End of training
Browse files- README.md +5 -3
- all_results.json +15 -0
- eval_results.json +9 -0
- train_results.json +9 -0
- trainer_state.json +2896 -0
README.md
CHANGED
|
@@ -2,6 +2,8 @@
|
|
| 2 |
license: apache-2.0
|
| 3 |
base_model: facebook/wav2vec2-xls-r-300m
|
| 4 |
tags:
|
|
|
|
|
|
|
| 5 |
- generated_from_trainer
|
| 6 |
metrics:
|
| 7 |
- wer
|
|
@@ -16,10 +18,10 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/abdouaziz/STT-FULFULDE/runs/779i64wa)
|
| 17 |
# wav2vec2-xls-r-300m-fula
|
| 18 |
|
| 19 |
-
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
-
- Loss: 0.
|
| 22 |
-
- Wer: 1.
|
| 23 |
|
| 24 |
## Model description
|
| 25 |
|
|
|
|
| 2 |
license: apache-2.0
|
| 3 |
base_model: facebook/wav2vec2-xls-r-300m
|
| 4 |
tags:
|
| 5 |
+
- automatic-speech-recognition
|
| 6 |
+
- abdouaziiz/fula_without_num
|
| 7 |
- generated_from_trainer
|
| 8 |
metrics:
|
| 9 |
- wer
|
|
|
|
| 18 |
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/abdouaziz/STT-FULFULDE/runs/779i64wa)
|
| 19 |
# wav2vec2-xls-r-300m-fula
|
| 20 |
|
| 21 |
+
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the ABDOUAZIIZ/FULA_WITHOUT_NUM - NA dataset.
|
| 22 |
It achieves the following results on the evaluation set:
|
| 23 |
+
- Loss: 0.4631
|
| 24 |
+
- Wer: 1.0057
|
| 25 |
|
| 26 |
## Model description
|
| 27 |
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.154639175257732,
|
| 3 |
+
"eval_loss": 0.46308374404907227,
|
| 4 |
+
"eval_runtime": 187.2843,
|
| 5 |
+
"eval_samples": 1225,
|
| 6 |
+
"eval_samples_per_second": 6.541,
|
| 7 |
+
"eval_steps_per_second": 0.822,
|
| 8 |
+
"eval_wer": 1.0057142857142858,
|
| 9 |
+
"total_flos": 5.52048529474528e+19,
|
| 10 |
+
"train_loss": 0.6429409675121307,
|
| 11 |
+
"train_runtime": 62307.8234,
|
| 12 |
+
"train_samples": 46550,
|
| 13 |
+
"train_samples_per_second": 22.413,
|
| 14 |
+
"train_steps_per_second": 0.934
|
| 15 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.154639175257732,
|
| 3 |
+
"eval_loss": 0.46308374404907227,
|
| 4 |
+
"eval_runtime": 187.2843,
|
| 5 |
+
"eval_samples": 1225,
|
| 6 |
+
"eval_samples_per_second": 6.541,
|
| 7 |
+
"eval_steps_per_second": 0.822,
|
| 8 |
+
"eval_wer": 1.0057142857142858
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.154639175257732,
|
| 3 |
+
"total_flos": 5.52048529474528e+19,
|
| 4 |
+
"train_loss": 0.6429409675121307,
|
| 5 |
+
"train_runtime": 62307.8234,
|
| 6 |
+
"train_samples": 46550,
|
| 7 |
+
"train_samples_per_second": 22.413,
|
| 8 |
+
"train_steps_per_second": 0.934
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2896 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.46308374404907227,
|
| 3 |
+
"best_model_checkpoint": "wav2vec2-xls-r-300m-fula/checkpoint-4000",
|
| 4 |
+
"epoch": 5.154639175257732,
|
| 5 |
+
"eval_steps": 2000,
|
| 6 |
+
"global_step": 10000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.01288659793814433,
|
| 13 |
+
"grad_norm": 10.90772533416748,
|
| 14 |
+
"learning_rate": 3.4499999999999996e-06,
|
| 15 |
+
"loss": 19.1295,
|
| 16 |
+
"step": 25
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.02577319587628866,
|
| 20 |
+
"grad_norm": 12.727262496948242,
|
| 21 |
+
"learning_rate": 7.2e-06,
|
| 22 |
+
"loss": 16.6655,
|
| 23 |
+
"step": 50
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.03865979381443299,
|
| 27 |
+
"grad_norm": 17.578950881958008,
|
| 28 |
+
"learning_rate": 1.0949999999999998e-05,
|
| 29 |
+
"loss": 13.462,
|
| 30 |
+
"step": 75
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.05154639175257732,
|
| 34 |
+
"grad_norm": 14.574068069458008,
|
| 35 |
+
"learning_rate": 1.47e-05,
|
| 36 |
+
"loss": 7.4496,
|
| 37 |
+
"step": 100
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.06443298969072164,
|
| 41 |
+
"grad_norm": 11.906006813049316,
|
| 42 |
+
"learning_rate": 1.8449999999999998e-05,
|
| 43 |
+
"loss": 6.1341,
|
| 44 |
+
"step": 125
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.07731958762886598,
|
| 48 |
+
"grad_norm": 10.743711471557617,
|
| 49 |
+
"learning_rate": 2.2199999999999998e-05,
|
| 50 |
+
"loss": 5.0479,
|
| 51 |
+
"step": 150
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.09020618556701031,
|
| 55 |
+
"grad_norm": 9.327588081359863,
|
| 56 |
+
"learning_rate": 2.5949999999999997e-05,
|
| 57 |
+
"loss": 4.86,
|
| 58 |
+
"step": 175
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.10309278350515463,
|
| 62 |
+
"grad_norm": 8.118748664855957,
|
| 63 |
+
"learning_rate": 2.97e-05,
|
| 64 |
+
"loss": 4.3188,
|
| 65 |
+
"step": 200
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.11597938144329897,
|
| 69 |
+
"grad_norm": 4.968245983123779,
|
| 70 |
+
"learning_rate": 3.345e-05,
|
| 71 |
+
"loss": 4.142,
|
| 72 |
+
"step": 225
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.12886597938144329,
|
| 76 |
+
"grad_norm": 3.5491788387298584,
|
| 77 |
+
"learning_rate": 3.7199999999999996e-05,
|
| 78 |
+
"loss": 3.7329,
|
| 79 |
+
"step": 250
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.14175257731958762,
|
| 83 |
+
"grad_norm": 2.228337049484253,
|
| 84 |
+
"learning_rate": 4.095e-05,
|
| 85 |
+
"loss": 3.5432,
|
| 86 |
+
"step": 275
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.15463917525773196,
|
| 90 |
+
"grad_norm": 1.3472025394439697,
|
| 91 |
+
"learning_rate": 4.4699999999999996e-05,
|
| 92 |
+
"loss": 3.3229,
|
| 93 |
+
"step": 300
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.16752577319587628,
|
| 97 |
+
"grad_norm": 0.8919354677200317,
|
| 98 |
+
"learning_rate": 4.845e-05,
|
| 99 |
+
"loss": 3.1852,
|
| 100 |
+
"step": 325
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.18041237113402062,
|
| 104 |
+
"grad_norm": 0.9615734815597534,
|
| 105 |
+
"learning_rate": 5.2199999999999995e-05,
|
| 106 |
+
"loss": 3.0339,
|
| 107 |
+
"step": 350
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.19329896907216496,
|
| 111 |
+
"grad_norm": 1.0601171255111694,
|
| 112 |
+
"learning_rate": 5.595e-05,
|
| 113 |
+
"loss": 2.9307,
|
| 114 |
+
"step": 375
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.20618556701030927,
|
| 118 |
+
"grad_norm": 1.2133561372756958,
|
| 119 |
+
"learning_rate": 5.97e-05,
|
| 120 |
+
"loss": 2.7395,
|
| 121 |
+
"step": 400
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.2190721649484536,
|
| 125 |
+
"grad_norm": 0.922024130821228,
|
| 126 |
+
"learning_rate": 6.345e-05,
|
| 127 |
+
"loss": 2.5765,
|
| 128 |
+
"step": 425
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.23195876288659795,
|
| 132 |
+
"grad_norm": 1.0528242588043213,
|
| 133 |
+
"learning_rate": 6.72e-05,
|
| 134 |
+
"loss": 2.2256,
|
| 135 |
+
"step": 450
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.24484536082474226,
|
| 139 |
+
"grad_norm": 1.39167058467865,
|
| 140 |
+
"learning_rate": 7.094999999999999e-05,
|
| 141 |
+
"loss": 1.7335,
|
| 142 |
+
"step": 475
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.25773195876288657,
|
| 146 |
+
"grad_norm": 1.182673692703247,
|
| 147 |
+
"learning_rate": 7.47e-05,
|
| 148 |
+
"loss": 1.2321,
|
| 149 |
+
"step": 500
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.2706185567010309,
|
| 153 |
+
"grad_norm": 0.9412080645561218,
|
| 154 |
+
"learning_rate": 7.845e-05,
|
| 155 |
+
"loss": 0.9694,
|
| 156 |
+
"step": 525
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.28350515463917525,
|
| 160 |
+
"grad_norm": 1.3718018531799316,
|
| 161 |
+
"learning_rate": 8.22e-05,
|
| 162 |
+
"loss": 0.9061,
|
| 163 |
+
"step": 550
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.2963917525773196,
|
| 167 |
+
"grad_norm": 0.7417821884155273,
|
| 168 |
+
"learning_rate": 8.594999999999999e-05,
|
| 169 |
+
"loss": 0.8102,
|
| 170 |
+
"step": 575
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.30927835051546393,
|
| 174 |
+
"grad_norm": 0.9690466523170471,
|
| 175 |
+
"learning_rate": 8.969999999999998e-05,
|
| 176 |
+
"loss": 0.7434,
|
| 177 |
+
"step": 600
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.32216494845360827,
|
| 181 |
+
"grad_norm": 0.8293886780738831,
|
| 182 |
+
"learning_rate": 9.345e-05,
|
| 183 |
+
"loss": 0.7381,
|
| 184 |
+
"step": 625
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.33505154639175255,
|
| 188 |
+
"grad_norm": 0.8486959934234619,
|
| 189 |
+
"learning_rate": 9.719999999999999e-05,
|
| 190 |
+
"loss": 0.6224,
|
| 191 |
+
"step": 650
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.3479381443298969,
|
| 195 |
+
"grad_norm": 0.7055938243865967,
|
| 196 |
+
"learning_rate": 0.00010095,
|
| 197 |
+
"loss": 0.6925,
|
| 198 |
+
"step": 675
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.36082474226804123,
|
| 202 |
+
"grad_norm": 0.7605512142181396,
|
| 203 |
+
"learning_rate": 0.00010469999999999998,
|
| 204 |
+
"loss": 0.6136,
|
| 205 |
+
"step": 700
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.37371134020618557,
|
| 209 |
+
"grad_norm": 0.6695568561553955,
|
| 210 |
+
"learning_rate": 0.00010845,
|
| 211 |
+
"loss": 0.6419,
|
| 212 |
+
"step": 725
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.3865979381443299,
|
| 216 |
+
"grad_norm": 0.8840702772140503,
|
| 217 |
+
"learning_rate": 0.00011219999999999999,
|
| 218 |
+
"loss": 0.6667,
|
| 219 |
+
"step": 750
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.39948453608247425,
|
| 223 |
+
"grad_norm": 0.7186158299446106,
|
| 224 |
+
"learning_rate": 0.00011595,
|
| 225 |
+
"loss": 0.6397,
|
| 226 |
+
"step": 775
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.41237113402061853,
|
| 230 |
+
"grad_norm": 0.6683000922203064,
|
| 231 |
+
"learning_rate": 0.0001197,
|
| 232 |
+
"loss": 0.5647,
|
| 233 |
+
"step": 800
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.4252577319587629,
|
| 237 |
+
"grad_norm": 0.8537669777870178,
|
| 238 |
+
"learning_rate": 0.00012345,
|
| 239 |
+
"loss": 0.5624,
|
| 240 |
+
"step": 825
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.4381443298969072,
|
| 244 |
+
"grad_norm": 1.0267616510391235,
|
| 245 |
+
"learning_rate": 0.00012719999999999997,
|
| 246 |
+
"loss": 0.624,
|
| 247 |
+
"step": 850
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.45103092783505155,
|
| 251 |
+
"grad_norm": 0.6453070640563965,
|
| 252 |
+
"learning_rate": 0.00013094999999999998,
|
| 253 |
+
"loss": 0.5539,
|
| 254 |
+
"step": 875
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.4639175257731959,
|
| 258 |
+
"grad_norm": 1.0308513641357422,
|
| 259 |
+
"learning_rate": 0.0001347,
|
| 260 |
+
"loss": 0.5728,
|
| 261 |
+
"step": 900
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.47680412371134023,
|
| 265 |
+
"grad_norm": 0.560512125492096,
|
| 266 |
+
"learning_rate": 0.00013845,
|
| 267 |
+
"loss": 0.5305,
|
| 268 |
+
"step": 925
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.4896907216494845,
|
| 272 |
+
"grad_norm": 0.7196568250656128,
|
| 273 |
+
"learning_rate": 0.0001422,
|
| 274 |
+
"loss": 0.5669,
|
| 275 |
+
"step": 950
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.5025773195876289,
|
| 279 |
+
"grad_norm": 0.5675578713417053,
|
| 280 |
+
"learning_rate": 0.00014595,
|
| 281 |
+
"loss": 0.5093,
|
| 282 |
+
"step": 975
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.5154639175257731,
|
| 286 |
+
"grad_norm": 0.8834163546562195,
|
| 287 |
+
"learning_rate": 0.00014969999999999998,
|
| 288 |
+
"loss": 0.5757,
|
| 289 |
+
"step": 1000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.5283505154639175,
|
| 293 |
+
"grad_norm": 0.5083448886871338,
|
| 294 |
+
"learning_rate": 0.00015344999999999996,
|
| 295 |
+
"loss": 0.5412,
|
| 296 |
+
"step": 1025
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.5412371134020618,
|
| 300 |
+
"grad_norm": 0.747595489025116,
|
| 301 |
+
"learning_rate": 0.0001572,
|
| 302 |
+
"loss": 0.5849,
|
| 303 |
+
"step": 1050
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.5541237113402062,
|
| 307 |
+
"grad_norm": 0.5015640258789062,
|
| 308 |
+
"learning_rate": 0.00016094999999999998,
|
| 309 |
+
"loss": 0.5049,
|
| 310 |
+
"step": 1075
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.5670103092783505,
|
| 314 |
+
"grad_norm": 0.7311388850212097,
|
| 315 |
+
"learning_rate": 0.0001647,
|
| 316 |
+
"loss": 0.4726,
|
| 317 |
+
"step": 1100
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.5798969072164949,
|
| 321 |
+
"grad_norm": 0.6942028999328613,
|
| 322 |
+
"learning_rate": 0.00016844999999999997,
|
| 323 |
+
"loss": 0.4934,
|
| 324 |
+
"step": 1125
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.5927835051546392,
|
| 328 |
+
"grad_norm": 0.7268182635307312,
|
| 329 |
+
"learning_rate": 0.00017219999999999998,
|
| 330 |
+
"loss": 0.5074,
|
| 331 |
+
"step": 1150
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.6056701030927835,
|
| 335 |
+
"grad_norm": 0.5172975063323975,
|
| 336 |
+
"learning_rate": 0.00017595,
|
| 337 |
+
"loss": 0.5293,
|
| 338 |
+
"step": 1175
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.6185567010309279,
|
| 342 |
+
"grad_norm": 0.5973320603370667,
|
| 343 |
+
"learning_rate": 0.00017969999999999998,
|
| 344 |
+
"loss": 0.5018,
|
| 345 |
+
"step": 1200
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.6314432989690721,
|
| 349 |
+
"grad_norm": 0.6601810455322266,
|
| 350 |
+
"learning_rate": 0.00018345,
|
| 351 |
+
"loss": 0.4953,
|
| 352 |
+
"step": 1225
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.6443298969072165,
|
| 356 |
+
"grad_norm": 0.7992896437644958,
|
| 357 |
+
"learning_rate": 0.0001872,
|
| 358 |
+
"loss": 0.4748,
|
| 359 |
+
"step": 1250
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.6572164948453608,
|
| 363 |
+
"grad_norm": 0.5774939060211182,
|
| 364 |
+
"learning_rate": 0.00019094999999999998,
|
| 365 |
+
"loss": 0.4659,
|
| 366 |
+
"step": 1275
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.6701030927835051,
|
| 370 |
+
"grad_norm": 0.7045702338218689,
|
| 371 |
+
"learning_rate": 0.0001947,
|
| 372 |
+
"loss": 0.4433,
|
| 373 |
+
"step": 1300
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.6829896907216495,
|
| 377 |
+
"grad_norm": 0.6067873239517212,
|
| 378 |
+
"learning_rate": 0.00019844999999999997,
|
| 379 |
+
"loss": 0.4872,
|
| 380 |
+
"step": 1325
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.6958762886597938,
|
| 384 |
+
"grad_norm": 1.0101662874221802,
|
| 385 |
+
"learning_rate": 0.0002022,
|
| 386 |
+
"loss": 0.4933,
|
| 387 |
+
"step": 1350
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.7087628865979382,
|
| 391 |
+
"grad_norm": 0.4532999098300934,
|
| 392 |
+
"learning_rate": 0.00020595,
|
| 393 |
+
"loss": 0.474,
|
| 394 |
+
"step": 1375
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.7216494845360825,
|
| 398 |
+
"grad_norm": 0.8769963383674622,
|
| 399 |
+
"learning_rate": 0.00020969999999999997,
|
| 400 |
+
"loss": 0.4923,
|
| 401 |
+
"step": 1400
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.7345360824742269,
|
| 405 |
+
"grad_norm": 0.38735705614089966,
|
| 406 |
+
"learning_rate": 0.00021344999999999998,
|
| 407 |
+
"loss": 0.5358,
|
| 408 |
+
"step": 1425
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.7474226804123711,
|
| 412 |
+
"grad_norm": 0.5298680663108826,
|
| 413 |
+
"learning_rate": 0.00021719999999999997,
|
| 414 |
+
"loss": 0.4961,
|
| 415 |
+
"step": 1450
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.7603092783505154,
|
| 419 |
+
"grad_norm": 0.6393166780471802,
|
| 420 |
+
"learning_rate": 0.00022095,
|
| 421 |
+
"loss": 0.5057,
|
| 422 |
+
"step": 1475
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.7731958762886598,
|
| 426 |
+
"grad_norm": 1.908589243888855,
|
| 427 |
+
"learning_rate": 0.0002247,
|
| 428 |
+
"loss": 0.5043,
|
| 429 |
+
"step": 1500
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.7860824742268041,
|
| 433 |
+
"grad_norm": 0.5039921402931213,
|
| 434 |
+
"learning_rate": 0.00022844999999999997,
|
| 435 |
+
"loss": 0.4767,
|
| 436 |
+
"step": 1525
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.7989690721649485,
|
| 440 |
+
"grad_norm": 0.5750266909599304,
|
| 441 |
+
"learning_rate": 0.00023219999999999998,
|
| 442 |
+
"loss": 0.5002,
|
| 443 |
+
"step": 1550
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.8118556701030928,
|
| 447 |
+
"grad_norm": 0.5906339883804321,
|
| 448 |
+
"learning_rate": 0.00023594999999999996,
|
| 449 |
+
"loss": 0.4776,
|
| 450 |
+
"step": 1575
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.8247422680412371,
|
| 454 |
+
"grad_norm": 0.5718595385551453,
|
| 455 |
+
"learning_rate": 0.0002397,
|
| 456 |
+
"loss": 0.4934,
|
| 457 |
+
"step": 1600
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.8376288659793815,
|
| 461 |
+
"grad_norm": 0.5280390381813049,
|
| 462 |
+
"learning_rate": 0.00024344999999999998,
|
| 463 |
+
"loss": 0.4533,
|
| 464 |
+
"step": 1625
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.8505154639175257,
|
| 468 |
+
"grad_norm": 1.016766905784607,
|
| 469 |
+
"learning_rate": 0.0002472,
|
| 470 |
+
"loss": 0.4795,
|
| 471 |
+
"step": 1650
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.8634020618556701,
|
| 475 |
+
"grad_norm": 0.4700835943222046,
|
| 476 |
+
"learning_rate": 0.00025095,
|
| 477 |
+
"loss": 0.4399,
|
| 478 |
+
"step": 1675
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.8762886597938144,
|
| 482 |
+
"grad_norm": 0.7027618885040283,
|
| 483 |
+
"learning_rate": 0.00025469999999999996,
|
| 484 |
+
"loss": 0.4801,
|
| 485 |
+
"step": 1700
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.8891752577319587,
|
| 489 |
+
"grad_norm": 0.41857969760894775,
|
| 490 |
+
"learning_rate": 0.00025844999999999997,
|
| 491 |
+
"loss": 0.4449,
|
| 492 |
+
"step": 1725
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.9020618556701031,
|
| 496 |
+
"grad_norm": 0.7206704020500183,
|
| 497 |
+
"learning_rate": 0.0002622,
|
| 498 |
+
"loss": 0.4648,
|
| 499 |
+
"step": 1750
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.9149484536082474,
|
| 503 |
+
"grad_norm": 0.485895037651062,
|
| 504 |
+
"learning_rate": 0.00026595,
|
| 505 |
+
"loss": 0.4598,
|
| 506 |
+
"step": 1775
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.9278350515463918,
|
| 510 |
+
"grad_norm": 0.7267158031463623,
|
| 511 |
+
"learning_rate": 0.0002697,
|
| 512 |
+
"loss": 0.462,
|
| 513 |
+
"step": 1800
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.9407216494845361,
|
| 517 |
+
"grad_norm": 0.4129009246826172,
|
| 518 |
+
"learning_rate": 0.00027344999999999995,
|
| 519 |
+
"loss": 0.4585,
|
| 520 |
+
"step": 1825
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.9536082474226805,
|
| 524 |
+
"grad_norm": 0.6467506885528564,
|
| 525 |
+
"learning_rate": 0.0002772,
|
| 526 |
+
"loss": 0.4539,
|
| 527 |
+
"step": 1850
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.9664948453608248,
|
| 531 |
+
"grad_norm": 0.6241980791091919,
|
| 532 |
+
"learning_rate": 0.00028094999999999997,
|
| 533 |
+
"loss": 0.4917,
|
| 534 |
+
"step": 1875
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.979381443298969,
|
| 538 |
+
"grad_norm": 0.817642867565155,
|
| 539 |
+
"learning_rate": 0.0002847,
|
| 540 |
+
"loss": 0.4337,
|
| 541 |
+
"step": 1900
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.9922680412371134,
|
| 545 |
+
"grad_norm": 1.18275785446167,
|
| 546 |
+
"learning_rate": 0.00028845,
|
| 547 |
+
"loss": 0.4591,
|
| 548 |
+
"step": 1925
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 1.0051546391752577,
|
| 552 |
+
"grad_norm": 0.5338313579559326,
|
| 553 |
+
"learning_rate": 0.00029219999999999995,
|
| 554 |
+
"loss": 0.5187,
|
| 555 |
+
"step": 1950
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 1.018041237113402,
|
| 559 |
+
"grad_norm": 0.40176838636398315,
|
| 560 |
+
"learning_rate": 0.00029595,
|
| 561 |
+
"loss": 0.4012,
|
| 562 |
+
"step": 1975
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 1.0309278350515463,
|
| 566 |
+
"grad_norm": 0.6469115018844604,
|
| 567 |
+
"learning_rate": 0.00029969999999999997,
|
| 568 |
+
"loss": 0.4752,
|
| 569 |
+
"step": 2000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 1.0309278350515463,
|
| 573 |
+
"eval_loss": 0.59544438123703,
|
| 574 |
+
"eval_runtime": 196.5387,
|
| 575 |
+
"eval_samples_per_second": 6.233,
|
| 576 |
+
"eval_steps_per_second": 0.784,
|
| 577 |
+
"eval_wer": 1.0,
|
| 578 |
+
"step": 2000
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 1.0438144329896908,
|
| 582 |
+
"grad_norm": 0.6261550188064575,
|
| 583 |
+
"learning_rate": 0.00029987722419928825,
|
| 584 |
+
"loss": 0.4102,
|
| 585 |
+
"step": 2025
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 1.056701030927835,
|
| 589 |
+
"grad_norm": 0.8379774689674377,
|
| 590 |
+
"learning_rate": 0.00029974377224199284,
|
| 591 |
+
"loss": 0.4658,
|
| 592 |
+
"step": 2050
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 1.0695876288659794,
|
| 596 |
+
"grad_norm": 0.503239631652832,
|
| 597 |
+
"learning_rate": 0.0002996103202846975,
|
| 598 |
+
"loss": 0.4167,
|
| 599 |
+
"step": 2075
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 1.0824742268041236,
|
| 603 |
+
"grad_norm": 0.3943336606025696,
|
| 604 |
+
"learning_rate": 0.00029947686832740213,
|
| 605 |
+
"loss": 0.4686,
|
| 606 |
+
"step": 2100
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 1.0953608247422681,
|
| 610 |
+
"grad_norm": 0.4782889783382416,
|
| 611 |
+
"learning_rate": 0.0002993434163701067,
|
| 612 |
+
"loss": 0.4003,
|
| 613 |
+
"step": 2125
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 1.1082474226804124,
|
| 617 |
+
"grad_norm": 0.37270641326904297,
|
| 618 |
+
"learning_rate": 0.00029920996441281137,
|
| 619 |
+
"loss": 0.4578,
|
| 620 |
+
"step": 2150
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 1.1211340206185567,
|
| 624 |
+
"grad_norm": 0.5629658699035645,
|
| 625 |
+
"learning_rate": 0.000299076512455516,
|
| 626 |
+
"loss": 0.3687,
|
| 627 |
+
"step": 2175
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 1.134020618556701,
|
| 631 |
+
"grad_norm": 0.3787396252155304,
|
| 632 |
+
"learning_rate": 0.0002989430604982206,
|
| 633 |
+
"loss": 0.4322,
|
| 634 |
+
"step": 2200
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 1.1469072164948453,
|
| 638 |
+
"grad_norm": 0.6377401947975159,
|
| 639 |
+
"learning_rate": 0.00029880960854092526,
|
| 640 |
+
"loss": 0.4403,
|
| 641 |
+
"step": 2225
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 1.1597938144329896,
|
| 645 |
+
"grad_norm": 0.37100082635879517,
|
| 646 |
+
"learning_rate": 0.00029867615658362985,
|
| 647 |
+
"loss": 0.453,
|
| 648 |
+
"step": 2250
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 1.172680412371134,
|
| 652 |
+
"grad_norm": 0.7276923060417175,
|
| 653 |
+
"learning_rate": 0.0002985427046263345,
|
| 654 |
+
"loss": 0.4145,
|
| 655 |
+
"step": 2275
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 1.1855670103092784,
|
| 659 |
+
"grad_norm": 0.8605408668518066,
|
| 660 |
+
"learning_rate": 0.00029840925266903914,
|
| 661 |
+
"loss": 0.4539,
|
| 662 |
+
"step": 2300
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 1.1984536082474226,
|
| 666 |
+
"grad_norm": 0.45449143648147583,
|
| 667 |
+
"learning_rate": 0.00029827580071174373,
|
| 668 |
+
"loss": 0.4031,
|
| 669 |
+
"step": 2325
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 1.211340206185567,
|
| 673 |
+
"grad_norm": 0.36433449387550354,
|
| 674 |
+
"learning_rate": 0.0002981423487544484,
|
| 675 |
+
"loss": 0.4548,
|
| 676 |
+
"step": 2350
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 1.2242268041237114,
|
| 680 |
+
"grad_norm": 0.48323678970336914,
|
| 681 |
+
"learning_rate": 0.000298008896797153,
|
| 682 |
+
"loss": 0.4236,
|
| 683 |
+
"step": 2375
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 1.2371134020618557,
|
| 687 |
+
"grad_norm": 0.6236255764961243,
|
| 688 |
+
"learning_rate": 0.0002978754448398576,
|
| 689 |
+
"loss": 0.438,
|
| 690 |
+
"step": 2400
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 1.25,
|
| 694 |
+
"grad_norm": 0.721978485584259,
|
| 695 |
+
"learning_rate": 0.00029774199288256227,
|
| 696 |
+
"loss": 0.3803,
|
| 697 |
+
"step": 2425
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 1.2628865979381443,
|
| 701 |
+
"grad_norm": 0.33121246099472046,
|
| 702 |
+
"learning_rate": 0.00029760854092526686,
|
| 703 |
+
"loss": 0.4759,
|
| 704 |
+
"step": 2450
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 1.2757731958762886,
|
| 708 |
+
"grad_norm": 0.6318331360816956,
|
| 709 |
+
"learning_rate": 0.0002974750889679715,
|
| 710 |
+
"loss": 0.3616,
|
| 711 |
+
"step": 2475
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 1.2886597938144329,
|
| 715 |
+
"grad_norm": 0.30613037943840027,
|
| 716 |
+
"learning_rate": 0.0002973416370106761,
|
| 717 |
+
"loss": 0.4355,
|
| 718 |
+
"step": 2500
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 1.3015463917525774,
|
| 722 |
+
"grad_norm": 0.6804624199867249,
|
| 723 |
+
"learning_rate": 0.00029720818505338075,
|
| 724 |
+
"loss": 0.4296,
|
| 725 |
+
"step": 2525
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 1.3144329896907216,
|
| 729 |
+
"grad_norm": 0.5847098231315613,
|
| 730 |
+
"learning_rate": 0.0002970747330960854,
|
| 731 |
+
"loss": 0.4774,
|
| 732 |
+
"step": 2550
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 1.327319587628866,
|
| 736 |
+
"grad_norm": 0.43900150060653687,
|
| 737 |
+
"learning_rate": 0.00029694128113879,
|
| 738 |
+
"loss": 0.4249,
|
| 739 |
+
"step": 2575
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 1.3402061855670104,
|
| 743 |
+
"grad_norm": 0.26748043298721313,
|
| 744 |
+
"learning_rate": 0.00029680782918149463,
|
| 745 |
+
"loss": 0.4713,
|
| 746 |
+
"step": 2600
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 1.3530927835051547,
|
| 750 |
+
"grad_norm": 0.5893319845199585,
|
| 751 |
+
"learning_rate": 0.0002966743772241993,
|
| 752 |
+
"loss": 0.3566,
|
| 753 |
+
"step": 2625
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 1.365979381443299,
|
| 757 |
+
"grad_norm": 0.2939490079879761,
|
| 758 |
+
"learning_rate": 0.00029654092526690387,
|
| 759 |
+
"loss": 0.427,
|
| 760 |
+
"step": 2650
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 1.3788659793814433,
|
| 764 |
+
"grad_norm": 0.6302582025527954,
|
| 765 |
+
"learning_rate": 0.0002964074733096085,
|
| 766 |
+
"loss": 0.413,
|
| 767 |
+
"step": 2675
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 1.3917525773195876,
|
| 771 |
+
"grad_norm": 0.5481074452400208,
|
| 772 |
+
"learning_rate": 0.0002962740213523131,
|
| 773 |
+
"loss": 0.5027,
|
| 774 |
+
"step": 2700
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 1.4046391752577319,
|
| 778 |
+
"grad_norm": 0.41078057885169983,
|
| 779 |
+
"learning_rate": 0.00029614056939501776,
|
| 780 |
+
"loss": 0.3935,
|
| 781 |
+
"step": 2725
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 1.4175257731958764,
|
| 785 |
+
"grad_norm": 0.5955342054367065,
|
| 786 |
+
"learning_rate": 0.0002960071174377224,
|
| 787 |
+
"loss": 0.4454,
|
| 788 |
+
"step": 2750
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 1.4304123711340206,
|
| 792 |
+
"grad_norm": 0.47366246581077576,
|
| 793 |
+
"learning_rate": 0.000295873665480427,
|
| 794 |
+
"loss": 0.3861,
|
| 795 |
+
"step": 2775
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 1.443298969072165,
|
| 799 |
+
"grad_norm": 0.5429800748825073,
|
| 800 |
+
"learning_rate": 0.00029574021352313164,
|
| 801 |
+
"loss": 0.4525,
|
| 802 |
+
"step": 2800
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 1.4561855670103092,
|
| 806 |
+
"grad_norm": 0.5089389681816101,
|
| 807 |
+
"learning_rate": 0.00029560676156583623,
|
| 808 |
+
"loss": 0.3727,
|
| 809 |
+
"step": 2825
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 1.4690721649484537,
|
| 813 |
+
"grad_norm": 0.3353477716445923,
|
| 814 |
+
"learning_rate": 0.0002954733096085409,
|
| 815 |
+
"loss": 0.4855,
|
| 816 |
+
"step": 2850
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 1.481958762886598,
|
| 820 |
+
"grad_norm": 0.6160407662391663,
|
| 821 |
+
"learning_rate": 0.0002953398576512455,
|
| 822 |
+
"loss": 0.3635,
|
| 823 |
+
"step": 2875
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 1.4948453608247423,
|
| 827 |
+
"grad_norm": 0.3237319886684418,
|
| 828 |
+
"learning_rate": 0.0002952064056939501,
|
| 829 |
+
"loss": 0.4285,
|
| 830 |
+
"step": 2900
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 1.5077319587628866,
|
| 834 |
+
"grad_norm": 0.450309693813324,
|
| 835 |
+
"learning_rate": 0.00029507295373665477,
|
| 836 |
+
"loss": 0.3759,
|
| 837 |
+
"step": 2925
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 1.5206185567010309,
|
| 841 |
+
"grad_norm": 0.3550674319267273,
|
| 842 |
+
"learning_rate": 0.0002949395017793594,
|
| 843 |
+
"loss": 0.4442,
|
| 844 |
+
"step": 2950
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 1.5335051546391751,
|
| 848 |
+
"grad_norm": 0.3989160656929016,
|
| 849 |
+
"learning_rate": 0.000294806049822064,
|
| 850 |
+
"loss": 0.4255,
|
| 851 |
+
"step": 2975
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 1.5463917525773194,
|
| 855 |
+
"grad_norm": 0.5774266719818115,
|
| 856 |
+
"learning_rate": 0.00029467259786476865,
|
| 857 |
+
"loss": 0.4526,
|
| 858 |
+
"step": 3000
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 1.559278350515464,
|
| 862 |
+
"grad_norm": 0.6264383792877197,
|
| 863 |
+
"learning_rate": 0.0002945391459074733,
|
| 864 |
+
"loss": 0.4097,
|
| 865 |
+
"step": 3025
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 1.5721649484536082,
|
| 869 |
+
"grad_norm": 0.33969295024871826,
|
| 870 |
+
"learning_rate": 0.0002944056939501779,
|
| 871 |
+
"loss": 0.4241,
|
| 872 |
+
"step": 3050
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 1.5850515463917527,
|
| 876 |
+
"grad_norm": 0.39299994707107544,
|
| 877 |
+
"learning_rate": 0.00029427224199288254,
|
| 878 |
+
"loss": 0.3778,
|
| 879 |
+
"step": 3075
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 1.597938144329897,
|
| 883 |
+
"grad_norm": 0.42388004064559937,
|
| 884 |
+
"learning_rate": 0.0002941387900355872,
|
| 885 |
+
"loss": 0.4004,
|
| 886 |
+
"step": 3100
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 1.6108247422680413,
|
| 890 |
+
"grad_norm": 0.9131516814231873,
|
| 891 |
+
"learning_rate": 0.0002940053380782918,
|
| 892 |
+
"loss": 0.3531,
|
| 893 |
+
"step": 3125
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 1.6237113402061856,
|
| 897 |
+
"grad_norm": 0.296908974647522,
|
| 898 |
+
"learning_rate": 0.0002938718861209964,
|
| 899 |
+
"loss": 0.4613,
|
| 900 |
+
"step": 3150
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 1.6365979381443299,
|
| 904 |
+
"grad_norm": 0.4583122134208679,
|
| 905 |
+
"learning_rate": 0.000293738434163701,
|
| 906 |
+
"loss": 0.3833,
|
| 907 |
+
"step": 3175
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 1.6494845360824741,
|
| 911 |
+
"grad_norm": 0.35052600502967834,
|
| 912 |
+
"learning_rate": 0.00029360498220640566,
|
| 913 |
+
"loss": 0.4432,
|
| 914 |
+
"step": 3200
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 1.6623711340206184,
|
| 918 |
+
"grad_norm": 0.532720685005188,
|
| 919 |
+
"learning_rate": 0.0002934715302491103,
|
| 920 |
+
"loss": 0.3635,
|
| 921 |
+
"step": 3225
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 1.675257731958763,
|
| 925 |
+
"grad_norm": 0.3807854652404785,
|
| 926 |
+
"learning_rate": 0.0002933380782918149,
|
| 927 |
+
"loss": 0.43,
|
| 928 |
+
"step": 3250
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 1.6881443298969072,
|
| 932 |
+
"grad_norm": 0.5288501381874084,
|
| 933 |
+
"learning_rate": 0.00029320462633451955,
|
| 934 |
+
"loss": 0.3449,
|
| 935 |
+
"step": 3275
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 1.7010309278350515,
|
| 939 |
+
"grad_norm": 0.3881712853908539,
|
| 940 |
+
"learning_rate": 0.0002930711743772242,
|
| 941 |
+
"loss": 0.4406,
|
| 942 |
+
"step": 3300
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 1.713917525773196,
|
| 946 |
+
"grad_norm": 0.42132484912872314,
|
| 947 |
+
"learning_rate": 0.0002929377224199288,
|
| 948 |
+
"loss": 0.3888,
|
| 949 |
+
"step": 3325
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 1.7268041237113403,
|
| 953 |
+
"grad_norm": 0.3974430561065674,
|
| 954 |
+
"learning_rate": 0.00029280427046263343,
|
| 955 |
+
"loss": 0.4265,
|
| 956 |
+
"step": 3350
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 1.7396907216494846,
|
| 960 |
+
"grad_norm": 0.4618494212627411,
|
| 961 |
+
"learning_rate": 0.0002926708185053381,
|
| 962 |
+
"loss": 0.3537,
|
| 963 |
+
"step": 3375
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 1.7525773195876289,
|
| 967 |
+
"grad_norm": 0.347777783870697,
|
| 968 |
+
"learning_rate": 0.00029253736654804267,
|
| 969 |
+
"loss": 0.4282,
|
| 970 |
+
"step": 3400
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 1.7654639175257731,
|
| 974 |
+
"grad_norm": 0.5885565280914307,
|
| 975 |
+
"learning_rate": 0.0002924039145907473,
|
| 976 |
+
"loss": 0.3694,
|
| 977 |
+
"step": 3425
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 1.7783505154639174,
|
| 981 |
+
"grad_norm": 0.5113171935081482,
|
| 982 |
+
"learning_rate": 0.00029227046263345197,
|
| 983 |
+
"loss": 0.5256,
|
| 984 |
+
"step": 3450
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 1.7912371134020617,
|
| 988 |
+
"grad_norm": 0.7724625468254089,
|
| 989 |
+
"learning_rate": 0.00029213701067615656,
|
| 990 |
+
"loss": 0.3649,
|
| 991 |
+
"step": 3475
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 1.8041237113402062,
|
| 995 |
+
"grad_norm": 0.44778281450271606,
|
| 996 |
+
"learning_rate": 0.0002920035587188612,
|
| 997 |
+
"loss": 0.4313,
|
| 998 |
+
"step": 3500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 1.8170103092783505,
|
| 1002 |
+
"grad_norm": 0.8251272439956665,
|
| 1003 |
+
"learning_rate": 0.00029187010676156585,
|
| 1004 |
+
"loss": 0.3839,
|
| 1005 |
+
"step": 3525
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 1.829896907216495,
|
| 1009 |
+
"grad_norm": 0.4858299493789673,
|
| 1010 |
+
"learning_rate": 0.00029173665480427044,
|
| 1011 |
+
"loss": 0.4489,
|
| 1012 |
+
"step": 3550
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 1.8427835051546393,
|
| 1016 |
+
"grad_norm": 0.5172144770622253,
|
| 1017 |
+
"learning_rate": 0.0002916032028469751,
|
| 1018 |
+
"loss": 0.409,
|
| 1019 |
+
"step": 3575
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 1.8556701030927836,
|
| 1023 |
+
"grad_norm": 0.34371522068977356,
|
| 1024 |
+
"learning_rate": 0.0002914697508896797,
|
| 1025 |
+
"loss": 0.4365,
|
| 1026 |
+
"step": 3600
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 1.8685567010309279,
|
| 1030 |
+
"grad_norm": 0.5957440137863159,
|
| 1031 |
+
"learning_rate": 0.00029133629893238433,
|
| 1032 |
+
"loss": 0.3611,
|
| 1033 |
+
"step": 3625
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 1.8814432989690721,
|
| 1037 |
+
"grad_norm": 0.3856901526451111,
|
| 1038 |
+
"learning_rate": 0.000291202846975089,
|
| 1039 |
+
"loss": 0.451,
|
| 1040 |
+
"step": 3650
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 1.8943298969072164,
|
| 1044 |
+
"grad_norm": 0.5961311459541321,
|
| 1045 |
+
"learning_rate": 0.00029106939501779357,
|
| 1046 |
+
"loss": 0.3934,
|
| 1047 |
+
"step": 3675
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 1.9072164948453607,
|
| 1051 |
+
"grad_norm": 0.4491939842700958,
|
| 1052 |
+
"learning_rate": 0.0002909359430604982,
|
| 1053 |
+
"loss": 0.4466,
|
| 1054 |
+
"step": 3700
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 1.920103092783505,
|
| 1058 |
+
"grad_norm": 0.41777607798576355,
|
| 1059 |
+
"learning_rate": 0.0002908024911032028,
|
| 1060 |
+
"loss": 0.3754,
|
| 1061 |
+
"step": 3725
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 1.9329896907216495,
|
| 1065 |
+
"grad_norm": 0.42550450563430786,
|
| 1066 |
+
"learning_rate": 0.00029066903914590745,
|
| 1067 |
+
"loss": 0.4027,
|
| 1068 |
+
"step": 3750
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 1.9458762886597938,
|
| 1072 |
+
"grad_norm": 0.427433043718338,
|
| 1073 |
+
"learning_rate": 0.0002905355871886121,
|
| 1074 |
+
"loss": 0.3603,
|
| 1075 |
+
"step": 3775
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 1.9587628865979383,
|
| 1079 |
+
"grad_norm": 0.8279537558555603,
|
| 1080 |
+
"learning_rate": 0.0002904021352313167,
|
| 1081 |
+
"loss": 0.4098,
|
| 1082 |
+
"step": 3800
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 1.9716494845360826,
|
| 1086 |
+
"grad_norm": 0.6759387850761414,
|
| 1087 |
+
"learning_rate": 0.00029026868327402134,
|
| 1088 |
+
"loss": 0.3756,
|
| 1089 |
+
"step": 3825
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 1.9845360824742269,
|
| 1093 |
+
"grad_norm": 0.29660704731941223,
|
| 1094 |
+
"learning_rate": 0.000290135231316726,
|
| 1095 |
+
"loss": 0.4591,
|
| 1096 |
+
"step": 3850
|
| 1097 |
+
},
|
| 1098 |
+
{
|
| 1099 |
+
"epoch": 1.9974226804123711,
|
| 1100 |
+
"grad_norm": 0.4846726655960083,
|
| 1101 |
+
"learning_rate": 0.0002900017793594306,
|
| 1102 |
+
"loss": 0.3944,
|
| 1103 |
+
"step": 3875
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 2.0103092783505154,
|
| 1107 |
+
"grad_norm": 0.4863591492176056,
|
| 1108 |
+
"learning_rate": 0.0002898683274021352,
|
| 1109 |
+
"loss": 0.3813,
|
| 1110 |
+
"step": 3900
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 2.0231958762886597,
|
| 1114 |
+
"grad_norm": 0.37275585532188416,
|
| 1115 |
+
"learning_rate": 0.0002897348754448398,
|
| 1116 |
+
"loss": 0.3753,
|
| 1117 |
+
"step": 3925
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 2.036082474226804,
|
| 1121 |
+
"grad_norm": 0.36628881096839905,
|
| 1122 |
+
"learning_rate": 0.00028960142348754446,
|
| 1123 |
+
"loss": 0.403,
|
| 1124 |
+
"step": 3950
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 2.0489690721649483,
|
| 1128 |
+
"grad_norm": 0.4523802697658539,
|
| 1129 |
+
"learning_rate": 0.0002894679715302491,
|
| 1130 |
+
"loss": 0.3601,
|
| 1131 |
+
"step": 3975
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 2.0618556701030926,
|
| 1135 |
+
"grad_norm": 0.4722582697868347,
|
| 1136 |
+
"learning_rate": 0.0002893345195729537,
|
| 1137 |
+
"loss": 0.3858,
|
| 1138 |
+
"step": 4000
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 2.0618556701030926,
|
| 1142 |
+
"eval_loss": 0.46308374404907227,
|
| 1143 |
+
"eval_runtime": 201.4429,
|
| 1144 |
+
"eval_samples_per_second": 6.081,
|
| 1145 |
+
"eval_steps_per_second": 0.764,
|
| 1146 |
+
"eval_wer": 1.0057142857142858,
|
| 1147 |
+
"step": 4000
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 2.0747422680412373,
|
| 1151 |
+
"grad_norm": 0.5041220784187317,
|
| 1152 |
+
"learning_rate": 0.00028920106761565835,
|
| 1153 |
+
"loss": 0.3518,
|
| 1154 |
+
"step": 4025
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 2.0876288659793816,
|
| 1158 |
+
"grad_norm": 0.48610439896583557,
|
| 1159 |
+
"learning_rate": 0.00028906761565836294,
|
| 1160 |
+
"loss": 0.3818,
|
| 1161 |
+
"step": 4050
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 2.100515463917526,
|
| 1165 |
+
"grad_norm": 0.6834833025932312,
|
| 1166 |
+
"learning_rate": 0.0002889341637010676,
|
| 1167 |
+
"loss": 0.3778,
|
| 1168 |
+
"step": 4075
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 2.11340206185567,
|
| 1172 |
+
"grad_norm": 0.687245786190033,
|
| 1173 |
+
"learning_rate": 0.00028880071174377224,
|
| 1174 |
+
"loss": 0.3771,
|
| 1175 |
+
"step": 4100
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 2.1262886597938144,
|
| 1179 |
+
"grad_norm": 0.706832230091095,
|
| 1180 |
+
"learning_rate": 0.00028866725978647683,
|
| 1181 |
+
"loss": 0.3868,
|
| 1182 |
+
"step": 4125
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 2.1391752577319587,
|
| 1186 |
+
"grad_norm": 0.7049499154090881,
|
| 1187 |
+
"learning_rate": 0.0002885338078291815,
|
| 1188 |
+
"loss": 0.3646,
|
| 1189 |
+
"step": 4150
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 2.152061855670103,
|
| 1193 |
+
"grad_norm": 0.40853607654571533,
|
| 1194 |
+
"learning_rate": 0.00028840035587188607,
|
| 1195 |
+
"loss": 0.3235,
|
| 1196 |
+
"step": 4175
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 2.1649484536082473,
|
| 1200 |
+
"grad_norm": 0.3675331771373749,
|
| 1201 |
+
"learning_rate": 0.0002882669039145907,
|
| 1202 |
+
"loss": 0.4079,
|
| 1203 |
+
"step": 4200
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 2.1778350515463916,
|
| 1207 |
+
"grad_norm": 1.3320142030715942,
|
| 1208 |
+
"learning_rate": 0.00028813345195729536,
|
| 1209 |
+
"loss": 0.3031,
|
| 1210 |
+
"step": 4225
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 2.1907216494845363,
|
| 1214 |
+
"grad_norm": 0.6416336297988892,
|
| 1215 |
+
"learning_rate": 0.00028799999999999995,
|
| 1216 |
+
"loss": 0.4087,
|
| 1217 |
+
"step": 4250
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 2.2036082474226806,
|
| 1221 |
+
"grad_norm": 0.43675994873046875,
|
| 1222 |
+
"learning_rate": 0.0002878665480427046,
|
| 1223 |
+
"loss": 0.3577,
|
| 1224 |
+
"step": 4275
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 2.216494845360825,
|
| 1228 |
+
"grad_norm": 0.3541963994503021,
|
| 1229 |
+
"learning_rate": 0.00028773309608540925,
|
| 1230 |
+
"loss": 0.3706,
|
| 1231 |
+
"step": 4300
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 2.229381443298969,
|
| 1235 |
+
"grad_norm": 0.4469320476055145,
|
| 1236 |
+
"learning_rate": 0.00028759964412811384,
|
| 1237 |
+
"loss": 0.3692,
|
| 1238 |
+
"step": 4325
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 2.2422680412371134,
|
| 1242 |
+
"grad_norm": 0.4056352376937866,
|
| 1243 |
+
"learning_rate": 0.0002874661921708185,
|
| 1244 |
+
"loss": 0.3659,
|
| 1245 |
+
"step": 4350
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 2.2551546391752577,
|
| 1249 |
+
"grad_norm": 0.3547820746898651,
|
| 1250 |
+
"learning_rate": 0.0002873327402135231,
|
| 1251 |
+
"loss": 0.3564,
|
| 1252 |
+
"step": 4375
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 2.268041237113402,
|
| 1256 |
+
"grad_norm": 0.32645678520202637,
|
| 1257 |
+
"learning_rate": 0.0002871992882562277,
|
| 1258 |
+
"loss": 0.3346,
|
| 1259 |
+
"step": 4400
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 2.2809278350515463,
|
| 1263 |
+
"grad_norm": 0.7449667453765869,
|
| 1264 |
+
"learning_rate": 0.00028706583629893237,
|
| 1265 |
+
"loss": 0.3696,
|
| 1266 |
+
"step": 4425
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 2.2938144329896906,
|
| 1270 |
+
"grad_norm": 0.4612623155117035,
|
| 1271 |
+
"learning_rate": 0.00028693238434163696,
|
| 1272 |
+
"loss": 0.3925,
|
| 1273 |
+
"step": 4450
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 2.306701030927835,
|
| 1277 |
+
"grad_norm": 0.6325463056564331,
|
| 1278 |
+
"learning_rate": 0.0002867989323843416,
|
| 1279 |
+
"loss": 0.3512,
|
| 1280 |
+
"step": 4475
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 2.319587628865979,
|
| 1284 |
+
"grad_norm": 0.29471156001091003,
|
| 1285 |
+
"learning_rate": 0.0002866654804270462,
|
| 1286 |
+
"loss": 0.3569,
|
| 1287 |
+
"step": 4500
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 2.332474226804124,
|
| 1291 |
+
"grad_norm": 1.076217770576477,
|
| 1292 |
+
"learning_rate": 0.00028653202846975085,
|
| 1293 |
+
"loss": 0.3375,
|
| 1294 |
+
"step": 4525
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 2.345360824742268,
|
| 1298 |
+
"grad_norm": 0.614162027835846,
|
| 1299 |
+
"learning_rate": 0.0002863985765124555,
|
| 1300 |
+
"loss": 0.4078,
|
| 1301 |
+
"step": 4550
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 2.3582474226804124,
|
| 1305 |
+
"grad_norm": 0.4514384865760803,
|
| 1306 |
+
"learning_rate": 0.0002862651245551601,
|
| 1307 |
+
"loss": 0.3758,
|
| 1308 |
+
"step": 4575
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 2.3711340206185567,
|
| 1312 |
+
"grad_norm": 0.314336359500885,
|
| 1313 |
+
"learning_rate": 0.00028613167259786473,
|
| 1314 |
+
"loss": 0.3765,
|
| 1315 |
+
"step": 4600
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 2.384020618556701,
|
| 1319 |
+
"grad_norm": 0.5032577514648438,
|
| 1320 |
+
"learning_rate": 0.0002859982206405694,
|
| 1321 |
+
"loss": 0.3226,
|
| 1322 |
+
"step": 4625
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 2.3969072164948453,
|
| 1326 |
+
"grad_norm": 0.6402880549430847,
|
| 1327 |
+
"learning_rate": 0.000285864768683274,
|
| 1328 |
+
"loss": 0.3986,
|
| 1329 |
+
"step": 4650
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 2.4097938144329896,
|
| 1333 |
+
"grad_norm": 0.5510269403457642,
|
| 1334 |
+
"learning_rate": 0.0002857313167259786,
|
| 1335 |
+
"loss": 0.3757,
|
| 1336 |
+
"step": 4675
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 2.422680412371134,
|
| 1340 |
+
"grad_norm": 0.5594025254249573,
|
| 1341 |
+
"learning_rate": 0.0002855978647686832,
|
| 1342 |
+
"loss": 0.4297,
|
| 1343 |
+
"step": 4700
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 2.4355670103092786,
|
| 1347 |
+
"grad_norm": 0.3621445894241333,
|
| 1348 |
+
"learning_rate": 0.00028546441281138786,
|
| 1349 |
+
"loss": 0.3336,
|
| 1350 |
+
"step": 4725
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 2.448453608247423,
|
| 1354 |
+
"grad_norm": 0.5683943033218384,
|
| 1355 |
+
"learning_rate": 0.0002853309608540925,
|
| 1356 |
+
"loss": 0.3395,
|
| 1357 |
+
"step": 4750
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 2.461340206185567,
|
| 1361 |
+
"grad_norm": 0.7661644816398621,
|
| 1362 |
+
"learning_rate": 0.0002851975088967971,
|
| 1363 |
+
"loss": 0.377,
|
| 1364 |
+
"step": 4775
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 2.4742268041237114,
|
| 1368 |
+
"grad_norm": 0.456636905670166,
|
| 1369 |
+
"learning_rate": 0.00028506405693950175,
|
| 1370 |
+
"loss": 0.3645,
|
| 1371 |
+
"step": 4800
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 2.4871134020618557,
|
| 1375 |
+
"grad_norm": 0.49251827597618103,
|
| 1376 |
+
"learning_rate": 0.0002849306049822064,
|
| 1377 |
+
"loss": 0.34,
|
| 1378 |
+
"step": 4825
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 2.5,
|
| 1382 |
+
"grad_norm": 0.32308971881866455,
|
| 1383 |
+
"learning_rate": 0.000284797153024911,
|
| 1384 |
+
"loss": 0.4092,
|
| 1385 |
+
"step": 4850
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 2.5128865979381443,
|
| 1389 |
+
"grad_norm": 0.9722476005554199,
|
| 1390 |
+
"learning_rate": 0.00028466370106761563,
|
| 1391 |
+
"loss": 0.3651,
|
| 1392 |
+
"step": 4875
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 2.5257731958762886,
|
| 1396 |
+
"grad_norm": 0.4552549123764038,
|
| 1397 |
+
"learning_rate": 0.0002845302491103203,
|
| 1398 |
+
"loss": 0.3837,
|
| 1399 |
+
"step": 4900
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 2.538659793814433,
|
| 1403 |
+
"grad_norm": 0.4148350656032562,
|
| 1404 |
+
"learning_rate": 0.00028439679715302487,
|
| 1405 |
+
"loss": 0.3931,
|
| 1406 |
+
"step": 4925
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 2.551546391752577,
|
| 1410 |
+
"grad_norm": 0.41644537448883057,
|
| 1411 |
+
"learning_rate": 0.0002842633451957295,
|
| 1412 |
+
"loss": 0.3677,
|
| 1413 |
+
"step": 4950
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 2.5644329896907214,
|
| 1417 |
+
"grad_norm": 0.3989255428314209,
|
| 1418 |
+
"learning_rate": 0.00028412989323843416,
|
| 1419 |
+
"loss": 0.3759,
|
| 1420 |
+
"step": 4975
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 2.5773195876288657,
|
| 1424 |
+
"grad_norm": 0.472091943025589,
|
| 1425 |
+
"learning_rate": 0.00028399644128113876,
|
| 1426 |
+
"loss": 0.4008,
|
| 1427 |
+
"step": 5000
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 2.5902061855670104,
|
| 1431 |
+
"grad_norm": 0.9913691282272339,
|
| 1432 |
+
"learning_rate": 0.0002838629893238434,
|
| 1433 |
+
"loss": 0.3555,
|
| 1434 |
+
"step": 5025
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 2.6030927835051547,
|
| 1438 |
+
"grad_norm": 0.425589382648468,
|
| 1439 |
+
"learning_rate": 0.000283729537366548,
|
| 1440 |
+
"loss": 0.3774,
|
| 1441 |
+
"step": 5050
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 2.615979381443299,
|
| 1445 |
+
"grad_norm": 0.7836791276931763,
|
| 1446 |
+
"learning_rate": 0.00028359608540925264,
|
| 1447 |
+
"loss": 0.317,
|
| 1448 |
+
"step": 5075
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 2.6288659793814433,
|
| 1452 |
+
"grad_norm": 0.7071629166603088,
|
| 1453 |
+
"learning_rate": 0.0002834626334519573,
|
| 1454 |
+
"loss": 0.4039,
|
| 1455 |
+
"step": 5100
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 2.6417525773195876,
|
| 1459 |
+
"grad_norm": 0.5756880640983582,
|
| 1460 |
+
"learning_rate": 0.0002833291814946619,
|
| 1461 |
+
"loss": 0.342,
|
| 1462 |
+
"step": 5125
|
| 1463 |
+
},
|
| 1464 |
+
{
|
| 1465 |
+
"epoch": 2.654639175257732,
|
| 1466 |
+
"grad_norm": 0.425029993057251,
|
| 1467 |
+
"learning_rate": 0.0002831957295373665,
|
| 1468 |
+
"loss": 0.3765,
|
| 1469 |
+
"step": 5150
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 2.667525773195876,
|
| 1473 |
+
"grad_norm": 0.5212023854255676,
|
| 1474 |
+
"learning_rate": 0.0002830622775800712,
|
| 1475 |
+
"loss": 0.3681,
|
| 1476 |
+
"step": 5175
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 2.680412371134021,
|
| 1480 |
+
"grad_norm": 0.39988288283348083,
|
| 1481 |
+
"learning_rate": 0.00028292882562277577,
|
| 1482 |
+
"loss": 0.3525,
|
| 1483 |
+
"step": 5200
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 2.693298969072165,
|
| 1487 |
+
"grad_norm": 0.31254854798316956,
|
| 1488 |
+
"learning_rate": 0.0002827953736654804,
|
| 1489 |
+
"loss": 0.3521,
|
| 1490 |
+
"step": 5225
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 2.7061855670103094,
|
| 1494 |
+
"grad_norm": 0.5564956665039062,
|
| 1495 |
+
"learning_rate": 0.00028266192170818506,
|
| 1496 |
+
"loss": 0.3602,
|
| 1497 |
+
"step": 5250
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 2.7190721649484537,
|
| 1501 |
+
"grad_norm": 0.45266616344451904,
|
| 1502 |
+
"learning_rate": 0.00028252846975088965,
|
| 1503 |
+
"loss": 0.3462,
|
| 1504 |
+
"step": 5275
|
| 1505 |
+
},
|
| 1506 |
+
{
|
| 1507 |
+
"epoch": 2.731958762886598,
|
| 1508 |
+
"grad_norm": 0.26552554965019226,
|
| 1509 |
+
"learning_rate": 0.0002823950177935943,
|
| 1510 |
+
"loss": 0.3787,
|
| 1511 |
+
"step": 5300
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"epoch": 2.7448453608247423,
|
| 1515 |
+
"grad_norm": 0.5850217938423157,
|
| 1516 |
+
"learning_rate": 0.00028226156583629894,
|
| 1517 |
+
"loss": 0.3289,
|
| 1518 |
+
"step": 5325
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 2.7577319587628866,
|
| 1522 |
+
"grad_norm": 0.47871604561805725,
|
| 1523 |
+
"learning_rate": 0.00028212811387900354,
|
| 1524 |
+
"loss": 0.4246,
|
| 1525 |
+
"step": 5350
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 2.770618556701031,
|
| 1529 |
+
"grad_norm": 0.5951977372169495,
|
| 1530 |
+
"learning_rate": 0.0002819946619217082,
|
| 1531 |
+
"loss": 0.3326,
|
| 1532 |
+
"step": 5375
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 2.783505154639175,
|
| 1536 |
+
"grad_norm": 0.5806294083595276,
|
| 1537 |
+
"learning_rate": 0.0002818612099644128,
|
| 1538 |
+
"loss": 0.3952,
|
| 1539 |
+
"step": 5400
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 2.7963917525773194,
|
| 1543 |
+
"grad_norm": 0.6218500733375549,
|
| 1544 |
+
"learning_rate": 0.0002817277580071174,
|
| 1545 |
+
"loss": 0.3222,
|
| 1546 |
+
"step": 5425
|
| 1547 |
+
},
|
| 1548 |
+
{
|
| 1549 |
+
"epoch": 2.8092783505154637,
|
| 1550 |
+
"grad_norm": 0.38604310154914856,
|
| 1551 |
+
"learning_rate": 0.00028159430604982207,
|
| 1552 |
+
"loss": 0.4408,
|
| 1553 |
+
"step": 5450
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 2.822164948453608,
|
| 1557 |
+
"grad_norm": 2.576910972595215,
|
| 1558 |
+
"learning_rate": 0.00028146085409252666,
|
| 1559 |
+
"loss": 0.355,
|
| 1560 |
+
"step": 5475
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 2.8350515463917527,
|
| 1564 |
+
"grad_norm": 0.28045961260795593,
|
| 1565 |
+
"learning_rate": 0.0002813274021352313,
|
| 1566 |
+
"loss": 0.3479,
|
| 1567 |
+
"step": 5500
|
| 1568 |
+
},
|
| 1569 |
+
{
|
| 1570 |
+
"epoch": 2.847938144329897,
|
| 1571 |
+
"grad_norm": 0.6178203225135803,
|
| 1572 |
+
"learning_rate": 0.00028119395017793596,
|
| 1573 |
+
"loss": 0.3097,
|
| 1574 |
+
"step": 5525
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 2.8608247422680413,
|
| 1578 |
+
"grad_norm": 0.36241769790649414,
|
| 1579 |
+
"learning_rate": 0.00028106049822064055,
|
| 1580 |
+
"loss": 0.3762,
|
| 1581 |
+
"step": 5550
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 2.8737113402061856,
|
| 1585 |
+
"grad_norm": 0.6459288597106934,
|
| 1586 |
+
"learning_rate": 0.0002809270462633452,
|
| 1587 |
+
"loss": 0.3261,
|
| 1588 |
+
"step": 5575
|
| 1589 |
+
},
|
| 1590 |
+
{
|
| 1591 |
+
"epoch": 2.88659793814433,
|
| 1592 |
+
"grad_norm": 0.458281546831131,
|
| 1593 |
+
"learning_rate": 0.0002807935943060498,
|
| 1594 |
+
"loss": 0.378,
|
| 1595 |
+
"step": 5600
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 2.899484536082474,
|
| 1599 |
+
"grad_norm": 0.4269348382949829,
|
| 1600 |
+
"learning_rate": 0.00028066014234875443,
|
| 1601 |
+
"loss": 0.3483,
|
| 1602 |
+
"step": 5625
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 2.9123711340206184,
|
| 1606 |
+
"grad_norm": 0.36018800735473633,
|
| 1607 |
+
"learning_rate": 0.0002805266903914591,
|
| 1608 |
+
"loss": 0.3592,
|
| 1609 |
+
"step": 5650
|
| 1610 |
+
},
|
| 1611 |
+
{
|
| 1612 |
+
"epoch": 2.925257731958763,
|
| 1613 |
+
"grad_norm": 0.6075245141983032,
|
| 1614 |
+
"learning_rate": 0.00028039323843416367,
|
| 1615 |
+
"loss": 0.3,
|
| 1616 |
+
"step": 5675
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 2.9381443298969074,
|
| 1620 |
+
"grad_norm": 0.3082279860973358,
|
| 1621 |
+
"learning_rate": 0.0002802597864768683,
|
| 1622 |
+
"loss": 0.4306,
|
| 1623 |
+
"step": 5700
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 2.9510309278350517,
|
| 1627 |
+
"grad_norm": 0.6344396471977234,
|
| 1628 |
+
"learning_rate": 0.0002801263345195729,
|
| 1629 |
+
"loss": 0.3795,
|
| 1630 |
+
"step": 5725
|
| 1631 |
+
},
|
| 1632 |
+
{
|
| 1633 |
+
"epoch": 2.963917525773196,
|
| 1634 |
+
"grad_norm": 0.3072182536125183,
|
| 1635 |
+
"learning_rate": 0.00027999288256227756,
|
| 1636 |
+
"loss": 0.3425,
|
| 1637 |
+
"step": 5750
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 2.9768041237113403,
|
| 1641 |
+
"grad_norm": 0.5076513886451721,
|
| 1642 |
+
"learning_rate": 0.0002798594306049822,
|
| 1643 |
+
"loss": 0.3214,
|
| 1644 |
+
"step": 5775
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 2.9896907216494846,
|
| 1648 |
+
"grad_norm": 0.34852054715156555,
|
| 1649 |
+
"learning_rate": 0.0002797259786476868,
|
| 1650 |
+
"loss": 0.3678,
|
| 1651 |
+
"step": 5800
|
| 1652 |
+
},
|
| 1653 |
+
{
|
| 1654 |
+
"epoch": 3.002577319587629,
|
| 1655 |
+
"grad_norm": 0.36334386467933655,
|
| 1656 |
+
"learning_rate": 0.00027959252669039144,
|
| 1657 |
+
"loss": 0.3984,
|
| 1658 |
+
"step": 5825
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 3.015463917525773,
|
| 1662 |
+
"grad_norm": 0.46211138367652893,
|
| 1663 |
+
"learning_rate": 0.00027945907473309604,
|
| 1664 |
+
"loss": 0.2912,
|
| 1665 |
+
"step": 5850
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 3.0283505154639174,
|
| 1669 |
+
"grad_norm": 0.3730103373527527,
|
| 1670 |
+
"learning_rate": 0.0002793256227758007,
|
| 1671 |
+
"loss": 0.337,
|
| 1672 |
+
"step": 5875
|
| 1673 |
+
},
|
| 1674 |
+
{
|
| 1675 |
+
"epoch": 3.0412371134020617,
|
| 1676 |
+
"grad_norm": 0.37066757678985596,
|
| 1677 |
+
"learning_rate": 0.00027919217081850533,
|
| 1678 |
+
"loss": 0.2959,
|
| 1679 |
+
"step": 5900
|
| 1680 |
+
},
|
| 1681 |
+
{
|
| 1682 |
+
"epoch": 3.054123711340206,
|
| 1683 |
+
"grad_norm": 0.5712897181510925,
|
| 1684 |
+
"learning_rate": 0.0002790587188612099,
|
| 1685 |
+
"loss": 0.3279,
|
| 1686 |
+
"step": 5925
|
| 1687 |
+
},
|
| 1688 |
+
{
|
| 1689 |
+
"epoch": 3.0670103092783507,
|
| 1690 |
+
"grad_norm": 0.39846915006637573,
|
| 1691 |
+
"learning_rate": 0.00027892526690391457,
|
| 1692 |
+
"loss": 0.3532,
|
| 1693 |
+
"step": 5950
|
| 1694 |
+
},
|
| 1695 |
+
{
|
| 1696 |
+
"epoch": 3.079896907216495,
|
| 1697 |
+
"grad_norm": 0.3366047143936157,
|
| 1698 |
+
"learning_rate": 0.0002787918149466192,
|
| 1699 |
+
"loss": 0.3486,
|
| 1700 |
+
"step": 5975
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 3.0927835051546393,
|
| 1704 |
+
"grad_norm": 0.5053852796554565,
|
| 1705 |
+
"learning_rate": 0.0002786583629893238,
|
| 1706 |
+
"loss": 0.3571,
|
| 1707 |
+
"step": 6000
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 3.0927835051546393,
|
| 1711 |
+
"eval_loss": 0.48639050126075745,
|
| 1712 |
+
"eval_runtime": 195.6249,
|
| 1713 |
+
"eval_samples_per_second": 6.262,
|
| 1714 |
+
"eval_steps_per_second": 0.787,
|
| 1715 |
+
"eval_wer": 1.0073469387755103,
|
| 1716 |
+
"step": 6000
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 3.1056701030927836,
|
| 1720 |
+
"grad_norm": 0.35252827405929565,
|
| 1721 |
+
"learning_rate": 0.00027852491103202845,
|
| 1722 |
+
"loss": 0.3401,
|
| 1723 |
+
"step": 6025
|
| 1724 |
+
},
|
| 1725 |
+
{
|
| 1726 |
+
"epoch": 3.118556701030928,
|
| 1727 |
+
"grad_norm": 0.3530094027519226,
|
| 1728 |
+
"learning_rate": 0.00027839145907473305,
|
| 1729 |
+
"loss": 0.3094,
|
| 1730 |
+
"step": 6050
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 3.131443298969072,
|
| 1734 |
+
"grad_norm": 0.9595320224761963,
|
| 1735 |
+
"learning_rate": 0.0002782580071174377,
|
| 1736 |
+
"loss": 0.3692,
|
| 1737 |
+
"step": 6075
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 3.1443298969072164,
|
| 1741 |
+
"grad_norm": 0.3585176467895508,
|
| 1742 |
+
"learning_rate": 0.00027812455516014234,
|
| 1743 |
+
"loss": 0.3351,
|
| 1744 |
+
"step": 6100
|
| 1745 |
+
},
|
| 1746 |
+
{
|
| 1747 |
+
"epoch": 3.1572164948453607,
|
| 1748 |
+
"grad_norm": 0.4432651996612549,
|
| 1749 |
+
"learning_rate": 0.00027799110320284693,
|
| 1750 |
+
"loss": 0.3494,
|
| 1751 |
+
"step": 6125
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 3.170103092783505,
|
| 1755 |
+
"grad_norm": 0.5367820858955383,
|
| 1756 |
+
"learning_rate": 0.0002778576512455516,
|
| 1757 |
+
"loss": 0.3311,
|
| 1758 |
+
"step": 6150
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 3.1829896907216493,
|
| 1762 |
+
"grad_norm": 0.3430980443954468,
|
| 1763 |
+
"learning_rate": 0.00027772419928825617,
|
| 1764 |
+
"loss": 0.3357,
|
| 1765 |
+
"step": 6175
|
| 1766 |
+
},
|
| 1767 |
+
{
|
| 1768 |
+
"epoch": 3.195876288659794,
|
| 1769 |
+
"grad_norm": 1.0889408588409424,
|
| 1770 |
+
"learning_rate": 0.0002775907473309608,
|
| 1771 |
+
"loss": 0.3056,
|
| 1772 |
+
"step": 6200
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 3.2087628865979383,
|
| 1776 |
+
"grad_norm": 0.3883308172225952,
|
| 1777 |
+
"learning_rate": 0.00027745729537366546,
|
| 1778 |
+
"loss": 0.3251,
|
| 1779 |
+
"step": 6225
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 3.2216494845360826,
|
| 1783 |
+
"grad_norm": 0.41912856698036194,
|
| 1784 |
+
"learning_rate": 0.00027732384341637006,
|
| 1785 |
+
"loss": 0.3292,
|
| 1786 |
+
"step": 6250
|
| 1787 |
+
},
|
| 1788 |
+
{
|
| 1789 |
+
"epoch": 3.234536082474227,
|
| 1790 |
+
"grad_norm": 0.45084699988365173,
|
| 1791 |
+
"learning_rate": 0.0002771903914590747,
|
| 1792 |
+
"loss": 0.3516,
|
| 1793 |
+
"step": 6275
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 3.247422680412371,
|
| 1797 |
+
"grad_norm": 0.4294661581516266,
|
| 1798 |
+
"learning_rate": 0.00027705693950177935,
|
| 1799 |
+
"loss": 0.3297,
|
| 1800 |
+
"step": 6300
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 3.2603092783505154,
|
| 1804 |
+
"grad_norm": 0.25462472438812256,
|
| 1805 |
+
"learning_rate": 0.00027692348754448394,
|
| 1806 |
+
"loss": 0.3513,
|
| 1807 |
+
"step": 6325
|
| 1808 |
+
},
|
| 1809 |
+
{
|
| 1810 |
+
"epoch": 3.2731958762886597,
|
| 1811 |
+
"grad_norm": 0.2990482747554779,
|
| 1812 |
+
"learning_rate": 0.0002767900355871886,
|
| 1813 |
+
"loss": 0.3472,
|
| 1814 |
+
"step": 6350
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 3.286082474226804,
|
| 1818 |
+
"grad_norm": 0.5474823713302612,
|
| 1819 |
+
"learning_rate": 0.0002766565836298932,
|
| 1820 |
+
"loss": 0.3404,
|
| 1821 |
+
"step": 6375
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 3.2989690721649483,
|
| 1825 |
+
"grad_norm": 0.31416311860084534,
|
| 1826 |
+
"learning_rate": 0.00027652313167259783,
|
| 1827 |
+
"loss": 0.3052,
|
| 1828 |
+
"step": 6400
|
| 1829 |
+
},
|
| 1830 |
+
{
|
| 1831 |
+
"epoch": 3.3118556701030926,
|
| 1832 |
+
"grad_norm": 0.6990143656730652,
|
| 1833 |
+
"learning_rate": 0.0002763896797153025,
|
| 1834 |
+
"loss": 0.3661,
|
| 1835 |
+
"step": 6425
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 3.3247422680412373,
|
| 1839 |
+
"grad_norm": 0.4336131811141968,
|
| 1840 |
+
"learning_rate": 0.00027625622775800707,
|
| 1841 |
+
"loss": 0.2757,
|
| 1842 |
+
"step": 6450
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 3.3376288659793816,
|
| 1846 |
+
"grad_norm": 0.35333672165870667,
|
| 1847 |
+
"learning_rate": 0.0002761227758007117,
|
| 1848 |
+
"loss": 0.3275,
|
| 1849 |
+
"step": 6475
|
| 1850 |
+
},
|
| 1851 |
+
{
|
| 1852 |
+
"epoch": 3.350515463917526,
|
| 1853 |
+
"grad_norm": 0.7743633985519409,
|
| 1854 |
+
"learning_rate": 0.0002759893238434163,
|
| 1855 |
+
"loss": 0.2855,
|
| 1856 |
+
"step": 6500
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 3.36340206185567,
|
| 1860 |
+
"grad_norm": 0.5316669344902039,
|
| 1861 |
+
"learning_rate": 0.00027585587188612095,
|
| 1862 |
+
"loss": 0.3742,
|
| 1863 |
+
"step": 6525
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 3.3762886597938144,
|
| 1867 |
+
"grad_norm": 0.3642534017562866,
|
| 1868 |
+
"learning_rate": 0.0002757277580071174,
|
| 1869 |
+
"loss": 0.3676,
|
| 1870 |
+
"step": 6550
|
| 1871 |
+
},
|
| 1872 |
+
{
|
| 1873 |
+
"epoch": 3.3891752577319587,
|
| 1874 |
+
"grad_norm": 0.6164928078651428,
|
| 1875 |
+
"learning_rate": 0.00027559430604982203,
|
| 1876 |
+
"loss": 0.3658,
|
| 1877 |
+
"step": 6575
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 3.402061855670103,
|
| 1881 |
+
"grad_norm": 0.3114078938961029,
|
| 1882 |
+
"learning_rate": 0.0002754608540925266,
|
| 1883 |
+
"loss": 0.2681,
|
| 1884 |
+
"step": 6600
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 3.4149484536082473,
|
| 1888 |
+
"grad_norm": 0.8610777258872986,
|
| 1889 |
+
"learning_rate": 0.00027532740213523127,
|
| 1890 |
+
"loss": 0.3607,
|
| 1891 |
+
"step": 6625
|
| 1892 |
+
},
|
| 1893 |
+
{
|
| 1894 |
+
"epoch": 3.4278350515463916,
|
| 1895 |
+
"grad_norm": 0.49727797508239746,
|
| 1896 |
+
"learning_rate": 0.0002751939501779359,
|
| 1897 |
+
"loss": 0.3306,
|
| 1898 |
+
"step": 6650
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 3.4407216494845363,
|
| 1902 |
+
"grad_norm": 0.3058234751224518,
|
| 1903 |
+
"learning_rate": 0.0002750604982206405,
|
| 1904 |
+
"loss": 0.3313,
|
| 1905 |
+
"step": 6675
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 3.4536082474226806,
|
| 1909 |
+
"grad_norm": 0.5317339301109314,
|
| 1910 |
+
"learning_rate": 0.00027492704626334516,
|
| 1911 |
+
"loss": 0.3316,
|
| 1912 |
+
"step": 6700
|
| 1913 |
+
},
|
| 1914 |
+
{
|
| 1915 |
+
"epoch": 3.466494845360825,
|
| 1916 |
+
"grad_norm": 0.4639209806919098,
|
| 1917 |
+
"learning_rate": 0.0002747935943060498,
|
| 1918 |
+
"loss": 0.3773,
|
| 1919 |
+
"step": 6725
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 3.479381443298969,
|
| 1923 |
+
"grad_norm": 0.28237494826316833,
|
| 1924 |
+
"learning_rate": 0.0002746601423487544,
|
| 1925 |
+
"loss": 0.3275,
|
| 1926 |
+
"step": 6750
|
| 1927 |
+
},
|
| 1928 |
+
{
|
| 1929 |
+
"epoch": 3.4922680412371134,
|
| 1930 |
+
"grad_norm": 0.44046372175216675,
|
| 1931 |
+
"learning_rate": 0.00027452669039145904,
|
| 1932 |
+
"loss": 0.3777,
|
| 1933 |
+
"step": 6775
|
| 1934 |
+
},
|
| 1935 |
+
{
|
| 1936 |
+
"epoch": 3.5051546391752577,
|
| 1937 |
+
"grad_norm": 0.40542760491371155,
|
| 1938 |
+
"learning_rate": 0.0002743932384341637,
|
| 1939 |
+
"loss": 0.3197,
|
| 1940 |
+
"step": 6800
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 3.518041237113402,
|
| 1944 |
+
"grad_norm": 0.587062418460846,
|
| 1945 |
+
"learning_rate": 0.0002742597864768683,
|
| 1946 |
+
"loss": 0.3441,
|
| 1947 |
+
"step": 6825
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 3.5309278350515463,
|
| 1951 |
+
"grad_norm": 0.34003278613090515,
|
| 1952 |
+
"learning_rate": 0.00027412633451957293,
|
| 1953 |
+
"loss": 0.3189,
|
| 1954 |
+
"step": 6850
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 3.5438144329896906,
|
| 1958 |
+
"grad_norm": 0.40320533514022827,
|
| 1959 |
+
"learning_rate": 0.0002739928825622776,
|
| 1960 |
+
"loss": 0.3115,
|
| 1961 |
+
"step": 6875
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 3.556701030927835,
|
| 1965 |
+
"grad_norm": 0.34437236189842224,
|
| 1966 |
+
"learning_rate": 0.00027385943060498217,
|
| 1967 |
+
"loss": 0.3344,
|
| 1968 |
+
"step": 6900
|
| 1969 |
+
},
|
| 1970 |
+
{
|
| 1971 |
+
"epoch": 3.569587628865979,
|
| 1972 |
+
"grad_norm": 0.42826181650161743,
|
| 1973 |
+
"learning_rate": 0.0002737259786476868,
|
| 1974 |
+
"loss": 0.3716,
|
| 1975 |
+
"step": 6925
|
| 1976 |
+
},
|
| 1977 |
+
{
|
| 1978 |
+
"epoch": 3.582474226804124,
|
| 1979 |
+
"grad_norm": 0.4200308620929718,
|
| 1980 |
+
"learning_rate": 0.00027359252669039146,
|
| 1981 |
+
"loss": 0.3511,
|
| 1982 |
+
"step": 6950
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 3.595360824742268,
|
| 1986 |
+
"grad_norm": 0.3897481858730316,
|
| 1987 |
+
"learning_rate": 0.00027345907473309605,
|
| 1988 |
+
"loss": 0.3651,
|
| 1989 |
+
"step": 6975
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 3.6082474226804124,
|
| 1993 |
+
"grad_norm": 0.3066980838775635,
|
| 1994 |
+
"learning_rate": 0.0002733256227758007,
|
| 1995 |
+
"loss": 0.314,
|
| 1996 |
+
"step": 7000
|
| 1997 |
+
},
|
| 1998 |
+
{
|
| 1999 |
+
"epoch": 3.6211340206185567,
|
| 2000 |
+
"grad_norm": 0.39002224802970886,
|
| 2001 |
+
"learning_rate": 0.0002731921708185053,
|
| 2002 |
+
"loss": 0.411,
|
| 2003 |
+
"step": 7025
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 3.634020618556701,
|
| 2007 |
+
"grad_norm": 0.8328560590744019,
|
| 2008 |
+
"learning_rate": 0.00027305871886120994,
|
| 2009 |
+
"loss": 0.3154,
|
| 2010 |
+
"step": 7050
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 3.6469072164948453,
|
| 2014 |
+
"grad_norm": 0.43831050395965576,
|
| 2015 |
+
"learning_rate": 0.0002729252669039146,
|
| 2016 |
+
"loss": 0.396,
|
| 2017 |
+
"step": 7075
|
| 2018 |
+
},
|
| 2019 |
+
{
|
| 2020 |
+
"epoch": 3.6597938144329896,
|
| 2021 |
+
"grad_norm": 0.3992260694503784,
|
| 2022 |
+
"learning_rate": 0.0002727918149466192,
|
| 2023 |
+
"loss": 0.3142,
|
| 2024 |
+
"step": 7100
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 3.6726804123711343,
|
| 2028 |
+
"grad_norm": 0.2941119074821472,
|
| 2029 |
+
"learning_rate": 0.0002726583629893238,
|
| 2030 |
+
"loss": 0.3527,
|
| 2031 |
+
"step": 7125
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 3.6855670103092786,
|
| 2035 |
+
"grad_norm": 0.4881301522254944,
|
| 2036 |
+
"learning_rate": 0.00027252491103202847,
|
| 2037 |
+
"loss": 0.3171,
|
| 2038 |
+
"step": 7150
|
| 2039 |
+
},
|
| 2040 |
+
{
|
| 2041 |
+
"epoch": 3.698453608247423,
|
| 2042 |
+
"grad_norm": 0.3898142874240875,
|
| 2043 |
+
"learning_rate": 0.00027239145907473306,
|
| 2044 |
+
"loss": 0.3817,
|
| 2045 |
+
"step": 7175
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 3.711340206185567,
|
| 2049 |
+
"grad_norm": 0.31580379605293274,
|
| 2050 |
+
"learning_rate": 0.0002722580071174377,
|
| 2051 |
+
"loss": 0.2595,
|
| 2052 |
+
"step": 7200
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 3.7242268041237114,
|
| 2056 |
+
"grad_norm": 0.3299170136451721,
|
| 2057 |
+
"learning_rate": 0.00027212455516014236,
|
| 2058 |
+
"loss": 0.3416,
|
| 2059 |
+
"step": 7225
|
| 2060 |
+
},
|
| 2061 |
+
{
|
| 2062 |
+
"epoch": 3.7371134020618557,
|
| 2063 |
+
"grad_norm": 0.6964473724365234,
|
| 2064 |
+
"learning_rate": 0.00027199110320284695,
|
| 2065 |
+
"loss": 0.2942,
|
| 2066 |
+
"step": 7250
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 3.75,
|
| 2070 |
+
"grad_norm": 0.3565264046192169,
|
| 2071 |
+
"learning_rate": 0.0002718576512455516,
|
| 2072 |
+
"loss": 0.3283,
|
| 2073 |
+
"step": 7275
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 3.7628865979381443,
|
| 2077 |
+
"grad_norm": 0.28134214878082275,
|
| 2078 |
+
"learning_rate": 0.00027172419928825624,
|
| 2079 |
+
"loss": 0.3203,
|
| 2080 |
+
"step": 7300
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 3.7757731958762886,
|
| 2084 |
+
"grad_norm": 0.4196174442768097,
|
| 2085 |
+
"learning_rate": 0.00027159074733096083,
|
| 2086 |
+
"loss": 0.3396,
|
| 2087 |
+
"step": 7325
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 3.788659793814433,
|
| 2091 |
+
"grad_norm": 0.4770311415195465,
|
| 2092 |
+
"learning_rate": 0.0002714572953736655,
|
| 2093 |
+
"loss": 0.3144,
|
| 2094 |
+
"step": 7350
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 3.801546391752577,
|
| 2098 |
+
"grad_norm": 0.29138994216918945,
|
| 2099 |
+
"learning_rate": 0.0002713238434163701,
|
| 2100 |
+
"loss": 0.3403,
|
| 2101 |
+
"step": 7375
|
| 2102 |
+
},
|
| 2103 |
+
{
|
| 2104 |
+
"epoch": 3.8144329896907214,
|
| 2105 |
+
"grad_norm": 0.47035351395606995,
|
| 2106 |
+
"learning_rate": 0.0002711903914590747,
|
| 2107 |
+
"loss": 0.2623,
|
| 2108 |
+
"step": 7400
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 3.8273195876288657,
|
| 2112 |
+
"grad_norm": 0.32780078053474426,
|
| 2113 |
+
"learning_rate": 0.00027105693950177937,
|
| 2114 |
+
"loss": 0.3863,
|
| 2115 |
+
"step": 7425
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 3.8402061855670104,
|
| 2119 |
+
"grad_norm": 0.4419979751110077,
|
| 2120 |
+
"learning_rate": 0.00027092348754448396,
|
| 2121 |
+
"loss": 0.3207,
|
| 2122 |
+
"step": 7450
|
| 2123 |
+
},
|
| 2124 |
+
{
|
| 2125 |
+
"epoch": 3.8530927835051547,
|
| 2126 |
+
"grad_norm": 0.28167861700057983,
|
| 2127 |
+
"learning_rate": 0.0002707900355871886,
|
| 2128 |
+
"loss": 0.3308,
|
| 2129 |
+
"step": 7475
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 3.865979381443299,
|
| 2133 |
+
"grad_norm": 0.3411625921726227,
|
| 2134 |
+
"learning_rate": 0.00027065658362989325,
|
| 2135 |
+
"loss": 0.3387,
|
| 2136 |
+
"step": 7500
|
| 2137 |
+
},
|
| 2138 |
+
{
|
| 2139 |
+
"epoch": 3.8788659793814433,
|
| 2140 |
+
"grad_norm": 0.31229549646377563,
|
| 2141 |
+
"learning_rate": 0.00027052313167259785,
|
| 2142 |
+
"loss": 0.408,
|
| 2143 |
+
"step": 7525
|
| 2144 |
+
},
|
| 2145 |
+
{
|
| 2146 |
+
"epoch": 3.8917525773195876,
|
| 2147 |
+
"grad_norm": 0.3603403866291046,
|
| 2148 |
+
"learning_rate": 0.0002703896797153025,
|
| 2149 |
+
"loss": 0.3259,
|
| 2150 |
+
"step": 7550
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 3.904639175257732,
|
| 2154 |
+
"grad_norm": 0.2832438051700592,
|
| 2155 |
+
"learning_rate": 0.0002702562277580071,
|
| 2156 |
+
"loss": 0.3423,
|
| 2157 |
+
"step": 7575
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 3.917525773195876,
|
| 2161 |
+
"grad_norm": 0.3435526192188263,
|
| 2162 |
+
"learning_rate": 0.00027012277580071173,
|
| 2163 |
+
"loss": 0.2835,
|
| 2164 |
+
"step": 7600
|
| 2165 |
+
},
|
| 2166 |
+
{
|
| 2167 |
+
"epoch": 3.930412371134021,
|
| 2168 |
+
"grad_norm": 0.3645428419113159,
|
| 2169 |
+
"learning_rate": 0.0002699893238434164,
|
| 2170 |
+
"loss": 0.3483,
|
| 2171 |
+
"step": 7625
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"epoch": 3.943298969072165,
|
| 2175 |
+
"grad_norm": 0.7062532901763916,
|
| 2176 |
+
"learning_rate": 0.00026985587188612097,
|
| 2177 |
+
"loss": 0.2586,
|
| 2178 |
+
"step": 7650
|
| 2179 |
+
},
|
| 2180 |
+
{
|
| 2181 |
+
"epoch": 3.9561855670103094,
|
| 2182 |
+
"grad_norm": 0.4606476426124573,
|
| 2183 |
+
"learning_rate": 0.0002697224199288256,
|
| 2184 |
+
"loss": 0.3939,
|
| 2185 |
+
"step": 7675
|
| 2186 |
+
},
|
| 2187 |
+
{
|
| 2188 |
+
"epoch": 3.9690721649484537,
|
| 2189 |
+
"grad_norm": 0.4727219045162201,
|
| 2190 |
+
"learning_rate": 0.0002695889679715302,
|
| 2191 |
+
"loss": 0.3151,
|
| 2192 |
+
"step": 7700
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 3.981958762886598,
|
| 2196 |
+
"grad_norm": 0.2774180471897125,
|
| 2197 |
+
"learning_rate": 0.00026945551601423486,
|
| 2198 |
+
"loss": 0.3519,
|
| 2199 |
+
"step": 7725
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 3.9948453608247423,
|
| 2203 |
+
"grad_norm": 0.4793704152107239,
|
| 2204 |
+
"learning_rate": 0.0002693220640569395,
|
| 2205 |
+
"loss": 0.3023,
|
| 2206 |
+
"step": 7750
|
| 2207 |
+
},
|
| 2208 |
+
{
|
| 2209 |
+
"epoch": 4.007731958762887,
|
| 2210 |
+
"grad_norm": 0.3699122965335846,
|
| 2211 |
+
"learning_rate": 0.0002691886120996441,
|
| 2212 |
+
"loss": 0.3664,
|
| 2213 |
+
"step": 7775
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"epoch": 4.020618556701031,
|
| 2217 |
+
"grad_norm": 0.6096347570419312,
|
| 2218 |
+
"learning_rate": 0.00026905516014234874,
|
| 2219 |
+
"loss": 0.2579,
|
| 2220 |
+
"step": 7800
|
| 2221 |
+
},
|
| 2222 |
+
{
|
| 2223 |
+
"epoch": 4.033505154639175,
|
| 2224 |
+
"grad_norm": 0.2851110100746155,
|
| 2225 |
+
"learning_rate": 0.00026892170818505333,
|
| 2226 |
+
"loss": 0.3347,
|
| 2227 |
+
"step": 7825
|
| 2228 |
+
},
|
| 2229 |
+
{
|
| 2230 |
+
"epoch": 4.046391752577319,
|
| 2231 |
+
"grad_norm": 0.5487604141235352,
|
| 2232 |
+
"learning_rate": 0.000268788256227758,
|
| 2233 |
+
"loss": 0.2676,
|
| 2234 |
+
"step": 7850
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 4.059278350515464,
|
| 2238 |
+
"grad_norm": 0.29913634061813354,
|
| 2239 |
+
"learning_rate": 0.0002686548042704626,
|
| 2240 |
+
"loss": 0.3061,
|
| 2241 |
+
"step": 7875
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 4.072164948453608,
|
| 2245 |
+
"grad_norm": 0.3672349750995636,
|
| 2246 |
+
"learning_rate": 0.0002685213523131672,
|
| 2247 |
+
"loss": 0.2632,
|
| 2248 |
+
"step": 7900
|
| 2249 |
+
},
|
| 2250 |
+
{
|
| 2251 |
+
"epoch": 4.085051546391752,
|
| 2252 |
+
"grad_norm": 0.5969862341880798,
|
| 2253 |
+
"learning_rate": 0.00026838790035587187,
|
| 2254 |
+
"loss": 0.2868,
|
| 2255 |
+
"step": 7925
|
| 2256 |
+
},
|
| 2257 |
+
{
|
| 2258 |
+
"epoch": 4.097938144329897,
|
| 2259 |
+
"grad_norm": 0.4409433901309967,
|
| 2260 |
+
"learning_rate": 0.0002682544483985765,
|
| 2261 |
+
"loss": 0.2907,
|
| 2262 |
+
"step": 7950
|
| 2263 |
+
},
|
| 2264 |
+
{
|
| 2265 |
+
"epoch": 4.110824742268041,
|
| 2266 |
+
"grad_norm": 0.4578838050365448,
|
| 2267 |
+
"learning_rate": 0.0002681209964412811,
|
| 2268 |
+
"loss": 0.3377,
|
| 2269 |
+
"step": 7975
|
| 2270 |
+
},
|
| 2271 |
+
{
|
| 2272 |
+
"epoch": 4.123711340206185,
|
| 2273 |
+
"grad_norm": 0.302491694688797,
|
| 2274 |
+
"learning_rate": 0.00026798754448398575,
|
| 2275 |
+
"loss": 0.2857,
|
| 2276 |
+
"step": 8000
|
| 2277 |
+
},
|
| 2278 |
+
{
|
| 2279 |
+
"epoch": 4.123711340206185,
|
| 2280 |
+
"eval_loss": 0.4647158980369568,
|
| 2281 |
+
"eval_runtime": 189.5598,
|
| 2282 |
+
"eval_samples_per_second": 6.462,
|
| 2283 |
+
"eval_steps_per_second": 0.812,
|
| 2284 |
+
"eval_wer": 1.0008163265306123,
|
| 2285 |
+
"step": 8000
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 4.13659793814433,
|
| 2289 |
+
"grad_norm": 0.5263584852218628,
|
| 2290 |
+
"learning_rate": 0.00026785409252669034,
|
| 2291 |
+
"loss": 0.3246,
|
| 2292 |
+
"step": 8025
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 4.149484536082475,
|
| 2296 |
+
"grad_norm": 0.48449578881263733,
|
| 2297 |
+
"learning_rate": 0.000267720640569395,
|
| 2298 |
+
"loss": 0.3163,
|
| 2299 |
+
"step": 8050
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 4.162371134020619,
|
| 2303 |
+
"grad_norm": 0.3020855784416199,
|
| 2304 |
+
"learning_rate": 0.00026758718861209964,
|
| 2305 |
+
"loss": 0.3248,
|
| 2306 |
+
"step": 8075
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 4.175257731958763,
|
| 2310 |
+
"grad_norm": 0.3554665148258209,
|
| 2311 |
+
"learning_rate": 0.00026745373665480423,
|
| 2312 |
+
"loss": 0.2863,
|
| 2313 |
+
"step": 8100
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 4.188144329896907,
|
| 2317 |
+
"grad_norm": 0.4055439531803131,
|
| 2318 |
+
"learning_rate": 0.0002673202846975089,
|
| 2319 |
+
"loss": 0.3289,
|
| 2320 |
+
"step": 8125
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 4.201030927835052,
|
| 2324 |
+
"grad_norm": 0.6498619914054871,
|
| 2325 |
+
"learning_rate": 0.00026718683274021347,
|
| 2326 |
+
"loss": 0.289,
|
| 2327 |
+
"step": 8150
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 4.213917525773196,
|
| 2331 |
+
"grad_norm": 0.23331980407238007,
|
| 2332 |
+
"learning_rate": 0.0002670533807829181,
|
| 2333 |
+
"loss": 0.2744,
|
| 2334 |
+
"step": 8175
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 4.22680412371134,
|
| 2338 |
+
"grad_norm": 0.5142727494239807,
|
| 2339 |
+
"learning_rate": 0.00026691992882562276,
|
| 2340 |
+
"loss": 0.3139,
|
| 2341 |
+
"step": 8200
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 4.239690721649485,
|
| 2345 |
+
"grad_norm": 0.4468501806259155,
|
| 2346 |
+
"learning_rate": 0.00026678647686832735,
|
| 2347 |
+
"loss": 0.3392,
|
| 2348 |
+
"step": 8225
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 4.252577319587629,
|
| 2352 |
+
"grad_norm": 0.36652007699012756,
|
| 2353 |
+
"learning_rate": 0.000266653024911032,
|
| 2354 |
+
"loss": 0.2499,
|
| 2355 |
+
"step": 8250
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 4.265463917525773,
|
| 2359 |
+
"grad_norm": 0.3004089891910553,
|
| 2360 |
+
"learning_rate": 0.0002665195729537366,
|
| 2361 |
+
"loss": 0.31,
|
| 2362 |
+
"step": 8275
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 4.278350515463917,
|
| 2366 |
+
"grad_norm": 0.4099670946598053,
|
| 2367 |
+
"learning_rate": 0.00026638612099644124,
|
| 2368 |
+
"loss": 0.2798,
|
| 2369 |
+
"step": 8300
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 4.291237113402062,
|
| 2373 |
+
"grad_norm": 0.26817041635513306,
|
| 2374 |
+
"learning_rate": 0.0002662526690391459,
|
| 2375 |
+
"loss": 0.3055,
|
| 2376 |
+
"step": 8325
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 4.304123711340206,
|
| 2380 |
+
"grad_norm": 0.4156699478626251,
|
| 2381 |
+
"learning_rate": 0.0002661192170818505,
|
| 2382 |
+
"loss": 0.2942,
|
| 2383 |
+
"step": 8350
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 4.31701030927835,
|
| 2387 |
+
"grad_norm": 2.611400842666626,
|
| 2388 |
+
"learning_rate": 0.0002659857651245551,
|
| 2389 |
+
"loss": 0.3659,
|
| 2390 |
+
"step": 8375
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 4.329896907216495,
|
| 2394 |
+
"grad_norm": 0.46959781646728516,
|
| 2395 |
+
"learning_rate": 0.00026585231316725977,
|
| 2396 |
+
"loss": 0.2982,
|
| 2397 |
+
"step": 8400
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 4.342783505154639,
|
| 2401 |
+
"grad_norm": 0.2867629826068878,
|
| 2402 |
+
"learning_rate": 0.00026571886120996436,
|
| 2403 |
+
"loss": 0.3511,
|
| 2404 |
+
"step": 8425
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 4.355670103092783,
|
| 2408 |
+
"grad_norm": 0.38449224829673767,
|
| 2409 |
+
"learning_rate": 0.000265585409252669,
|
| 2410 |
+
"loss": 0.2826,
|
| 2411 |
+
"step": 8450
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 4.368556701030927,
|
| 2415 |
+
"grad_norm": 0.4524473249912262,
|
| 2416 |
+
"learning_rate": 0.0002654519572953736,
|
| 2417 |
+
"loss": 0.313,
|
| 2418 |
+
"step": 8475
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 4.381443298969073,
|
| 2422 |
+
"grad_norm": 0.3989889919757843,
|
| 2423 |
+
"learning_rate": 0.00026531850533807825,
|
| 2424 |
+
"loss": 0.2825,
|
| 2425 |
+
"step": 8500
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 4.394329896907217,
|
| 2429 |
+
"grad_norm": 0.3007287085056305,
|
| 2430 |
+
"learning_rate": 0.0002651850533807829,
|
| 2431 |
+
"loss": 0.3369,
|
| 2432 |
+
"step": 8525
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 4.407216494845361,
|
| 2436 |
+
"grad_norm": 0.7760800123214722,
|
| 2437 |
+
"learning_rate": 0.0002650516014234875,
|
| 2438 |
+
"loss": 0.2835,
|
| 2439 |
+
"step": 8550
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 4.420103092783505,
|
| 2443 |
+
"grad_norm": 0.27871614694595337,
|
| 2444 |
+
"learning_rate": 0.00026491814946619214,
|
| 2445 |
+
"loss": 0.3172,
|
| 2446 |
+
"step": 8575
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 4.43298969072165,
|
| 2450 |
+
"grad_norm": 0.48607179522514343,
|
| 2451 |
+
"learning_rate": 0.0002647846975088968,
|
| 2452 |
+
"loss": 0.303,
|
| 2453 |
+
"step": 8600
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 4.445876288659794,
|
| 2457 |
+
"grad_norm": 0.3540396988391876,
|
| 2458 |
+
"learning_rate": 0.0002646512455516014,
|
| 2459 |
+
"loss": 0.355,
|
| 2460 |
+
"step": 8625
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 4.458762886597938,
|
| 2464 |
+
"grad_norm": 1.2004367113113403,
|
| 2465 |
+
"learning_rate": 0.000264517793594306,
|
| 2466 |
+
"loss": 0.3197,
|
| 2467 |
+
"step": 8650
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 4.471649484536083,
|
| 2471 |
+
"grad_norm": 0.2405807226896286,
|
| 2472 |
+
"learning_rate": 0.00026438434163701067,
|
| 2473 |
+
"loss": 0.3179,
|
| 2474 |
+
"step": 8675
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 4.484536082474227,
|
| 2478 |
+
"grad_norm": 0.36553826928138733,
|
| 2479 |
+
"learning_rate": 0.00026425088967971526,
|
| 2480 |
+
"loss": 0.2793,
|
| 2481 |
+
"step": 8700
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 4.497422680412371,
|
| 2485 |
+
"grad_norm": 0.26435115933418274,
|
| 2486 |
+
"learning_rate": 0.0002641174377224199,
|
| 2487 |
+
"loss": 0.2849,
|
| 2488 |
+
"step": 8725
|
| 2489 |
+
},
|
| 2490 |
+
{
|
| 2491 |
+
"epoch": 4.510309278350515,
|
| 2492 |
+
"grad_norm": 0.6196132302284241,
|
| 2493 |
+
"learning_rate": 0.00026398398576512455,
|
| 2494 |
+
"loss": 0.3267,
|
| 2495 |
+
"step": 8750
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 4.52319587628866,
|
| 2499 |
+
"grad_norm": 0.2987557649612427,
|
| 2500 |
+
"learning_rate": 0.00026385053380782915,
|
| 2501 |
+
"loss": 0.3152,
|
| 2502 |
+
"step": 8775
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 4.536082474226804,
|
| 2506 |
+
"grad_norm": 0.4445992112159729,
|
| 2507 |
+
"learning_rate": 0.0002637170818505338,
|
| 2508 |
+
"loss": 0.3,
|
| 2509 |
+
"step": 8800
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 4.548969072164948,
|
| 2513 |
+
"grad_norm": 0.4059930741786957,
|
| 2514 |
+
"learning_rate": 0.00026358362989323844,
|
| 2515 |
+
"loss": 0.3395,
|
| 2516 |
+
"step": 8825
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 4.561855670103093,
|
| 2520 |
+
"grad_norm": 0.522637665271759,
|
| 2521 |
+
"learning_rate": 0.00026345017793594303,
|
| 2522 |
+
"loss": 0.2783,
|
| 2523 |
+
"step": 8850
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 4.574742268041237,
|
| 2527 |
+
"grad_norm": 0.1963280737400055,
|
| 2528 |
+
"learning_rate": 0.0002633167259786477,
|
| 2529 |
+
"loss": 0.2979,
|
| 2530 |
+
"step": 8875
|
| 2531 |
+
},
|
| 2532 |
+
{
|
| 2533 |
+
"epoch": 4.587628865979381,
|
| 2534 |
+
"grad_norm": 0.7622120976448059,
|
| 2535 |
+
"learning_rate": 0.00026318327402135227,
|
| 2536 |
+
"loss": 0.3204,
|
| 2537 |
+
"step": 8900
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 4.600515463917525,
|
| 2541 |
+
"grad_norm": 0.2749217450618744,
|
| 2542 |
+
"learning_rate": 0.0002630498220640569,
|
| 2543 |
+
"loss": 0.3365,
|
| 2544 |
+
"step": 8925
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 4.61340206185567,
|
| 2548 |
+
"grad_norm": 0.5664824843406677,
|
| 2549 |
+
"learning_rate": 0.00026291637010676156,
|
| 2550 |
+
"loss": 0.2772,
|
| 2551 |
+
"step": 8950
|
| 2552 |
+
},
|
| 2553 |
+
{
|
| 2554 |
+
"epoch": 4.626288659793815,
|
| 2555 |
+
"grad_norm": 0.3661792278289795,
|
| 2556 |
+
"learning_rate": 0.00026278291814946616,
|
| 2557 |
+
"loss": 0.3314,
|
| 2558 |
+
"step": 8975
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 4.639175257731958,
|
| 2562 |
+
"grad_norm": 0.37070122361183167,
|
| 2563 |
+
"learning_rate": 0.0002626494661921708,
|
| 2564 |
+
"loss": 0.2991,
|
| 2565 |
+
"step": 9000
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 4.652061855670103,
|
| 2569 |
+
"grad_norm": 0.4963271915912628,
|
| 2570 |
+
"learning_rate": 0.00026251601423487545,
|
| 2571 |
+
"loss": 0.3312,
|
| 2572 |
+
"step": 9025
|
| 2573 |
+
},
|
| 2574 |
+
{
|
| 2575 |
+
"epoch": 4.664948453608248,
|
| 2576 |
+
"grad_norm": 0.5193173885345459,
|
| 2577 |
+
"learning_rate": 0.00026238256227758004,
|
| 2578 |
+
"loss": 0.2914,
|
| 2579 |
+
"step": 9050
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 4.677835051546392,
|
| 2583 |
+
"grad_norm": 0.3158729076385498,
|
| 2584 |
+
"learning_rate": 0.0002622491103202847,
|
| 2585 |
+
"loss": 0.3317,
|
| 2586 |
+
"step": 9075
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"epoch": 4.690721649484536,
|
| 2590 |
+
"grad_norm": 0.37290090322494507,
|
| 2591 |
+
"learning_rate": 0.00026211565836298934,
|
| 2592 |
+
"loss": 0.2612,
|
| 2593 |
+
"step": 9100
|
| 2594 |
+
},
|
| 2595 |
+
{
|
| 2596 |
+
"epoch": 4.703608247422681,
|
| 2597 |
+
"grad_norm": 0.334089070558548,
|
| 2598 |
+
"learning_rate": 0.00026198220640569393,
|
| 2599 |
+
"loss": 0.3235,
|
| 2600 |
+
"step": 9125
|
| 2601 |
+
},
|
| 2602 |
+
{
|
| 2603 |
+
"epoch": 4.716494845360825,
|
| 2604 |
+
"grad_norm": 0.3259222209453583,
|
| 2605 |
+
"learning_rate": 0.0002618487544483986,
|
| 2606 |
+
"loss": 0.2952,
|
| 2607 |
+
"step": 9150
|
| 2608 |
+
},
|
| 2609 |
+
{
|
| 2610 |
+
"epoch": 4.729381443298969,
|
| 2611 |
+
"grad_norm": 0.2776556611061096,
|
| 2612 |
+
"learning_rate": 0.0002617153024911032,
|
| 2613 |
+
"loss": 0.3405,
|
| 2614 |
+
"step": 9175
|
| 2615 |
+
},
|
| 2616 |
+
{
|
| 2617 |
+
"epoch": 4.742268041237113,
|
| 2618 |
+
"grad_norm": 0.33884450793266296,
|
| 2619 |
+
"learning_rate": 0.0002615818505338078,
|
| 2620 |
+
"loss": 0.2813,
|
| 2621 |
+
"step": 9200
|
| 2622 |
+
},
|
| 2623 |
+
{
|
| 2624 |
+
"epoch": 4.755154639175258,
|
| 2625 |
+
"grad_norm": 0.48323071002960205,
|
| 2626 |
+
"learning_rate": 0.00026144839857651246,
|
| 2627 |
+
"loss": 0.3044,
|
| 2628 |
+
"step": 9225
|
| 2629 |
+
},
|
| 2630 |
+
{
|
| 2631 |
+
"epoch": 4.768041237113402,
|
| 2632 |
+
"grad_norm": 0.32996150851249695,
|
| 2633 |
+
"learning_rate": 0.00026131494661921705,
|
| 2634 |
+
"loss": 0.2817,
|
| 2635 |
+
"step": 9250
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 4.780927835051546,
|
| 2639 |
+
"grad_norm": 0.3732437491416931,
|
| 2640 |
+
"learning_rate": 0.0002611814946619217,
|
| 2641 |
+
"loss": 0.3165,
|
| 2642 |
+
"step": 9275
|
| 2643 |
+
},
|
| 2644 |
+
{
|
| 2645 |
+
"epoch": 4.793814432989691,
|
| 2646 |
+
"grad_norm": 0.5148000717163086,
|
| 2647 |
+
"learning_rate": 0.00026104804270462635,
|
| 2648 |
+
"loss": 0.2636,
|
| 2649 |
+
"step": 9300
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 4.806701030927835,
|
| 2653 |
+
"grad_norm": 0.3256881833076477,
|
| 2654 |
+
"learning_rate": 0.00026091459074733094,
|
| 2655 |
+
"loss": 0.3677,
|
| 2656 |
+
"step": 9325
|
| 2657 |
+
},
|
| 2658 |
+
{
|
| 2659 |
+
"epoch": 4.819587628865979,
|
| 2660 |
+
"grad_norm": 0.5248320698738098,
|
| 2661 |
+
"learning_rate": 0.0002607811387900356,
|
| 2662 |
+
"loss": 0.2959,
|
| 2663 |
+
"step": 9350
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 4.832474226804123,
|
| 2667 |
+
"grad_norm": 0.25363728404045105,
|
| 2668 |
+
"learning_rate": 0.0002606476868327402,
|
| 2669 |
+
"loss": 0.355,
|
| 2670 |
+
"step": 9375
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"epoch": 4.845360824742268,
|
| 2674 |
+
"grad_norm": 1.110967755317688,
|
| 2675 |
+
"learning_rate": 0.0002605142348754448,
|
| 2676 |
+
"loss": 0.3208,
|
| 2677 |
+
"step": 9400
|
| 2678 |
+
},
|
| 2679 |
+
{
|
| 2680 |
+
"epoch": 4.858247422680412,
|
| 2681 |
+
"grad_norm": 0.3510916531085968,
|
| 2682 |
+
"learning_rate": 0.00026038078291814947,
|
| 2683 |
+
"loss": 0.3421,
|
| 2684 |
+
"step": 9425
|
| 2685 |
+
},
|
| 2686 |
+
{
|
| 2687 |
+
"epoch": 4.871134020618557,
|
| 2688 |
+
"grad_norm": 0.33312925696372986,
|
| 2689 |
+
"learning_rate": 0.00026024733096085406,
|
| 2690 |
+
"loss": 0.3046,
|
| 2691 |
+
"step": 9450
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 4.8840206185567006,
|
| 2695 |
+
"grad_norm": 0.32453814148902893,
|
| 2696 |
+
"learning_rate": 0.0002601138790035587,
|
| 2697 |
+
"loss": 0.3294,
|
| 2698 |
+
"step": 9475
|
| 2699 |
+
},
|
| 2700 |
+
{
|
| 2701 |
+
"epoch": 4.896907216494846,
|
| 2702 |
+
"grad_norm": 0.3632301092147827,
|
| 2703 |
+
"learning_rate": 0.0002599804270462633,
|
| 2704 |
+
"loss": 0.3261,
|
| 2705 |
+
"step": 9500
|
| 2706 |
+
},
|
| 2707 |
+
{
|
| 2708 |
+
"epoch": 4.90979381443299,
|
| 2709 |
+
"grad_norm": 0.3007452189922333,
|
| 2710 |
+
"learning_rate": 0.00025984697508896795,
|
| 2711 |
+
"loss": 0.3524,
|
| 2712 |
+
"step": 9525
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"epoch": 4.922680412371134,
|
| 2716 |
+
"grad_norm": 0.47182586789131165,
|
| 2717 |
+
"learning_rate": 0.0002597135231316726,
|
| 2718 |
+
"loss": 0.3287,
|
| 2719 |
+
"step": 9550
|
| 2720 |
+
},
|
| 2721 |
+
{
|
| 2722 |
+
"epoch": 4.935567010309279,
|
| 2723 |
+
"grad_norm": 0.3959861397743225,
|
| 2724 |
+
"learning_rate": 0.0002595800711743772,
|
| 2725 |
+
"loss": 0.3134,
|
| 2726 |
+
"step": 9575
|
| 2727 |
+
},
|
| 2728 |
+
{
|
| 2729 |
+
"epoch": 4.948453608247423,
|
| 2730 |
+
"grad_norm": 0.28187814354896545,
|
| 2731 |
+
"learning_rate": 0.00025944661921708183,
|
| 2732 |
+
"loss": 0.3154,
|
| 2733 |
+
"step": 9600
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 4.961340206185567,
|
| 2737 |
+
"grad_norm": 0.32686540484428406,
|
| 2738 |
+
"learning_rate": 0.0002593131672597865,
|
| 2739 |
+
"loss": 0.3464,
|
| 2740 |
+
"step": 9625
|
| 2741 |
+
},
|
| 2742 |
+
{
|
| 2743 |
+
"epoch": 4.974226804123711,
|
| 2744 |
+
"grad_norm": 0.46848079562187195,
|
| 2745 |
+
"learning_rate": 0.0002591797153024911,
|
| 2746 |
+
"loss": 0.2753,
|
| 2747 |
+
"step": 9650
|
| 2748 |
+
},
|
| 2749 |
+
{
|
| 2750 |
+
"epoch": 4.987113402061856,
|
| 2751 |
+
"grad_norm": 0.25264793634414673,
|
| 2752 |
+
"learning_rate": 0.0002590462633451957,
|
| 2753 |
+
"loss": 0.3403,
|
| 2754 |
+
"step": 9675
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 5.0,
|
| 2758 |
+
"grad_norm": 0.4163696765899658,
|
| 2759 |
+
"learning_rate": 0.0002589128113879003,
|
| 2760 |
+
"loss": 0.2669,
|
| 2761 |
+
"step": 9700
|
| 2762 |
+
},
|
| 2763 |
+
{
|
| 2764 |
+
"epoch": 5.012886597938144,
|
| 2765 |
+
"grad_norm": 0.29742294549942017,
|
| 2766 |
+
"learning_rate": 0.00025877935943060496,
|
| 2767 |
+
"loss": 0.3203,
|
| 2768 |
+
"step": 9725
|
| 2769 |
+
},
|
| 2770 |
+
{
|
| 2771 |
+
"epoch": 5.025773195876289,
|
| 2772 |
+
"grad_norm": 0.8651419281959534,
|
| 2773 |
+
"learning_rate": 0.0002586459074733096,
|
| 2774 |
+
"loss": 0.2589,
|
| 2775 |
+
"step": 9750
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 5.038659793814433,
|
| 2779 |
+
"grad_norm": 0.2916441261768341,
|
| 2780 |
+
"learning_rate": 0.0002585124555160142,
|
| 2781 |
+
"loss": 0.3,
|
| 2782 |
+
"step": 9775
|
| 2783 |
+
},
|
| 2784 |
+
{
|
| 2785 |
+
"epoch": 5.051546391752577,
|
| 2786 |
+
"grad_norm": 1.4987293481826782,
|
| 2787 |
+
"learning_rate": 0.00025837900355871884,
|
| 2788 |
+
"loss": 0.3148,
|
| 2789 |
+
"step": 9800
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 5.064432989690721,
|
| 2793 |
+
"grad_norm": 0.46689650416374207,
|
| 2794 |
+
"learning_rate": 0.00025824555160142344,
|
| 2795 |
+
"loss": 0.2935,
|
| 2796 |
+
"step": 9825
|
| 2797 |
+
},
|
| 2798 |
+
{
|
| 2799 |
+
"epoch": 5.077319587628866,
|
| 2800 |
+
"grad_norm": 0.3726617991924286,
|
| 2801 |
+
"learning_rate": 0.0002581120996441281,
|
| 2802 |
+
"loss": 0.2871,
|
| 2803 |
+
"step": 9850
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"epoch": 5.09020618556701,
|
| 2807 |
+
"grad_norm": 0.29355180263519287,
|
| 2808 |
+
"learning_rate": 0.00025797864768683273,
|
| 2809 |
+
"loss": 0.2422,
|
| 2810 |
+
"step": 9875
|
| 2811 |
+
},
|
| 2812 |
+
{
|
| 2813 |
+
"epoch": 5.103092783505154,
|
| 2814 |
+
"grad_norm": 0.6249547600746155,
|
| 2815 |
+
"learning_rate": 0.0002578451957295373,
|
| 2816 |
+
"loss": 0.2749,
|
| 2817 |
+
"step": 9900
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 5.115979381443299,
|
| 2821 |
+
"grad_norm": 0.42613521218299866,
|
| 2822 |
+
"learning_rate": 0.00025771174377224197,
|
| 2823 |
+
"loss": 0.2998,
|
| 2824 |
+
"step": 9925
|
| 2825 |
+
},
|
| 2826 |
+
{
|
| 2827 |
+
"epoch": 5.128865979381443,
|
| 2828 |
+
"grad_norm": 0.4285449981689453,
|
| 2829 |
+
"learning_rate": 0.00025757829181494656,
|
| 2830 |
+
"loss": 0.3241,
|
| 2831 |
+
"step": 9950
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 5.141752577319588,
|
| 2835 |
+
"grad_norm": 0.2756471037864685,
|
| 2836 |
+
"learning_rate": 0.0002574448398576512,
|
| 2837 |
+
"loss": 0.2788,
|
| 2838 |
+
"step": 9975
|
| 2839 |
+
},
|
| 2840 |
+
{
|
| 2841 |
+
"epoch": 5.154639175257732,
|
| 2842 |
+
"grad_norm": 0.650787889957428,
|
| 2843 |
+
"learning_rate": 0.00025731138790035586,
|
| 2844 |
+
"loss": 0.2628,
|
| 2845 |
+
"step": 10000
|
| 2846 |
+
},
|
| 2847 |
+
{
|
| 2848 |
+
"epoch": 5.154639175257732,
|
| 2849 |
+
"eval_loss": 0.5844400525093079,
|
| 2850 |
+
"eval_runtime": 196.2011,
|
| 2851 |
+
"eval_samples_per_second": 6.244,
|
| 2852 |
+
"eval_steps_per_second": 0.785,
|
| 2853 |
+
"eval_wer": 1.006530612244898,
|
| 2854 |
+
"step": 10000
|
| 2855 |
+
},
|
| 2856 |
+
{
|
| 2857 |
+
"epoch": 5.154639175257732,
|
| 2858 |
+
"step": 10000,
|
| 2859 |
+
"total_flos": 5.52048529474528e+19,
|
| 2860 |
+
"train_loss": 0.6429409675121307,
|
| 2861 |
+
"train_runtime": 62307.8234,
|
| 2862 |
+
"train_samples_per_second": 22.413,
|
| 2863 |
+
"train_steps_per_second": 0.934
|
| 2864 |
+
}
|
| 2865 |
+
],
|
| 2866 |
+
"logging_steps": 25,
|
| 2867 |
+
"max_steps": 58200,
|
| 2868 |
+
"num_input_tokens_seen": 0,
|
| 2869 |
+
"num_train_epochs": 30,
|
| 2870 |
+
"save_steps": 2000,
|
| 2871 |
+
"stateful_callbacks": {
|
| 2872 |
+
"EarlyStoppingCallback": {
|
| 2873 |
+
"args": {
|
| 2874 |
+
"early_stopping_patience": 3,
|
| 2875 |
+
"early_stopping_threshold": 0.0
|
| 2876 |
+
},
|
| 2877 |
+
"attributes": {
|
| 2878 |
+
"early_stopping_patience_counter": 0
|
| 2879 |
+
}
|
| 2880 |
+
},
|
| 2881 |
+
"TrainerControl": {
|
| 2882 |
+
"args": {
|
| 2883 |
+
"should_epoch_stop": false,
|
| 2884 |
+
"should_evaluate": false,
|
| 2885 |
+
"should_log": false,
|
| 2886 |
+
"should_save": true,
|
| 2887 |
+
"should_training_stop": true
|
| 2888 |
+
},
|
| 2889 |
+
"attributes": {}
|
| 2890 |
+
}
|
| 2891 |
+
},
|
| 2892 |
+
"total_flos": 5.52048529474528e+19,
|
| 2893 |
+
"train_batch_size": 12,
|
| 2894 |
+
"trial_name": null,
|
| 2895 |
+
"trial_params": null
|
| 2896 |
+
}
|