loss": 1.0132,
|
| 1668 |
+
"step": 237
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 2.2666666666666666,
|
| 1672 |
+
"grad_norm": 0.28395461762897617,
|
| 1673 |
+
"learning_rate": 5.331880701878165e-05,
|
| 1674 |
+
"loss": 1.0116,
|
| 1675 |
+
"step": 238
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 2.276190476190476,
|
| 1679 |
+
"grad_norm": 0.27391953654101664,
|
| 1680 |
+
"learning_rate": 5.3067469355536525e-05,
|
| 1681 |
+
"loss": 1.0183,
|
| 1682 |
+
"step": 239
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 2.2857142857142856,
|
| 1686 |
+
"grad_norm": 0.22816906449119856,
|
| 1687 |
+
"learning_rate": 5.2815552789511426e-05,
|
| 1688 |
+
"loss": 1.0031,
|
| 1689 |
+
"step": 240
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 2.295238095238095,
|
| 1693 |
+
"grad_norm": 0.2370061459930583,
|
| 1694 |
+
"learning_rate": 5.256306848087796e-05,
|
| 1695 |
+
"loss": 1.0243,
|
| 1696 |
+
"step": 241
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 2.3047619047619046,
|
| 1700 |
+
"grad_norm": 0.2679568465187226,
|
| 1701 |
+
"learning_rate": 5.2310027614959316e-05,
|
| 1702 |
+
"loss": 1.003,
|
| 1703 |
+
"step": 242
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 2.314285714285714,
|
| 1707 |
+
"grad_norm": 0.26159769690103984,
|
| 1708 |
+
"learning_rate": 5.20564414017348e-05,
|
| 1709 |
+
"loss": 1.0286,
|
| 1710 |
+
"step": 243
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 2.323809523809524,
|
| 1714 |
+
"grad_norm": 0.2277499325426611,
|
| 1715 |
+
"learning_rate": 5.1802321075343135e-05,
|
| 1716 |
+
"loss": 1.0158,
|
| 1717 |
+
"step": 244
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 2.3333333333333335,
|
| 1721 |
+
"grad_norm": 0.2765055373878013,
|
| 1722 |
+
"learning_rate": 5.1547677893584846e-05,
|
| 1723 |
+
"loss": 1.0176,
|
| 1724 |
+
"step": 245
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 2.342857142857143,
|
| 1728 |
+
"grad_norm": 0.2707671528392851,
|
| 1729 |
+
"learning_rate": 5.129252313742353e-05,
|
| 1730 |
+
"loss": 1.012,
|
| 1731 |
+
"step": 246
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 2.3523809523809525,
|
| 1735 |
+
"grad_norm": 0.24663987368382112,
|
| 1736 |
+
"learning_rate": 5.103686811048603e-05,
|
| 1737 |
+
"loss": 1.005,
|
| 1738 |
+
"step": 247
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 2.361904761904762,
|
| 1742 |
+
"grad_norm": 0.3119601934853684,
|
| 1743 |
+
"learning_rate": 5.078072413856174e-05,
|
| 1744 |
+
"loss": 0.9982,
|
| 1745 |
+
"step": 248
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 2.3714285714285714,
|
| 1749 |
+
"grad_norm": 0.28361246368044496,
|
| 1750 |
+
"learning_rate": 5.052410256910085e-05,
|
| 1751 |
+
"loss": 1.0106,
|
| 1752 |
+
"step": 249
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 2.380952380952381,
|
| 1756 |
+
"grad_norm": 0.210984716029959,
|
| 1757 |
+
"learning_rate": 5.026701477071161e-05,
|
| 1758 |
+
"loss": 1.0249,
|
| 1759 |
+
"step": 250
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 2.3904761904761904,
|
| 1763 |
+
"grad_norm": 0.2639314428580076,
|
| 1764 |
+
"learning_rate": 5.00094721326567e-05,
|
| 1765 |
+
"loss": 1.0133,
|
| 1766 |
+
"step": 251
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 2.4,
|
| 1770 |
+
"grad_norm": 0.2077340280233697,
|
| 1771 |
+
"learning_rate": 4.9751486064348695e-05,
|
| 1772 |
+
"loss": 1.0166,
|
| 1773 |
+
"step": 252
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 2.4095238095238094,
|
| 1777 |
+
"grad_norm": 0.20906067378182863,
|
| 1778 |
+
"learning_rate": 4.9493067994844606e-05,
|
| 1779 |
+
"loss": 1.0017,
|
| 1780 |
+
"step": 253
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 2.419047619047619,
|
| 1784 |
+
"grad_norm": 0.31463242576457345,
|
| 1785 |
+
"learning_rate": 4.9234229372339525e-05,
|
| 1786 |
+
"loss": 1.0167,
|
| 1787 |
+
"step": 254
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 2.4285714285714284,
|
| 1791 |
+
"grad_norm": 0.23613151956098302,
|
| 1792 |
+
"learning_rate": 4.897498166365953e-05,
|
| 1793 |
+
"loss": 1.0148,
|
| 1794 |
+
"step": 255
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 2.4380952380952383,
|
| 1798 |
+
"grad_norm": 0.2809994241715782,
|
| 1799 |
+
"learning_rate": 4.8715336353753616e-05,
|
| 1800 |
+
"loss": 1.0168,
|
| 1801 |
+
"step": 256
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 2.447619047619048,
|
| 1805 |
+
"grad_norm": 0.28140513046492216,
|
| 1806 |
+
"learning_rate": 4.845530494518498e-05,
|
| 1807 |
+
"loss": 1.0174,
|
| 1808 |
+
"step": 257
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 2.4571428571428573,
|
| 1812 |
+
"grad_norm": 0.26072644777277915,
|
| 1813 |
+
"learning_rate": 4.819489895762135e-05,
|
| 1814 |
+
"loss": 1.003,
|
| 1815 |
+
"step": 258
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 2.466666666666667,
|
| 1819 |
+
"grad_norm": 0.24673982217299864,
|
| 1820 |
+
"learning_rate": 4.7934129927324717e-05,
|
| 1821 |
+
"loss": 1.0098,
|
| 1822 |
+
"step": 259
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 2.4761904761904763,
|
| 1826 |
+
"grad_norm": 0.2336437136943171,
|
| 1827 |
+
"learning_rate": 4.7673009406640264e-05,
|
| 1828 |
+
"loss": 1.0186,
|
| 1829 |
+
"step": 260
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 2.4857142857142858,
|
| 1833 |
+
"grad_norm": 0.29231870944215277,
|
| 1834 |
+
"learning_rate": 4.741154896348458e-05,
|
| 1835 |
+
"loss": 1.0068,
|
| 1836 |
+
"step": 261
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 2.4952380952380953,
|
| 1840 |
+
"grad_norm": 0.2911376621948982,
|
| 1841 |
+
"learning_rate": 4.714976018083315e-05,
|
| 1842 |
+
"loss": 1.0224,
|
| 1843 |
+
"step": 262
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 2.5047619047619047,
|
| 1847 |
+
"grad_norm": 0.3338692874924189,
|
| 1848 |
+
"learning_rate": 4.6887654656207255e-05,
|
| 1849 |
+
"loss": 1.0195,
|
| 1850 |
+
"step": 263
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 2.5142857142857142,
|
| 1854 |
+
"grad_norm": 0.2541097184148386,
|
| 1855 |
+
"learning_rate": 4.66252440011602e-05,
|
| 1856 |
+
"loss": 1.003,
|
| 1857 |
+
"step": 264
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 2.5238095238095237,
|
| 1861 |
+
"grad_norm": 0.24770604612278468,
|
| 1862 |
+
"learning_rate": 4.6362539840762886e-05,
|
| 1863 |
+
"loss": 1.0223,
|
| 1864 |
+
"step": 265
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 2.533333333333333,
|
| 1868 |
+
"grad_norm": 0.2191655449635504,
|
| 1869 |
+
"learning_rate": 4.60995538130888e-05,
|
| 1870 |
+
"loss": 1.0145,
|
| 1871 |
+
"step": 266
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 2.5428571428571427,
|
| 1875 |
+
"grad_norm": 0.2790975433249823,
|
| 1876 |
+
"learning_rate": 4.5836297568698475e-05,
|
| 1877 |
+
"loss": 1.0126,
|
| 1878 |
+
"step": 267
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 2.552380952380952,
|
| 1882 |
+
"grad_norm": 0.28428156686414346,
|
| 1883 |
+
"learning_rate": 4.557278277012329e-05,
|
| 1884 |
+
"loss": 1.0295,
|
| 1885 |
+
"step": 268
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 2.5619047619047617,
|
| 1889 |
+
"grad_norm": 0.3233959922060247,
|
| 1890 |
+
"learning_rate": 4.5309021091348885e-05,
|
| 1891 |
+
"loss": 1.0174,
|
| 1892 |
+
"step": 269
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 2.571428571428571,
|
| 1896 |
+
"grad_norm": 0.2956250139389734,
|
| 1897 |
+
"learning_rate": 4.504502421729795e-05,
|
| 1898 |
+
"loss": 1.0215,
|
| 1899 |
+
"step": 270
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 2.580952380952381,
|
| 1903 |
+
"grad_norm": 0.2009944046883891,
|
| 1904 |
+
"learning_rate": 4.478080384331255e-05,
|
| 1905 |
+
"loss": 1.011,
|
| 1906 |
+
"step": 271
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 2.5904761904761906,
|
| 1910 |
+
"grad_norm": 0.18530962666368314,
|
| 1911 |
+
"learning_rate": 4.4516371674636074e-05,
|
| 1912 |
+
"loss": 1.0181,
|
| 1913 |
+
"step": 272
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 2.6,
|
| 1917 |
+
"grad_norm": 0.19572463370442308,
|
| 1918 |
+
"learning_rate": 4.425173942589462e-05,
|
| 1919 |
+
"loss": 1.0222,
|
| 1920 |
+
"step": 273
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 2.6095238095238096,
|
| 1924 |
+
"grad_norm": 0.20133706322572675,
|
| 1925 |
+
"learning_rate": 4.398691882057804e-05,
|
| 1926 |
+
"loss": 1.0238,
|
| 1927 |
+
"step": 274
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 2.619047619047619,
|
| 1931 |
+
"grad_norm": 0.1696896722933609,
|
| 1932 |
+
"learning_rate": 4.372192159052058e-05,
|
| 1933 |
+
"loss": 1.0121,
|
| 1934 |
+
"step": 275
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 2.6285714285714286,
|
| 1938 |
+
"grad_norm": 0.5599358185228015,
|
| 1939 |
+
"learning_rate": 4.3456759475381183e-05,
|
| 1940 |
+
"loss": 1.0468,
|
| 1941 |
+
"step": 276
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 2.638095238095238,
|
| 1945 |
+
"grad_norm": 0.20466020884323685,
|
| 1946 |
+
"learning_rate": 4.3191444222123326e-05,
|
| 1947 |
+
"loss": 1.0211,
|
| 1948 |
+
"step": 277
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 2.6476190476190475,
|
| 1952 |
+
"grad_norm": 0.18418573847215977,
|
| 1953 |
+
"learning_rate": 4.2925987584494706e-05,
|
| 1954 |
+
"loss": 1.013,
|
| 1955 |
+
"step": 278
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 2.657142857142857,
|
| 1959 |
+
"grad_norm": 0.20944834576237945,
|
| 1960 |
+
"learning_rate": 4.26604013225065e-05,
|
| 1961 |
+
"loss": 1.0165,
|
| 1962 |
+
"step": 279
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 2.6666666666666665,
|
| 1966 |
+
"grad_norm": 0.17888267966598845,
|
| 1967 |
+
"learning_rate": 4.239469720191234e-05,
|
| 1968 |
+
"loss": 1.0285,
|
| 1969 |
+
"step": 280
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 2.6761904761904765,
|
| 1973 |
+
"grad_norm": 0.24111408627710695,
|
| 1974 |
+
"learning_rate": 4.2128886993687145e-05,
|
| 1975 |
+
"loss": 1.011,
|
| 1976 |
+
"step": 281
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 2.685714285714286,
|
| 1980 |
+
"grad_norm": 0.25283589971223985,
|
| 1981 |
+
"learning_rate": 4.186298247350567e-05,
|
| 1982 |
+
"loss": 1.0021,
|
| 1983 |
+
"step": 282
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 2.6952380952380954,
|
| 1987 |
+
"grad_norm": 0.19532012129248275,
|
| 1988 |
+
"learning_rate": 4.159699542122071e-05,
|
| 1989 |
+
"loss": 1.0113,
|
| 1990 |
+
"step": 283
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 2.704761904761905,
|
| 1994 |
+
"grad_norm": 0.2297669069644532,
|
| 1995 |
+
"learning_rate": 4.133093762034137e-05,
|
| 1996 |
+
"loss": 1.0032,
|
| 1997 |
+
"step": 284
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 2.7142857142857144,
|
| 2001 |
+
"grad_norm": 0.2465633017960609,
|
| 2002 |
+
"learning_rate": 4.1064820857511e-05,
|
| 2003 |
+
"loss": 0.9943,
|
| 2004 |
+
"step": 285
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 2.723809523809524,
|
| 2008 |
+
"grad_norm": 0.24245129662503742,
|
| 2009 |
+
"learning_rate": 4.079865692198499e-05,
|
| 2010 |
+
"loss": 1.0164,
|
| 2011 |
+
"step": 286
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 2.7333333333333334,
|
| 2015 |
+
"grad_norm": 0.22833439764975844,
|
| 2016 |
+
"learning_rate": 4.053245760510856e-05,
|
| 2017 |
+
"loss": 1.0161,
|
| 2018 |
+
"step": 287
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 2.742857142857143,
|
| 2022 |
+
"grad_norm": 0.22230175915043607,
|
| 2023 |
+
"learning_rate": 4.026623469979436e-05,
|
| 2024 |
+
"loss": 1.0091,
|
| 2025 |
+
"step": 288
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 2.7523809523809524,
|
| 2029 |
+
"grad_norm": 0.2781762656155117,
|
| 2030 |
+
"learning_rate": 4e-05,
|
| 2031 |
+
"loss": 1.0017,
|
| 2032 |
+
"step": 289
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 2.761904761904762,
|
| 2036 |
+
"grad_norm": 0.2864072028925091,
|
| 2037 |
+
"learning_rate": 3.9733765300205654e-05,
|
| 2038 |
+
"loss": 1.0075,
|
| 2039 |
+
"step": 290
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 2.7714285714285714,
|
| 2043 |
+
"grad_norm": 0.22898372701526665,
|
| 2044 |
+
"learning_rate": 3.946754239489146e-05,
|
| 2045 |
+
"loss": 1.0122,
|
| 2046 |
+
"step": 291
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 2.780952380952381,
|
| 2050 |
+
"grad_norm": 0.21256356082358066,
|
| 2051 |
+
"learning_rate": 3.9201343078015026e-05,
|
| 2052 |
+
"loss": 1.008,
|
| 2053 |
+
"step": 292
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 2.7904761904761903,
|
| 2057 |
+
"grad_norm": 0.2818845650719653,
|
| 2058 |
+
"learning_rate": 3.8935179142489016e-05,
|
| 2059 |
+
"loss": 0.9986,
|
| 2060 |
+
"step": 293
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 2.8,
|
| 2064 |
+
"grad_norm": 0.23141186471770828,
|
| 2065 |
+
"learning_rate": 3.866906237965865e-05,
|
| 2066 |
+
"loss": 1.0165,
|
| 2067 |
+
"step": 294
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 2.8095238095238093,
|
| 2071 |
+
"grad_norm": 0.23784103111468072,
|
| 2072 |
+
"learning_rate": 3.840300457877931e-05,
|
| 2073 |
+
"loss": 1.0226,
|
| 2074 |
+
"step": 295
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 2.819047619047619,
|
| 2078 |
+
"grad_norm": 0.20078316045157482,
|
| 2079 |
+
"learning_rate": 3.813701752649435e-05,
|
| 2080 |
+
"loss": 1.0015,
|
| 2081 |
+
"step": 296
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 2.8285714285714287,
|
| 2085 |
+
"grad_norm": 0.24338393846551132,
|
| 2086 |
+
"learning_rate": 3.787111300631287e-05,
|
| 2087 |
+
"loss": 1.0067,
|
| 2088 |
+
"step": 297
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 2.8380952380952382,
|
| 2092 |
+
"grad_norm": 0.20074428570812403,
|
| 2093 |
+
"learning_rate": 3.7605302798087686e-05,
|
| 2094 |
+
"loss": 1.0122,
|
| 2095 |
+
"step": 298
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 2.8476190476190477,
|
| 2099 |
+
"grad_norm": 0.20714128390839137,
|
| 2100 |
+
"learning_rate": 3.7339598677493515e-05,
|
| 2101 |
+
"loss": 1.0035,
|
| 2102 |
+
"step": 299
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 2.857142857142857,
|
| 2106 |
+
"grad_norm": 0.17387463464530994,
|
| 2107 |
+
"learning_rate": 3.7074012415505294e-05,
|
| 2108 |
+
"loss": 1.0106,
|
| 2109 |
+
"step": 300
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 2.8666666666666667,
|
| 2113 |
+
"grad_norm": 0.19763448671477987,
|
| 2114 |
+
"learning_rate": 3.6808555777876673e-05,
|
| 2115 |
+
"loss": 1.0071,
|
| 2116 |
+
"step": 301
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 2.876190476190476,
|
| 2120 |
+
"grad_norm": 0.193551861844252,
|
| 2121 |
+
"learning_rate": 3.654324052461883e-05,
|
| 2122 |
+
"loss": 1.0149,
|
| 2123 |
+
"step": 302
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 2.8857142857142857,
|
| 2127 |
+
"grad_norm": 0.20465104040557303,
|
| 2128 |
+
"learning_rate": 3.6278078409479424e-05,
|
| 2129 |
+
"loss": 1.0132,
|
| 2130 |
+
"step": 303
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 2.895238095238095,
|
| 2134 |
+
"grad_norm": 0.1858868750629034,
|
| 2135 |
+
"learning_rate": 3.6013081179421965e-05,
|
| 2136 |
+
"loss": 1.0113,
|
| 2137 |
+
"step": 304
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 2.9047619047619047,
|
| 2141 |
+
"grad_norm": 0.19231783181407783,
|
| 2142 |
+
"learning_rate": 3.5748260574105395e-05,
|
| 2143 |
+
"loss": 1.0223,
|
| 2144 |
+
"step": 305
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 2.914285714285714,
|
| 2148 |
+
"grad_norm": 0.20811259594067136,
|
| 2149 |
+
"learning_rate": 3.548362832536393e-05,
|
| 2150 |
+
"loss": 1.031,
|
| 2151 |
+
"step": 306
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 2.923809523809524,
|
| 2155 |
+
"grad_norm": 0.17374784949143135,
|
| 2156 |
+
"learning_rate": 3.5219196156687454e-05,
|
| 2157 |
+
"loss": 1.0143,
|
| 2158 |
+
"step": 307
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 2.9333333333333336,
|
| 2162 |
+
"grad_norm": 0.18941317471635122,
|
| 2163 |
+
"learning_rate": 3.495497578270206e-05,
|
| 2164 |
+
"loss": 1.0067,
|
| 2165 |
+
"step": 308
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 2.942857142857143,
|
| 2169 |
+
"grad_norm": 0.2175969420603483,
|
| 2170 |
+
"learning_rate": 3.469097890865113e-05,
|
| 2171 |
+
"loss": 1.0149,
|
| 2172 |
+
"step": 309
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 2.9523809523809526,
|
| 2176 |
+
"grad_norm": 0.20305721352150802,
|
| 2177 |
+
"learning_rate": 3.442721722987673e-05,
|
| 2178 |
+
"loss": 1.0148,
|
| 2179 |
+
"step": 310
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 2.961904761904762,
|
| 2183 |
+
"grad_norm": 0.1735412980023502,
|
| 2184 |
+
"learning_rate": 3.416370243130154e-05,
|
| 2185 |
+
"loss": 1.0308,
|
| 2186 |
+
"step": 311
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 2.9714285714285715,
|
| 2190 |
+
"grad_norm": 0.23287418335200405,
|
| 2191 |
+
"learning_rate": 3.390044618691121e-05,
|
| 2192 |
+
"loss": 1.0181,
|
| 2193 |
+
"step": 312
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 2.980952380952381,
|
| 2197 |
+
"grad_norm": 0.1731215573089734,
|
| 2198 |
+
"learning_rate": 3.363746015923713e-05,
|
| 2199 |
+
"loss": 1.0141,
|
| 2200 |
+
"step": 313
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 2.9904761904761905,
|
| 2204 |
+
"grad_norm": 0.36664679839921727,
|
| 2205 |
+
"learning_rate": 3.337475599883981e-05,
|
| 2206 |
+
"loss": 1.0252,
|
| 2207 |
+
"step": 314
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 3.0,
|
| 2211 |
+
"grad_norm": 0.20808906149822945,
|
| 2212 |
+
"learning_rate": 3.3112345343792765e-05,
|
| 2213 |
+
"loss": 1.0094,
|
| 2214 |
+
"step": 315
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 3.0095238095238095,
|
| 2218 |
+
"grad_norm": 0.2386737684696705,
|
| 2219 |
+
"learning_rate": 3.285023981916687e-05,
|
| 2220 |
+
"loss": 0.985,
|
| 2221 |
+
"step": 316
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 3.019047619047619,
|
| 2225 |
+
"grad_norm": 0.21859428661149227,
|
| 2226 |
+
"learning_rate": 3.2588451036515435e-05,
|
| 2227 |
+
"loss": 0.9875,
|
| 2228 |
+
"step": 317
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 3.0285714285714285,
|
| 2232 |
+
"grad_norm": 0.21686946170692808,
|
| 2233 |
+
"learning_rate": 3.2326990593359756e-05,
|
| 2234 |
+
"loss": 0.9648,
|
| 2235 |
+
"step": 318
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 3.038095238095238,
|
| 2239 |
+
"grad_norm": 0.23585519844179503,
|
| 2240 |
+
"learning_rate": 3.206587007267528e-05,
|
| 2241 |
+
"loss": 0.9804,
|
| 2242 |
+
"step": 319
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 3.0476190476190474,
|
| 2246 |
+
"grad_norm": 0.23471049638102,
|
| 2247 |
+
"learning_rate": 3.1805101042378665e-05,
|
| 2248 |
+
"loss": 0.9764,
|
| 2249 |
+
"step": 320
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 3.057142857142857,
|
| 2253 |
+
"grad_norm": 0.24601337714676638,
|
| 2254 |
+
"learning_rate": 3.154469505481503e-05,
|
| 2255 |
+
"loss": 0.9787,
|
| 2256 |
+
"step": 321
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 3.066666666666667,
|
| 2260 |
+
"grad_norm": 0.24680145517708438,
|
| 2261 |
+
"learning_rate": 3.128466364624638e-05,
|
| 2262 |
+
"loss": 0.9751,
|
| 2263 |
+
"step": 322
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 3.0761904761904764,
|
| 2267 |
+
"grad_norm": 0.20266157663662596,
|
| 2268 |
+
"learning_rate": 3.1025018336340484e-05,
|
| 2269 |
+
"loss": 0.9786,
|
| 2270 |
+
"step": 323
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 3.085714285714286,
|
| 2274 |
+
"grad_norm": 0.23172246726109874,
|
| 2275 |
+
"learning_rate": 3.076577062766049e-05,
|
| 2276 |
+
"loss": 0.9797,
|
| 2277 |
+
"step": 324
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 3.0952380952380953,
|
| 2281 |
+
"grad_norm": 0.20530652182231443,
|
| 2282 |
+
"learning_rate": 3.0506932005155407e-05,
|
| 2283 |
+
"loss": 0.9776,
|
| 2284 |
+
"step": 325
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 3.104761904761905,
|
| 2288 |
+
"grad_norm": 0.19541728346574888,
|
| 2289 |
+
"learning_rate": 3.024851393565132e-05,
|
| 2290 |
+
"loss": 0.9747,
|
| 2291 |
+
"step": 326
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 3.1142857142857143,
|
| 2295 |
+
"grad_norm": 0.24999279824271964,
|
| 2296 |
+
"learning_rate": 2.999052786734331e-05,
|
| 2297 |
+
"loss": 0.988,
|
| 2298 |
+
"step": 327
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 3.123809523809524,
|
| 2302 |
+
"grad_norm": 0.22057462842837305,
|
| 2303 |
+
"learning_rate": 2.9732985229288397e-05,
|
| 2304 |
+
"loss": 0.9758,
|
| 2305 |
+
"step": 328
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 3.1333333333333333,
|
| 2309 |
+
"grad_norm": 0.21976681898897724,
|
| 2310 |
+
"learning_rate": 2.9475897430899157e-05,
|
| 2311 |
+
"loss": 0.9778,
|
| 2312 |
+
"step": 329
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 3.142857142857143,
|
| 2316 |
+
"grad_norm": 0.22563434482551034,
|
| 2317 |
+
"learning_rate": 2.921927586143827e-05,
|
| 2318 |
+
"loss": 0.989,
|
| 2319 |
+
"step": 330
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 3.1523809523809523,
|
| 2323 |
+
"grad_norm": 0.1951899724158065,
|
| 2324 |
+
"learning_rate": 2.8963131889513986e-05,
|
| 2325 |
+
"loss": 0.9723,
|
| 2326 |
+
"step": 331
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 3.1619047619047618,
|
| 2330 |
+
"grad_norm": 0.31930450117417375,
|
| 2331 |
+
"learning_rate": 2.870747686257649e-05,
|
| 2332 |
+
"loss": 0.9743,
|
| 2333 |
+
"step": 332
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 3.1714285714285713,
|
| 2337 |
+
"grad_norm": 0.24347478004659553,
|
| 2338 |
+
"learning_rate": 2.845232210641517e-05,
|
| 2339 |
+
"loss": 0.9658,
|
| 2340 |
+
"step": 333
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 3.1809523809523808,
|
| 2344 |
+
"grad_norm": 0.22607819816932653,
|
| 2345 |
+
"learning_rate": 2.8197678924656886e-05,
|
| 2346 |
+
"loss": 0.9666,
|
| 2347 |
+
"step": 334
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 3.1904761904761907,
|
| 2351 |
+
"grad_norm": 0.26083022022974084,
|
| 2352 |
+
"learning_rate": 2.7943558598265218e-05,
|
| 2353 |
+
"loss": 0.9875,
|
| 2354 |
+
"step": 335
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 3.2,
|
| 2358 |
+
"grad_norm": 0.20430972542099413,
|
| 2359 |
+
"learning_rate": 2.7689972385040697e-05,
|
| 2360 |
+
"loss": 0.9817,
|
| 2361 |
+
"step": 336
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 3.2095238095238097,
|
| 2365 |
+
"grad_norm": 0.2478055729854012,
|
| 2366 |
+
"learning_rate": 2.743693151912206e-05,
|
| 2367 |
+
"loss": 0.9833,
|
| 2368 |
+
"step": 337
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 3.219047619047619,
|
| 2372 |
+
"grad_norm": 0.20670211431468252,
|
| 2373 |
+
"learning_rate": 2.718444721048859e-05,
|
| 2374 |
+
"loss": 0.9768,
|
| 2375 |
+
"step": 338
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 3.2285714285714286,
|
| 2379 |
+
"grad_norm": 0.258133896429674,
|
| 2380 |
+
"learning_rate": 2.693253064446348e-05,
|
| 2381 |
+
"loss": 0.9663,
|
| 2382 |
+
"step": 339
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 3.238095238095238,
|
| 2386 |
+
"grad_norm": 0.22888650123481227,
|
| 2387 |
+
"learning_rate": 2.6681192981218348e-05,
|
| 2388 |
+
"loss": 0.9815,
|
| 2389 |
+
"step": 340
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 3.2476190476190476,
|
| 2393 |
+
"grad_norm": 0.2052454770699904,
|
| 2394 |
+
"learning_rate": 2.6430445355278788e-05,
|
| 2395 |
+
"loss": 0.9752,
|
| 2396 |
+
"step": 341
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 3.257142857142857,
|
| 2400 |
+
"grad_norm": 0.2161754437788185,
|
| 2401 |
+
"learning_rate": 2.6180298875031098e-05,
|
| 2402 |
+
"loss": 0.9688,
|
| 2403 |
+
"step": 342
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 3.2666666666666666,
|
| 2407 |
+
"grad_norm": 0.2389481202074638,
|
| 2408 |
+
"learning_rate": 2.59307646222302e-05,
|
| 2409 |
+
"loss": 0.9903,
|
| 2410 |
+
"step": 343
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 3.276190476190476,
|
| 2414 |
+
"grad_norm": 0.1688931879244891,
|
| 2415 |
+
"learning_rate": 2.5681853651508704e-05,
|
| 2416 |
+
"loss": 0.9906,
|
| 2417 |
+
"step": 344
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 3.2857142857142856,
|
| 2421 |
+
"grad_norm": 0.2709186008027029,
|
| 2422 |
+
"learning_rate": 2.5433576989887115e-05,
|
| 2423 |
+
"loss": 0.9634,
|
| 2424 |
+
"step": 345
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 3.295238095238095,
|
| 2428 |
+
"grad_norm": 0.17314495292558174,
|
| 2429 |
+
"learning_rate": 2.5185945636285416e-05,
|
| 2430 |
+
"loss": 0.9805,
|
| 2431 |
+
"step": 346
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 3.3047619047619046,
|
| 2435 |
+
"grad_norm": 0.226865700017327,
|
| 2436 |
+
"learning_rate": 2.4938970561035753e-05,
|
| 2437 |
+
"loss": 0.9745,
|
| 2438 |
+
"step": 347
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 3.314285714285714,
|
| 2442 |
+
"grad_norm": 0.22027328808623828,
|
| 2443 |
+
"learning_rate": 2.4692662705396412e-05,
|
| 2444 |
+
"loss": 0.9808,
|
| 2445 |
+
"step": 348
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 3.323809523809524,
|
| 2449 |
+
"grad_norm": 0.21347101273276106,
|
| 2450 |
+
"learning_rate": 2.444703298106718e-05,
|
| 2451 |
+
"loss": 0.9742,
|
| 2452 |
+
"step": 349
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 3.3333333333333335,
|
| 2456 |
+
"grad_norm": 0.6099297681729771,
|
| 2457 |
+
"learning_rate": 2.420209226970588e-05,
|
| 2458 |
+
"loss": 1.0004,
|
| 2459 |
+
"step": 350
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 3.342857142857143,
|
| 2463 |
+
"grad_norm": 0.18262102730707186,
|
| 2464 |
+
"learning_rate": 2.395785142244634e-05,
|
| 2465 |
+
"loss": 0.9785,
|
| 2466 |
+
"step": 351
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 3.3523809523809525,
|
| 2470 |
+
"grad_norm": 0.19086404836286397,
|
| 2471 |
+
"learning_rate": 2.3714321259417662e-05,
|
| 2472 |
+
"loss": 0.9784,
|
| 2473 |
+
"step": 352
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 3.361904761904762,
|
| 2477 |
+
"grad_norm": 0.2101963610417534,
|
| 2478 |
+
"learning_rate": 2.3471512569264884e-05,
|
| 2479 |
+
"loss": 0.9703,
|
| 2480 |
+
"step": 353
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 3.3714285714285714,
|
| 2484 |
+
"grad_norm": 0.20089915550579932,
|
| 2485 |
+
"learning_rate": 2.3229436108671014e-05,
|
| 2486 |
+
"loss": 0.9753,
|
| 2487 |
+
"step": 354
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 3.380952380952381,
|
| 2491 |
+
"grad_norm": 0.17710318859904475,
|
| 2492 |
+
"learning_rate": 2.298810260188054e-05,
|
| 2493 |
+
"loss": 0.9717,
|
| 2494 |
+
"step": 355
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 3.3904761904761904,
|
| 2498 |
+
"grad_norm": 0.1505659891122774,
|
| 2499 |
+
"learning_rate": 2.274752274022428e-05,
|
| 2500 |
+
"loss": 0.9778,
|
| 2501 |
+
"step": 356
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 3.4,
|
| 2505 |
+
"grad_norm": 0.18038983533749448,
|
| 2506 |
+
"learning_rate": 2.250770718164579e-05,
|
| 2507 |
+
"loss": 0.9854,
|
| 2508 |
+
"step": 357
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 3.4095238095238094,
|
| 2512 |
+
"grad_norm": 0.18174732799541482,
|
| 2513 |
+
"learning_rate": 2.2268666550229173e-05,
|
| 2514 |
+
"loss": 0.9787,
|
| 2515 |
+
"step": 358
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 3.419047619047619,
|
| 2519 |
+
"grad_norm": 0.15466454580054184,
|
| 2520 |
+
"learning_rate": 2.203041143572845e-05,
|
| 2521 |
+
"loss": 0.9771,
|
| 2522 |
+
"step": 359
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 3.4285714285714284,
|
| 2526 |
+
"grad_norm": 0.17410077565262463,
|
| 2527 |
+
"learning_rate": 2.1792952393098394e-05,
|
| 2528 |
+
"loss": 0.9876,
|
| 2529 |
+
"step": 360
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 3.4380952380952383,
|
| 2533 |
+
"grad_norm": 0.17696206014651153,
|
| 2534 |
+
"learning_rate": 2.155629994202696e-05,
|
| 2535 |
+
"loss": 0.9844,
|
| 2536 |
+
"step": 361
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 3.447619047619048,
|
| 2540 |
+
"grad_norm": 0.15599328687176803,
|
| 2541 |
+
"learning_rate": 2.1320464566469233e-05,
|
| 2542 |
+
"loss": 0.9827,
|
| 2543 |
+
"step": 362
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 3.4571428571428573,
|
| 2547 |
+
"grad_norm": 0.18924852148069596,
|
| 2548 |
+
"learning_rate": 2.1085456714183002e-05,
|
| 2549 |
+
"loss": 0.9737,
|
| 2550 |
+
"step": 363
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 3.466666666666667,
|
| 2554 |
+
"grad_norm": 0.21383476826326972,
|
| 2555 |
+
"learning_rate": 2.0851286796265838e-05,
|
| 2556 |
+
"loss": 0.9817,
|
| 2557 |
+
"step": 364
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 3.4761904761904763,
|
| 2561 |
+
"grad_norm": 0.17321070429571578,
|
| 2562 |
+
"learning_rate": 2.0617965186694e-05,
|
| 2563 |
+
"loss": 0.984,
|
| 2564 |
+
"step": 365
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 3.4857142857142858,
|
| 2568 |
+
"grad_norm": 0.17184638099703334,
|
| 2569 |
+
"learning_rate": 2.0385502221862742e-05,
|
| 2570 |
+
"loss": 0.9878,
|
| 2571 |
+
"step": 366
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 3.4952380952380953,
|
| 2575 |
+
"grad_norm": 0.17908426969585328,
|
| 2576 |
+
"learning_rate": 2.015390820012847e-05,
|
| 2577 |
+
"loss": 0.9842,
|
| 2578 |
+
"step": 367
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 3.5047619047619047,
|
| 2582 |
+
"grad_norm": 0.1640535311292707,
|
| 2583 |
+
"learning_rate": 1.9923193381352468e-05,
|
| 2584 |
+
"loss": 0.9641,
|
| 2585 |
+
"step": 368
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 3.5142857142857142,
|
| 2589 |
+
"grad_norm": 0.16868052713135487,
|
| 2590 |
+
"learning_rate": 1.9693367986446415e-05,
|
| 2591 |
+
"loss": 0.9616,
|
| 2592 |
+
"step": 369
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 3.5238095238095237,
|
| 2596 |
+
"grad_norm": 0.14316347468811785,
|
| 2597 |
+
"learning_rate": 1.9464442196919525e-05,
|
| 2598 |
+
"loss": 0.9854,
|
| 2599 |
+
"step": 370
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 3.533333333333333,
|
| 2603 |
+
"grad_norm": 0.13603829907464166,
|
| 2604 |
+
"learning_rate": 1.9236426154427583e-05,
|
| 2605 |
+
"loss": 0.9741,
|
| 2606 |
+
"step": 371
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 3.5428571428571427,
|
| 2610 |
+
"grad_norm": 0.13833016109922058,
|
| 2611 |
+
"learning_rate": 1.9009329960323594e-05,
|
| 2612 |
+
"loss": 0.9911,
|
| 2613 |
+
"step": 372
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 3.552380952380952,
|
| 2617 |
+
"grad_norm": 0.14395757212272578,
|
| 2618 |
+
"learning_rate": 1.8783163675210307e-05,
|
| 2619 |
+
"loss": 0.9837,
|
| 2620 |
+
"step": 373
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 3.5619047619047617,
|
| 2624 |
+
"grad_norm": 0.16846410866461603,
|
| 2625 |
+
"learning_rate": 1.8557937318494507e-05,
|
| 2626 |
+
"loss": 0.9744,
|
| 2627 |
+
"step": 374
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 3.571428571428571,
|
| 2631 |
+
"grad_norm": 0.1523793105601579,
|
| 2632 |
+
"learning_rate": 1.8333660867943163e-05,
|
| 2633 |
+
"loss": 0.971,
|
| 2634 |
+
"step": 375
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 3.580952380952381,
|
| 2638 |
+
"grad_norm": 0.13785660562080743,
|
| 2639 |
+
"learning_rate": 1.8110344259241398e-05,
|
| 2640 |
+
"loss": 0.9766,
|
| 2641 |
+
"step": 376
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 3.5904761904761906,
|
| 2645 |
+
"grad_norm": 0.1325642554414296,
|
| 2646 |
+
"learning_rate": 1.7887997385552278e-05,
|
| 2647 |
+
"loss": 0.9649,
|
| 2648 |
+
"step": 377
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 3.6,
|
| 2652 |
+
"grad_norm": 0.15775432838130118,
|
| 2653 |
+
"learning_rate": 1.766663009707861e-05,
|
| 2654 |
+
"loss": 0.9689,
|
| 2655 |
+
"step": 378
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 3.6095238095238096,
|
| 2659 |
+
"grad_norm": 0.12446439908125384,
|
| 2660 |
+
"learning_rate": 1.7446252200626555e-05,
|
| 2661 |
+
"loss": 0.9722,
|
| 2662 |
+
"step": 379
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 3.619047619047619,
|
| 2666 |
+
"grad_norm": 0.13869302526796073,
|
| 2667 |
+
"learning_rate": 1.7226873459171142e-05,
|
| 2668 |
+
"loss": 0.9651,
|
| 2669 |
+
"step": 380
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 3.6285714285714286,
|
| 2673 |
+
"grad_norm": 0.12813837269442743,
|
| 2674 |
+
"learning_rate": 1.700850359142373e-05,
|
| 2675 |
+
"loss": 0.9659,
|
| 2676 |
+
"step": 381
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 3.638095238095238,
|
| 2680 |
+
"grad_norm": 0.13933232700434076,
|
| 2681 |
+
"learning_rate": 1.679115227140155e-05,
|
| 2682 |
+
"loss": 0.9731,
|
| 2683 |
+
"step": 382
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 3.6476190476190475,
|
| 2687 |
+
"grad_norm": 0.14485112162942804,
|
| 2688 |
+
"learning_rate": 1.6574829127999067e-05,
|
| 2689 |
+
"loss": 0.9854,
|
| 2690 |
+
"step": 383
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 3.657142857142857,
|
| 2694 |
+
"grad_norm": 0.1390597866481306,
|
| 2695 |
+
"learning_rate": 1.6359543744561438e-05,
|
| 2696 |
+
"loss": 0.9706,
|
| 2697 |
+
"step": 384
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 3.6666666666666665,
|
| 2701 |
+
"grad_norm": 0.14156616905744906,
|
| 2702 |
+
"learning_rate": 1.614530565845994e-05,
|
| 2703 |
+
"loss": 0.9741,
|
| 2704 |
+
"step": 385
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 3.6761904761904765,
|
| 2708 |
+
"grad_norm": 0.14789503098270715,
|
| 2709 |
+
"learning_rate": 1.5932124360669473e-05,
|
| 2710 |
+
"loss": 0.9792,
|
| 2711 |
+
"step": 386
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 3.685714285714286,
|
| 2715 |
+
"grad_norm": 0.14078688051407673,
|
| 2716 |
+
"learning_rate": 1.5720009295348103e-05,
|
| 2717 |
+
"loss": 0.9805,
|
| 2718 |
+
"step": 387
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 3.6952380952380954,
|
| 2722 |
+
"grad_norm": 0.149662490393232,
|
| 2723 |
+
"learning_rate": 1.5508969859418617e-05,
|
| 2724 |
+
"loss": 0.9708,
|
| 2725 |
+
"step": 388
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 3.704761904761905,
|
| 2729 |
+
"grad_norm": 0.15565972406127485,
|
| 2730 |
+
"learning_rate": 1.529901540215233e-05,
|
| 2731 |
+
"loss": 0.9907,
|
| 2732 |
+
"step": 389
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 3.7142857142857144,
|
| 2736 |
+
"grad_norm": 0.14457925539378302,
|
| 2737 |
+
"learning_rate": 1.5090155224754823e-05,
|
| 2738 |
+
"loss": 0.972,
|
| 2739 |
+
"step": 390
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 3.723809523809524,
|
| 2743 |
+
"grad_norm": 0.15980853988420965,
|
| 2744 |
+
"learning_rate": 1.4882398579953928e-05,
|
| 2745 |
+
"loss": 0.9858,
|
| 2746 |
+
"step": 391
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 3.7333333333333334,
|
| 2750 |
+
"grad_norm": 0.1353513800841828,
|
| 2751 |
+
"learning_rate": 1.4675754671589801e-05,
|
| 2752 |
+
"loss": 0.9843,
|
| 2753 |
+
"step": 392
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 3.742857142857143,
|
| 2757 |
+
"grad_norm": 0.13602936334115348,
|
| 2758 |
+
"learning_rate": 1.4470232654207208e-05,
|
| 2759 |
+
"loss": 0.9819,
|
| 2760 |
+
"step": 393
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 3.7523809523809524,
|
| 2764 |
+
"grad_norm": 0.14448055061579274,
|
| 2765 |
+
"learning_rate": 1.4265841632649915e-05,
|
| 2766 |
+
"loss": 0.9722,
|
| 2767 |
+
"step": 394
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 3.761904761904762,
|
| 2771 |
+
"grad_norm": 0.13590836391844754,
|
| 2772 |
+
"learning_rate": 1.40625906616574e-05,
|
| 2773 |
+
"loss": 0.9658,
|
| 2774 |
+
"step": 395
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 3.7714285714285714,
|
| 2778 |
+
"grad_norm": 0.1435455827754151,
|
| 2779 |
+
"learning_rate": 1.3860488745463694e-05,
|
| 2780 |
+
"loss": 0.9794,
|
| 2781 |
+
"step": 396
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 3.780952380952381,
|
| 2785 |
+
"grad_norm": 0.1397532798524967,
|
| 2786 |
+
"learning_rate": 1.365954483739846e-05,
|
| 2787 |
+
"loss": 0.9829,
|
| 2788 |
+
"step": 397
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 3.7904761904761903,
|
| 2792 |
+
"grad_norm": 0.12978394192415188,
|
| 2793 |
+
"learning_rate": 1.3459767839490386e-05,
|
| 2794 |
+
"loss": 0.9721,
|
| 2795 |
+
"step": 398
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 3.8,
|
| 2799 |
+
"grad_norm": 0.14356015101687936,
|
| 2800 |
+
"learning_rate": 1.326116660207279e-05,
|
| 2801 |
+
"loss": 0.9835,
|
| 2802 |
+
"step": 399
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 3.8095238095238093,
|
| 2806 |
+
"grad_norm": 0.13247764252770453,
|
| 2807 |
+
"learning_rate": 1.3063749923391557e-05,
|
| 2808 |
+
"loss": 0.9765,
|
| 2809 |
+
"step": 400
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 3.819047619047619,
|
| 2813 |
+
"grad_norm": 0.13968884823844,
|
| 2814 |
+
"learning_rate": 1.2867526549215356e-05,
|
| 2815 |
+
"loss": 0.9806,
|
| 2816 |
+
"step": 401
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 3.8285714285714287,
|
| 2820 |
+
"grad_norm": 0.15020881590521024,
|
| 2821 |
+
"learning_rate": 1.2672505172448201e-05,
|
| 2822 |
+
"loss": 0.9743,
|
| 2823 |
+
"step": 402
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 3.8380952380952382,
|
| 2827 |
+
"grad_norm": 0.12600562838248064,
|
| 2828 |
+
"learning_rate": 1.2478694432744342e-05,
|
| 2829 |
+
"loss": 0.9747,
|
| 2830 |
+
"step": 403
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 3.8476190476190477,
|
| 2834 |
+
"grad_norm": 0.1253894515680319,
|
| 2835 |
+
"learning_rate": 1.2286102916125534e-05,
|
| 2836 |
+
"loss": 0.9734,
|
| 2837 |
+
"step": 404
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 3.857142857142857,
|
| 2841 |
+
"grad_norm": 0.14320490323799898,
|
| 2842 |
+
"learning_rate": 1.2094739154600616e-05,
|
| 2843 |
+
"loss": 0.9816,
|
| 2844 |
+
"step": 405
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 3.8666666666666667,
|
| 2848 |
+
"grad_norm": 0.12399745336033988,
|
| 2849 |
+
"learning_rate": 1.1904611625787612e-05,
|
| 2850 |
+
"loss": 0.9884,
|
| 2851 |
+
"step": 406
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 3.876190476190476,
|
| 2855 |
+
"grad_norm": 0.13417523628822833,
|
| 2856 |
+
"learning_rate": 1.1715728752538103e-05,
|
| 2857 |
+
"loss": 0.9783,
|
| 2858 |
+
"step": 407
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 3.8857142857142857,
|
| 2862 |
+
"grad_norm": 0.15683701346116036,
|
| 2863 |
+
"learning_rate": 1.1528098902564109e-05,
|
| 2864 |
+
"loss": 0.9733,
|
| 2865 |
+
"step": 408
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 3.895238095238095,
|
| 2869 |
+
"grad_norm": 0.13713534522259704,
|
| 2870 |
+
"learning_rate": 1.1341730388067393e-05,
|
| 2871 |
+
"loss": 0.9818,
|
| 2872 |
+
"step": 409
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 3.9047619047619047,
|
| 2876 |
+
"grad_norm": 0.13758248343689114,
|
| 2877 |
+
"learning_rate": 1.1156631465371213e-05,
|
| 2878 |
+
"loss": 0.9688,
|
| 2879 |
+
"step": 410
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 3.914285714285714,
|
| 2883 |
+
"grad_norm": 0.14115308612974595,
|
| 2884 |
+
"learning_rate": 1.0972810334554565e-05,
|
| 2885 |
+
"loss": 0.9797,
|
| 2886 |
+
"step": 411
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 3.923809523809524,
|
| 2890 |
+
"grad_norm": 0.5870477835443715,
|
| 2891 |
+
"learning_rate": 1.0790275139088879e-05,
|
| 2892 |
+
"loss": 0.982,
|
| 2893 |
+
"step": 412
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 3.9333333333333336,
|
| 2897 |
+
"grad_norm": 0.12922641407686744,
|
| 2898 |
+
"learning_rate": 1.0609033965477318e-05,
|
| 2899 |
+
"loss": 0.9805,
|
| 2900 |
+
"step": 413
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 3.942857142857143,
|
| 2904 |
+
"grad_norm": 0.14422492668411402,
|
| 2905 |
+
"learning_rate": 1.0429094842896484e-05,
|
| 2906 |
+
"loss": 0.9795,
|
| 2907 |
+
"step": 414
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 3.9523809523809526,
|
| 2911 |
+
"grad_norm": 0.14945738902129274,
|
| 2912 |
+
"learning_rate": 1.0250465742840743e-05,
|
| 2913 |
+
"loss": 0.9736,
|
| 2914 |
+
"step": 415
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 3.961904761904762,
|
| 2918 |
+
"grad_norm": 0.12244500468761615,
|
| 2919 |
+
"learning_rate": 1.007315457876907e-05,
|
| 2920 |
+
"loss": 0.9811,
|
| 2921 |
+
"step": 416
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 3.9714285714285715,
|
| 2925 |
+
"grad_norm": 0.13445713568795264,
|
| 2926 |
+
"learning_rate": 9.897169205754461e-06,
|
| 2927 |
+
"loss": 0.9846,
|
| 2928 |
+
"step": 417
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 3.980952380952381,
|
| 2932 |
+
"grad_norm": 0.13743382078481892,
|
| 2933 |
+
"learning_rate": 9.722517420135977e-06,
|
| 2934 |
+
"loss": 0.9688,
|
| 2935 |
+
"step": 418
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 3.9904761904761905,
|
| 2939 |
+
"grad_norm": 0.1233003022413971,
|
| 2940 |
+
"learning_rate": 9.549206959173331e-06,
|
| 2941 |
+
"loss": 0.9845,
|
| 2942 |
+
"step": 419
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 4.0,
|
| 2946 |
+
"grad_norm": 0.1263829549351974,
|
| 2947 |
+
"learning_rate": 9.377245500704135e-06,
|
| 2948 |
+
"loss": 0.9928,
|
| 2949 |
+
"step": 420
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 4.0095238095238095,
|
| 2953 |
+
"grad_norm": 0.17390084497636504,
|
| 2954 |
+
"learning_rate": 9.206640662803746e-06,
|
| 2955 |
+
"loss": 0.9623,
|
| 2956 |
+
"step": 421
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 4.019047619047619,
|
| 2960 |
+
"grad_norm": 0.1431373191141913,
|
| 2961 |
+
"learning_rate": 9.037400003447808e-06,
|
| 2962 |
+
"loss": 0.9616,
|
| 2963 |
+
"step": 422
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 4.0285714285714285,
|
| 2967 |
+
"grad_norm": 0.13455551888638445,
|
| 2968 |
+
"learning_rate": 8.869531020177367e-06,
|
| 2969 |
+
"loss": 0.9475,
|
| 2970 |
+
"step": 423
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 4.038095238095238,
|
| 2974 |
+
"grad_norm": 0.14362146854953411,
|
| 2975 |
+
"learning_rate": 8.703041149766797e-06,
|
| 2976 |
+
"loss": 0.9541,
|
| 2977 |
+
"step": 424
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 4.0476190476190474,
|
| 2981 |
+
"grad_norm": 0.1452952254226326,
|
| 2982 |
+
"learning_rate": 8.537937767894303e-06,
|
| 2983 |
+
"loss": 0.9548,
|
| 2984 |
+
"step": 425
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 4.057142857142857,
|
| 2988 |
+
"grad_norm": 0.156138958795091,
|
| 2989 |
+
"learning_rate": 8.374228188815157e-06,
|
| 2990 |
+
"loss": 0.9597,
|
| 2991 |
+
"step": 426
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 4.066666666666666,
|
| 2995 |
+
"grad_norm": 0.15807376487720612,
|
| 2996 |
+
"learning_rate": 8.211919665037697e-06,
|
| 2997 |
+
"loss": 0.944,
|
| 2998 |
+
"step": 427
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 4.076190476190476,
|
| 3002 |
+
"grad_norm": 0.1428349305587997,
|
| 3003 |
+
"learning_rate": 8.051019387002035e-06,
|
| 3004 |
+
"loss": 0.957,
|
| 3005 |
+
"step": 428
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 4.085714285714285,
|
| 3009 |
+
"grad_norm": 0.156467940239568,
|
| 3010 |
+
"learning_rate": 7.891534482761463e-06,
|
| 3011 |
+
"loss": 0.9519,
|
| 3012 |
+
"step": 429
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 4.095238095238095,
|
| 3016 |
+
"grad_norm": 0.15091456328015054,
|
| 3017 |
+
"learning_rate": 7.733472017666739e-06,
|
| 3018 |
+
"loss": 0.9579,
|
| 3019 |
+
"step": 430
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 4.104761904761904,
|
| 3023 |
+
"grad_norm": 0.15455184255548396,
|
| 3024 |
+
"learning_rate": 7.57683899405305e-06,
|
| 3025 |
+
"loss": 0.9614,
|
| 3026 |
+
"step": 431
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 4.114285714285714,
|
| 3030 |
+
"grad_norm": 0.23250611326493095,
|
| 3031 |
+
"learning_rate": 7.42164235092981e-06,
|
| 3032 |
+
"loss": 0.9777,
|
| 3033 |
+
"step": 432
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 4.123809523809523,
|
| 3037 |
+
"grad_norm": 0.15146330742839126,
|
| 3038 |
+
"learning_rate": 7.26788896367324e-06,
|
| 3039 |
+
"loss": 0.9519,
|
| 3040 |
+
"step": 433
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 4.133333333333334,
|
| 3044 |
+
"grad_norm": 0.14062980079515083,
|
| 3045 |
+
"learning_rate": 7.115585643721798e-06,
|
| 3046 |
+
"loss": 0.961,
|
| 3047 |
+
"step": 434
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 4.142857142857143,
|
| 3051 |
+
"grad_norm": 0.13654994504554915,
|
| 3052 |
+
"learning_rate": 6.964739138274433e-06,
|
| 3053 |
+
"loss": 0.9638,
|
| 3054 |
+
"step": 435
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 4.152380952380953,
|
| 3058 |
+
"grad_norm": 0.13405718763219662,
|
| 3059 |
+
"learning_rate": 6.815356129991624e-06,
|
| 3060 |
+
"loss": 0.9523,
|
| 3061 |
+
"step": 436
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 4.161904761904762,
|
| 3065 |
+
"grad_norm": 0.1355347721482518,
|
| 3066 |
+
"learning_rate": 6.667443236699398e-06,
|
| 3067 |
+
"loss": 0.9572,
|
| 3068 |
+
"step": 437
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 4.171428571428572,
|
| 3072 |
+
"grad_norm": 0.3193405192419738,
|
| 3073 |
+
"learning_rate": 6.521007011096143e-06,
|
| 3074 |
+
"loss": 0.965,
|
| 3075 |
+
"step": 438
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 4.180952380952381,
|
| 3079 |
+
"grad_norm": 0.12077417580448366,
|
| 3080 |
+
"learning_rate": 6.376053940462279e-06,
|
| 3081 |
+
"loss": 0.959,
|
| 3082 |
+
"step": 439
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 4.190476190476191,
|
| 3086 |
+
"grad_norm": 0.11975543200056642,
|
| 3087 |
+
"learning_rate": 6.232590446372864e-06,
|
| 3088 |
+
"loss": 0.9463,
|
| 3089 |
+
"step": 440
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 4.2,
|
| 3093 |
+
"grad_norm": 0.12394898025158416,
|
| 3094 |
+
"learning_rate": 6.090622884413164e-06,
|
| 3095 |
+
"loss": 0.9599,
|
| 3096 |
+
"step": 441
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 4.20952380952381,
|
| 3100 |
+
"grad_norm": 0.13246928966220897,
|
| 3101 |
+
"learning_rate": 5.95015754389705e-06,
|
| 3102 |
+
"loss": 0.9587,
|
| 3103 |
+
"step": 442
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 4.219047619047619,
|
| 3107 |
+
"grad_norm": 0.17218887548328696,
|
| 3108 |
+
"learning_rate": 5.811200647588386e-06,
|
| 3109 |
+
"loss": 0.9483,
|
| 3110 |
+
"step": 443
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 4.228571428571429,
|
| 3114 |
+
"grad_norm": 0.12269437836537307,
|
| 3115 |
+
"learning_rate": 5.673758351425358e-06,
|
| 3116 |
+
"loss": 0.9652,
|
| 3117 |
+
"step": 444
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 4.238095238095238,
|
| 3121 |
+
"grad_norm": 0.1150641697364405,
|
| 3122 |
+
"learning_rate": 5.537836744247753e-06,
|
| 3123 |
+
"loss": 0.9481,
|
| 3124 |
+
"step": 445
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 4.247619047619048,
|
| 3128 |
+
"grad_norm": 0.12282362749817176,
|
| 3129 |
+
"learning_rate": 5.403441847527227e-06,
|
| 3130 |
+
"loss": 0.9565,
|
| 3131 |
+
"step": 446
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 4.257142857142857,
|
| 3135 |
+
"grad_norm": 0.12605101659916942,
|
| 3136 |
+
"learning_rate": 5.270579615100518e-06,
|
| 3137 |
+
"loss": 0.9476,
|
| 3138 |
+
"step": 447
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 4.266666666666667,
|
| 3142 |
+
"grad_norm": 0.12368987938067505,
|
| 3143 |
+
"learning_rate": 5.139255932905731e-06,
|
| 3144 |
+
"loss": 0.9538,
|
| 3145 |
+
"step": 448
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 4.276190476190476,
|
| 3149 |
+
"grad_norm": 0.11271067268917899,
|
| 3150 |
+
"learning_rate": 5.009476618721549e-06,
|
| 3151 |
+
"loss": 0.9776,
|
| 3152 |
+
"step": 449
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 4.285714285714286,
|
| 3156 |
+
"grad_norm": 0.1136633818810862,
|
| 3157 |
+
"learning_rate": 4.881247421909523e-06,
|
| 3158 |
+
"loss": 0.962,
|
| 3159 |
+
"step": 450
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 4.295238095238095,
|
| 3163 |
+
"grad_norm": 0.127022521830231,
|
| 3164 |
+
"learning_rate": 4.754574023159335e-06,
|
| 3165 |
+
"loss": 0.9574,
|
| 3166 |
+
"step": 451
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 4.304761904761905,
|
| 3170 |
+
"grad_norm": 0.11992208253243061,
|
| 3171 |
+
"learning_rate": 4.629462034237193e-06,
|
| 3172 |
+
"loss": 0.9606,
|
| 3173 |
+
"step": 452
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 4.314285714285714,
|
| 3177 |
+
"grad_norm": 0.11191987434816011,
|
| 3178 |
+
"learning_rate": 4.505916997737143e-06,
|
| 3179 |
+
"loss": 0.9489,
|
| 3180 |
+
"step": 453
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 4.3238095238095235,
|
| 3184 |
+
"grad_norm": 0.1183918815452489,
|
| 3185 |
+
"learning_rate": 4.383944386835617e-06,
|
| 3186 |
+
"loss": 0.9516,
|
| 3187 |
+
"step": 454
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 4.333333333333333,
|
| 3191 |
+
"grad_norm": 0.11408502654065866,
|
| 3192 |
+
"learning_rate": 4.263549605048898e-06,
|
| 3193 |
+
"loss": 0.9407,
|
| 3194 |
+
"step": 455
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 4.3428571428571425,
|
| 3198 |
+
"grad_norm": 0.11790658741334677,
|
| 3199 |
+
"learning_rate": 4.144737985993774e-06,
|
| 3200 |
+
"loss": 0.9539,
|
| 3201 |
+
"step": 456
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 4.352380952380952,
|
| 3205 |
+
"grad_norm": 0.12878580244190324,
|
| 3206 |
+
"learning_rate": 4.027514793151235e-06,
|
| 3207 |
+
"loss": 0.961,
|
| 3208 |
+
"step": 457
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 4.3619047619047615,
|
| 3212 |
+
"grad_norm": 0.12558845969705756,
|
| 3213 |
+
"learning_rate": 3.9118852196333e-06,
|
| 3214 |
+
"loss": 0.9681,
|
| 3215 |
+
"step": 458
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 4.371428571428572,
|
| 3219 |
+
"grad_norm": 0.11339758568248902,
|
| 3220 |
+
"learning_rate": 3.7978543879529704e-06,
|
| 3221 |
+
"loss": 0.9544,
|
| 3222 |
+
"step": 459
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 4.380952380952381,
|
| 3226 |
+
"grad_norm": 0.12502688051375377,
|
| 3227 |
+
"learning_rate": 3.6854273497972705e-06,
|
| 3228 |
+
"loss": 0.9573,
|
| 3229 |
+
"step": 460
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 4.390476190476191,
|
| 3233 |
+
"grad_norm": 0.10534529012828957,
|
| 3234 |
+
"learning_rate": 3.574609085803471e-06,
|
| 3235 |
+
"loss": 0.9369,
|
| 3236 |
+
"step": 461
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 4.4,
|
| 3240 |
+
"grad_norm": 0.10245960822897354,
|
| 3241 |
+
"learning_rate": 3.4654045053384456e-06,
|
| 3242 |
+
"loss": 0.9501,
|
| 3243 |
+
"step": 462
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 4.40952380952381,
|
| 3247 |
+
"grad_norm": 0.11767956608151667,
|
| 3248 |
+
"learning_rate": 3.3578184462811714e-06,
|
| 3249 |
+
"loss": 0.9564,
|
| 3250 |
+
"step": 463
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 4.419047619047619,
|
| 3254 |
+
"grad_norm": 0.12275952669713826,
|
| 3255 |
+
"learning_rate": 3.2518556748083817e-06,
|
| 3256 |
+
"loss": 0.9536,
|
| 3257 |
+
"step": 464
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 4.428571428571429,
|
| 3261 |
+
"grad_norm": 0.10820465034590014,
|
| 3262 |
+
"learning_rate": 3.1475208851834815e-06,
|
| 3263 |
+
"loss": 0.9736,
|
| 3264 |
+
"step": 465
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 4.438095238095238,
|
| 3268 |
+
"grad_norm": 0.10999235496173602,
|
| 3269 |
+
"learning_rate": 3.0448186995485307e-06,
|
| 3270 |
+
"loss": 0.9661,
|
| 3271 |
+
"step": 466
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 4.447619047619048,
|
| 3275 |
+
"grad_norm": 0.10346043147687037,
|
| 3276 |
+
"learning_rate": 2.9437536677194976e-06,
|
| 3277 |
+
"loss": 0.9567,
|
| 3278 |
+
"step": 467
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 4.457142857142857,
|
| 3282 |
+
"grad_norm": 0.10719225041373819,
|
| 3283 |
+
"learning_rate": 2.844330266984705e-06,
|
| 3284 |
+
"loss": 0.9456,
|
| 3285 |
+
"step": 468
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 4.466666666666667,
|
| 3289 |
+
"grad_norm": 0.10494961443465013,
|
| 3290 |
+
"learning_rate": 2.746552901906463e-06,
|
| 3291 |
+
"loss": 0.9639,
|
| 3292 |
+
"step": 469
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 4.476190476190476,
|
| 3296 |
+
"grad_norm": 0.10157859382678862,
|
| 3297 |
+
"learning_rate": 2.650425904125964e-06,
|
| 3298 |
+
"loss": 0.9562,
|
| 3299 |
+
"step": 470
|
| 3300 |
+
},
|
| 3301 |
+
{
|
| 3302 |
+
"epoch": 4.485714285714286,
|
| 3303 |
+
"grad_norm": 0.09988194579990615,
|
| 3304 |
+
"learning_rate": 2.55595353217136e-06,
|
| 3305 |
+
"loss": 0.9488,
|
| 3306 |
+
"step": 471
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 4.495238095238095,
|
| 3310 |
+
"grad_norm": 0.09993285343775345,
|
| 3311 |
+
"learning_rate": 2.463139971269133e-06,
|
| 3312 |
+
"loss": 0.9599,
|
| 3313 |
+
"step": 472
|
| 3314 |
+
},
|
| 3315 |
+
{
|
| 3316 |
+
"epoch": 4.504761904761905,
|
| 3317 |
+
"grad_norm": 0.09844176642750738,
|
| 3318 |
+
"learning_rate": 2.371989333158671e-06,
|
| 3319 |
+
"loss": 0.9615,
|
| 3320 |
+
"step": 473
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 4.514285714285714,
|
| 3324 |
+
"grad_norm": 0.10190707779708189,
|
| 3325 |
+
"learning_rate": 2.2825056559101145e-06,
|
| 3326 |
+
"loss": 0.9563,
|
| 3327 |
+
"step": 474
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 4.523809523809524,
|
| 3331 |
+
"grad_norm": 0.10131001370817404,
|
| 3332 |
+
"learning_rate": 2.194692903745459e-06,
|
| 3333 |
+
"loss": 0.9587,
|
| 3334 |
+
"step": 475
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 4.533333333333333,
|
| 3338 |
+
"grad_norm": 0.10251872435272494,
|
| 3339 |
+
"learning_rate": 2.1085549668629567e-06,
|
| 3340 |
+
"loss": 0.9489,
|
| 3341 |
+
"step": 476
|
| 3342 |
+
},
|
| 3343 |
+
{
|
| 3344 |
+
"epoch": 4.542857142857143,
|
| 3345 |
+
"grad_norm": 0.10043227346155645,
|
| 3346 |
+
"learning_rate": 2.0240956612647487e-06,
|
| 3347 |
+
"loss": 0.9634,
|
| 3348 |
+
"step": 477
|
| 3349 |
+
},
|
| 3350 |
+
{
|
| 3351 |
+
"epoch": 4.552380952380952,
|
| 3352 |
+
"grad_norm": 0.09956161225983029,
|
| 3353 |
+
"learning_rate": 1.9413187285878355e-06,
|
| 3354 |
+
"loss": 0.9468,
|
| 3355 |
+
"step": 478
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 4.561904761904762,
|
| 3359 |
+
"grad_norm": 0.11045616365540861,
|
| 3360 |
+
"learning_rate": 1.8602278359383063e-06,
|
| 3361 |
+
"loss": 0.9526,
|
| 3362 |
+
"step": 479
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 4.571428571428571,
|
| 3366 |
+
"grad_norm": 0.09834202928593194,
|
| 3367 |
+
"learning_rate": 1.78082657572888e-06,
|
| 3368 |
+
"loss": 0.9609,
|
| 3369 |
+
"step": 480
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 4.580952380952381,
|
| 3373 |
+
"grad_norm": 0.1147748598722428,
|
| 3374 |
+
"learning_rate": 1.7031184655197818e-06,
|
| 3375 |
+
"loss": 0.9412,
|
| 3376 |
+
"step": 481
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 4.59047619047619,
|
| 3380 |
+
"grad_norm": 0.10211354358709429,
|
| 3381 |
+
"learning_rate": 1.6271069478628644e-06,
|
| 3382 |
+
"loss": 0.9557,
|
| 3383 |
+
"step": 482
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 4.6,
|
| 3387 |
+
"grad_norm": 0.1146778979125764,
|
| 3388 |
+
"learning_rate": 1.5527953901491466e-06,
|
| 3389 |
+
"loss": 0.9468,
|
| 3390 |
+
"step": 483
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 4.609523809523809,
|
| 3394 |
+
"grad_norm": 0.10809368043798627,
|
| 3395 |
+
"learning_rate": 1.48018708445961e-06,
|
| 3396 |
+
"loss": 0.9693,
|
| 3397 |
+
"step": 484
|
| 3398 |
+
},
|
| 3399 |
+
{
|
| 3400 |
+
"epoch": 4.619047619047619,
|
| 3401 |
+
"grad_norm": 0.09749254775799188,
|
| 3402 |
+
"learning_rate": 1.409285247419363e-06,
|
| 3403 |
+
"loss": 0.9524,
|
| 3404 |
+
"step": 485
|
| 3405 |
+
},
|
| 3406 |
+
{
|
| 3407 |
+
"epoch": 4.628571428571428,
|
| 3408 |
+
"grad_norm": 0.09704895874324644,
|
| 3409 |
+
"learning_rate": 1.3400930200551331e-06,
|
| 3410 |
+
"loss": 0.9607,
|
| 3411 |
+
"step": 486
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"epoch": 4.638095238095238,
|
| 3415 |
+
"grad_norm": 0.09581374811618217,
|
| 3416 |
+
"learning_rate": 1.2726134676561385e-06,
|
| 3417 |
+
"loss": 0.9747,
|
| 3418 |
+
"step": 487
|
| 3419 |
+
},
|
| 3420 |
+
{
|
| 3421 |
+
"epoch": 4.647619047619048,
|
| 3422 |
+
"grad_norm": 0.09600971666081125,
|
| 3423 |
+
"learning_rate": 1.2068495796382495e-06,
|
| 3424 |
+
"loss": 0.9483,
|
| 3425 |
+
"step": 488
|
| 3426 |
+
},
|
| 3427 |
+
{
|
| 3428 |
+
"epoch": 4.6571428571428575,
|
| 3429 |
+
"grad_norm": 0.10065951842357326,
|
| 3430 |
+
"learning_rate": 1.142804269411606e-06,
|
| 3431 |
+
"loss": 0.9652,
|
| 3432 |
+
"step": 489
|
| 3433 |
+
},
|
| 3434 |
+
{
|
| 3435 |
+
"epoch": 4.666666666666667,
|
| 3436 |
+
"grad_norm": 0.1012630059377017,
|
| 3437 |
+
"learning_rate": 1.0804803742515068e-06,
|
| 3438 |
+
"loss": 0.9417,
|
| 3439 |
+
"step": 490
|
| 3440 |
+
},
|
| 3441 |
+
{
|
| 3442 |
+
"epoch": 4.6761904761904765,
|
| 3443 |
+
"grad_norm": 0.10102648935759229,
|
| 3444 |
+
"learning_rate": 1.0198806551727557e-06,
|
| 3445 |
+
"loss": 0.9575,
|
| 3446 |
+
"step": 491
|
| 3447 |
+
},
|
| 3448 |
+
{
|
| 3449 |
+
"epoch": 4.685714285714286,
|
| 3450 |
+
"grad_norm": 0.10109454230992841,
|
| 3451 |
+
"learning_rate": 9.610077968072962e-07,
|
| 3452 |
+
"loss": 0.966,
|
| 3453 |
+
"step": 492
|
| 3454 |
+
},
|
| 3455 |
+
{
|
| 3456 |
+
"epoch": 4.695238095238095,
|
| 3457 |
+
"grad_norm": 0.09805229033614825,
|
| 3458 |
+
"learning_rate": 9.038644072853331e-07,
|
| 3459 |
+
"loss": 0.9671,
|
| 3460 |
+
"step": 493
|
| 3461 |
+
},
|
| 3462 |
+
{
|
| 3463 |
+
"epoch": 4.704761904761905,
|
| 3464 |
+
"grad_norm": 0.09726857715744779,
|
| 3465 |
+
"learning_rate": 8.484530181197504e-07,
|
| 3466 |
+
"loss": 0.9477,
|
| 3467 |
+
"step": 494
|
| 3468 |
+
},
|
| 3469 |
+
{
|
| 3470 |
+
"epoch": 4.714285714285714,
|
| 3471 |
+
"grad_norm": 0.1012887727099665,
|
| 3472 |
+
"learning_rate": 7.947760840939688e-07,
|
| 3473 |
+
"loss": 0.9572,
|
| 3474 |
+
"step": 495
|
| 3475 |
+
},
|
| 3476 |
+
{
|
| 3477 |
+
"epoch": 4.723809523809524,
|
| 3478 |
+
"grad_norm": 0.09689224625015615,
|
| 3479 |
+
"learning_rate": 7.428359831532117e-07,
|
| 3480 |
+
"loss": 0.9385,
|
| 3481 |
+
"step": 496
|
| 3482 |
+
},
|
| 3483 |
+
{
|
| 3484 |
+
"epoch": 4.733333333333333,
|
| 3485 |
+
"grad_norm": 0.09894505285476389,
|
| 3486 |
+
"learning_rate": 6.926350162991613e-07,
|
| 3487 |
+
"loss": 0.9502,
|
| 3488 |
+
"step": 497
|
| 3489 |
+
},
|
| 3490 |
+
{
|
| 3491 |
+
"epoch": 4.742857142857143,
|
| 3492 |
+
"grad_norm": 0.10241361126956,
|
| 3493 |
+
"learning_rate": 6.441754074879925e-07,
|
| 3494 |
+
"loss": 0.9495,
|
| 3495 |
+
"step": 498
|
| 3496 |
+
},
|
| 3497 |
+
{
|
| 3498 |
+
"epoch": 4.752380952380952,
|
| 3499 |
+
"grad_norm": 0.09443159904089143,
|
| 3500 |
+
"learning_rate": 5.974593035318777e-07,
|
| 3501 |
+
"loss": 0.9536,
|
| 3502 |
+
"step": 499
|
| 3503 |
+
},
|
| 3504 |
+
{
|
| 3505 |
+
"epoch": 4.761904761904762,
|
| 3506 |
+
"grad_norm": 0.10189499647323387,
|
| 3507 |
+
"learning_rate": 5.524887740038676e-07,
|
| 3508 |
+
"loss": 0.9472,
|
| 3509 |
+
"step": 500
|
| 3510 |
+
},
|
| 3511 |
+
{
|
| 3512 |
+
"epoch": 4.771428571428571,
|
| 3513 |
+
"grad_norm": 0.10773088780281193,
|
| 3514 |
+
"learning_rate": 5.092658111462179e-07,
|
| 3515 |
+
"loss": 0.9473,
|
| 3516 |
+
"step": 501
|
| 3517 |
+
},
|
| 3518 |
+
{
|
| 3519 |
+
"epoch": 4.780952380952381,
|
| 3520 |
+
"grad_norm": 0.3240978802797332,
|
| 3521 |
+
"learning_rate": 4.6779232978211297e-07,
|
| 3522 |
+
"loss": 0.9784,
|
| 3523 |
+
"step": 502
|
| 3524 |
+
},
|
| 3525 |
+
{
|
| 3526 |
+
"epoch": 4.79047619047619,
|
| 3527 |
+
"grad_norm": 0.09242163553460452,
|
| 3528 |
+
"learning_rate": 4.280701672308585e-07,
|
| 3529 |
+
"loss": 0.9583,
|
| 3530 |
+
"step": 503
|
| 3531 |
+
},
|
| 3532 |
+
{
|
| 3533 |
+
"epoch": 4.8,
|
| 3534 |
+
"grad_norm": 0.09118495736223464,
|
| 3535 |
+
"learning_rate": 3.901010832264662e-07,
|
| 3536 |
+
"loss": 0.9473,
|
| 3537 |
+
"step": 504
|
| 3538 |
+
},
|
| 3539 |
+
{
|
| 3540 |
+
"epoch": 4.809523809523809,
|
| 3541 |
+
"grad_norm": 0.09466091553337481,
|
| 3542 |
+
"learning_rate": 3.5388675983971664e-07,
|
| 3543 |
+
"loss": 0.9523,
|
| 3544 |
+
"step": 505
|
| 3545 |
+
},
|
| 3546 |
+
{
|
| 3547 |
+
"epoch": 4.819047619047619,
|
| 3548 |
+
"grad_norm": 0.10089901559145306,
|
| 3549 |
+
"learning_rate": 3.1942880140360955e-07,
|
| 3550 |
+
"loss": 0.9591,
|
| 3551 |
+
"step": 506
|
| 3552 |
+
},
|
| 3553 |
+
{
|
| 3554 |
+
"epoch": 4.828571428571428,
|
| 3555 |
+
"grad_norm": 0.10487799317926302,
|
| 3556 |
+
"learning_rate": 2.867287344423364e-07,
|
| 3557 |
+
"loss": 0.9764,
|
| 3558 |
+
"step": 507
|
| 3559 |
+
},
|
| 3560 |
+
{
|
| 3561 |
+
"epoch": 4.838095238095238,
|
| 3562 |
+
"grad_norm": 0.0945247266110506,
|
| 3563 |
+
"learning_rate": 2.557880076036101e-07,
|
| 3564 |
+
"loss": 0.9492,
|
| 3565 |
+
"step": 508
|
| 3566 |
+
},
|
| 3567 |
+
{
|
| 3568 |
+
"epoch": 4.847619047619047,
|
| 3569 |
+
"grad_norm": 0.09269936781651757,
|
| 3570 |
+
"learning_rate": 2.2660799159451629e-07,
|
| 3571 |
+
"loss": 0.9579,
|
| 3572 |
+
"step": 509
|
| 3573 |
+
},
|
| 3574 |
+
{
|
| 3575 |
+
"epoch": 4.857142857142857,
|
| 3576 |
+
"grad_norm": 0.09460038465065274,
|
| 3577 |
+
"learning_rate": 1.991899791207663e-07,
|
| 3578 |
+
"loss": 0.9638,
|
| 3579 |
+
"step": 510
|
| 3580 |
+
},
|
| 3581 |
+
{
|
| 3582 |
+
"epoch": 4.866666666666667,
|
| 3583 |
+
"grad_norm": 0.0982949463276317,
|
| 3584 |
+
"learning_rate": 1.7353518482946308e-07,
|
| 3585 |
+
"loss": 0.9535,
|
| 3586 |
+
"step": 511
|
| 3587 |
+
},
|
| 3588 |
+
{
|
| 3589 |
+
"epoch": 4.876190476190477,
|
| 3590 |
+
"grad_norm": 0.09485743697787292,
|
| 3591 |
+
"learning_rate": 1.4964474525525075e-07,
|
| 3592 |
+
"loss": 0.9599,
|
| 3593 |
+
"step": 512
|
| 3594 |
+
},
|
| 3595 |
+
{
|
| 3596 |
+
"epoch": 4.885714285714286,
|
| 3597 |
+
"grad_norm": 0.0963473582791281,
|
| 3598 |
+
"learning_rate": 1.2751971876999504e-07,
|
| 3599 |
+
"loss": 0.962,
|
| 3600 |
+
"step": 513
|
| 3601 |
+
},
|
| 3602 |
+
{
|
| 3603 |
+
"epoch": 4.895238095238096,
|
| 3604 |
+
"grad_norm": 0.09723177864800757,
|
| 3605 |
+
"learning_rate": 1.0716108553588289e-07,
|
| 3606 |
+
"loss": 0.9398,
|
| 3607 |
+
"step": 514
|
| 3608 |
+
},
|
| 3609 |
+
{
|
| 3610 |
+
"epoch": 4.904761904761905,
|
| 3611 |
+
"grad_norm": 0.09412117317242105,
|
| 3612 |
+
"learning_rate": 8.856974746199954e-08,
|
| 3613 |
+
"loss": 0.9578,
|
| 3614 |
+
"step": 515
|
| 3615 |
+
},
|
| 3616 |
+
{
|
| 3617 |
+
"epoch": 4.914285714285715,
|
| 3618 |
+
"grad_norm": 0.08864592036283941,
|
| 3619 |
+
"learning_rate": 7.174652816437811e-08,
|
| 3620 |
+
"loss": 0.9599,
|
| 3621 |
+
"step": 516
|
| 3622 |
+
},
|
| 3623 |
+
{
|
| 3624 |
+
"epoch": 4.923809523809524,
|
| 3625 |
+
"grad_norm": 0.09083413925650892,
|
| 3626 |
+
"learning_rate": 5.669217292952223e-08,
|
| 3627 |
+
"loss": 0.9512,
|
| 3628 |
+
"step": 517
|
| 3629 |
+
},
|
| 3630 |
+
{
|
| 3631 |
+
"epoch": 4.933333333333334,
|
| 3632 |
+
"grad_norm": 0.09669535232036294,
|
| 3633 |
+
"learning_rate": 4.3407348681361314e-08,
|
| 3634 |
+
"loss": 0.9524,
|
| 3635 |
+
"step": 518
|
| 3636 |
+
},
|
| 3637 |
+
{
|
| 3638 |
+
"epoch": 4.942857142857143,
|
| 3639 |
+
"grad_norm": 0.09367376797157739,
|
| 3640 |
+
"learning_rate": 3.189264395172753e-08,
|
| 3641 |
+
"loss": 0.9438,
|
| 3642 |
+
"step": 519
|
| 3643 |
+
},
|
| 3644 |
+
{
|
| 3645 |
+
"epoch": 4.9523809523809526,
|
| 3646 |
+
"grad_norm": 0.09753187429670476,
|
| 3647 |
+
"learning_rate": 2.214856885427885e-08,
|
| 3648 |
+
"loss": 0.9513,
|
| 3649 |
+
"step": 520
|
| 3650 |
+
},
|
| 3651 |
+
{
|
| 3652 |
+
"epoch": 4.961904761904762,
|
| 3653 |
+
"grad_norm": 0.09256693337711726,
|
| 3654 |
+
"learning_rate": 1.4175555061894942e-08,
|
| 3655 |
+
"loss": 0.9504,
|
| 3656 |
+
"step": 521
|
| 3657 |
+
},
|
| 3658 |
+
{
|
| 3659 |
+
"epoch": 4.9714285714285715,
|
| 3660 |
+
"grad_norm": 0.09471054416033763,
|
| 3661 |
+
"learning_rate": 7.973955787567988e-09,
|
| 3662 |
+
"loss": 0.9648,
|
| 3663 |
+
"step": 522
|
| 3664 |
+
},
|
| 3665 |
+
{
|
| 3666 |
+
"epoch": 4.980952380952381,
|
| 3667 |
+
"grad_norm": 0.09324529832027807,
|
| 3668 |
+
"learning_rate": 3.544045768730797e-09,
|
| 3669 |
+
"loss": 0.9516,
|
| 3670 |
+
"step": 523
|
| 3671 |
+
},
|
| 3672 |
+
{
|
| 3673 |
+
"epoch": 4.9904761904761905,
|
| 3674 |
+
"grad_norm": 0.09644781364770945,
|
| 3675 |
+
"learning_rate": 8.860212551020653e-10,
|
| 3676 |
+
"loss": 0.9408,
|
| 3677 |
+
"step": 524
|
| 3678 |
+
},
|
| 3679 |
+
{
|
| 3680 |
+
"epoch": 5.0,
|
| 3681 |
+
"grad_norm": 0.09523670709888162,
|
| 3682 |
+
"learning_rate": 0.0,
|
| 3683 |
+
"loss": 0.9441,
|
| 3684 |
+
"step": 525
|
| 3685 |
+
},
|
| 3686 |
+
{
|
| 3687 |
+
"epoch": 5.0,
|
| 3688 |
+
"step": 525,
|
| 3689 |
+
"total_flos": 1.350981278339013e+19,
|
| 3690 |
+
"train_loss": 1.0422365560985747,
|
| 3691 |
+
"train_runtime": 52983.8768,
|
| 3692 |
+
"train_samples_per_second": 5.072,
|
| 3693 |
+
"train_steps_per_second": 0.01
|
| 3694 |
+
}
|
| 3695 |
+
],
|
| 3696 |
+
"logging_steps": 1,
|
| 3697 |
+
"max_steps": 525,
|
| 3698 |
+
"num_input_tokens_seen": 0,
|
| 3699 |
+
"num_train_epochs": 5,
|
| 3700 |
+
"save_steps": 500,
|
| 3701 |
+
"stateful_callbacks": {
|
| 3702 |
+
"TrainerControl": {
|
| 3703 |
+
"args": {
|
| 3704 |
+
"should_epoch_stop": false,
|
| 3705 |
+
"should_evaluate": false,
|
| 3706 |
+
"should_log": false,
|
| 3707 |
+
"should_save": true,
|
| 3708 |
+
"should_training_stop": true
|
| 3709 |
+
},
|
| 3710 |
+
"attributes": {}
|
| 3711 |
+
}
|
| 3712 |
+
},
|
| 3713 |
+
"total_flos": 1.350981278339013e+19,
|
| 3714 |
+
"train_batch_size": 1,
|
| 3715 |
+
"trial_name": null,
|
| 3716 |
+
"trial_params": null
|
| 3717 |
+
}
|
training_loss.png
ADDED
|