End of training
Browse files- README.md +6 -3
- adapter.bem.safetensors +3 -0
- all_results.json +15 -0
- eval_results.json +9 -0
- train_results.json +9 -0
- trainer_state.json +938 -0
README.md
CHANGED
|
@@ -3,6 +3,9 @@ library_name: transformers
|
|
| 3 |
license: cc-by-nc-4.0
|
| 4 |
base_model: mms-meta/mms-zeroshot-300m
|
| 5 |
tags:
|
|
|
|
|
|
|
|
|
|
| 6 |
- generated_from_trainer
|
| 7 |
metrics:
|
| 8 |
- wer
|
|
@@ -16,10 +19,10 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
|
| 17 |
# mms-zeroshot-300m-bem
|
| 18 |
|
| 19 |
-
This model is a fine-tuned version of [mms-meta/mms-zeroshot-300m](https://huggingface.co/mms-meta/mms-zeroshot-300m) on
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
-
- Loss: 0.
|
| 22 |
-
- Wer: 0.
|
| 23 |
|
| 24 |
## Model description
|
| 25 |
|
|
|
|
| 3 |
license: cc-by-nc-4.0
|
| 4 |
base_model: mms-meta/mms-zeroshot-300m
|
| 5 |
tags:
|
| 6 |
+
- automatic-speech-recognition
|
| 7 |
+
- BembaSpeech
|
| 8 |
+
- mms
|
| 9 |
- generated_from_trainer
|
| 10 |
metrics:
|
| 11 |
- wer
|
|
|
|
| 19 |
|
| 20 |
# mms-zeroshot-300m-bem
|
| 21 |
|
| 22 |
+
This model is a fine-tuned version of [mms-meta/mms-zeroshot-300m](https://huggingface.co/mms-meta/mms-zeroshot-300m) on the BEMBASPEECH - BEM dataset.
|
| 23 |
It achieves the following results on the evaluation set:
|
| 24 |
+
- Loss: 0.1787
|
| 25 |
+
- Wer: 0.3583
|
| 26 |
|
| 27 |
## Model description
|
| 28 |
|
adapter.bem.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c7a58094c83a05c3e25756ebeb3a6147a50aa370bf77f26c63e99bad9d9e6317
|
| 3 |
+
size 3586804
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_loss": 0.17866507172584534,
|
| 4 |
+
"eval_runtime": 77.7914,
|
| 5 |
+
"eval_samples": 1479,
|
| 6 |
+
"eval_samples_per_second": 19.012,
|
| 7 |
+
"eval_steps_per_second": 2.378,
|
| 8 |
+
"eval_wer": 0.3583289171524466,
|
| 9 |
+
"total_flos": 2.1329860467760157e+19,
|
| 10 |
+
"train_loss": 0.48163046043254915,
|
| 11 |
+
"train_runtime": 15300.4436,
|
| 12 |
+
"train_samples": 11245,
|
| 13 |
+
"train_samples_per_second": 7.349,
|
| 14 |
+
"train_steps_per_second": 1.838
|
| 15 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_loss": 0.17866507172584534,
|
| 4 |
+
"eval_runtime": 77.7914,
|
| 5 |
+
"eval_samples": 1479,
|
| 6 |
+
"eval_samples_per_second": 19.012,
|
| 7 |
+
"eval_steps_per_second": 2.378,
|
| 8 |
+
"eval_wer": 0.3583289171524466
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"total_flos": 2.1329860467760157e+19,
|
| 4 |
+
"train_loss": 0.48163046043254915,
|
| 5 |
+
"train_runtime": 15300.4436,
|
| 6 |
+
"train_samples": 11245,
|
| 7 |
+
"train_samples_per_second": 7.349,
|
| 8 |
+
"train_steps_per_second": 1.838
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,938 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 28120,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.17780938833570412,
|
| 13 |
+
"grad_norm": 1.8023470640182495,
|
| 14 |
+
"learning_rate": 0.0002988,
|
| 15 |
+
"loss": 4.6629,
|
| 16 |
+
"step": 500
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.17780938833570412,
|
| 20 |
+
"eval_loss": 0.3540094494819641,
|
| 21 |
+
"eval_runtime": 77.5992,
|
| 22 |
+
"eval_samples_per_second": 19.059,
|
| 23 |
+
"eval_steps_per_second": 2.384,
|
| 24 |
+
"eval_wer": 0.5421303656597775,
|
| 25 |
+
"step": 500
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.35561877667140823,
|
| 29 |
+
"grad_norm": 1.411382794380188,
|
| 30 |
+
"learning_rate": 0.0002945908761766835,
|
| 31 |
+
"loss": 0.6579,
|
| 32 |
+
"step": 1000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.35561877667140823,
|
| 36 |
+
"eval_loss": 0.25881391763687134,
|
| 37 |
+
"eval_runtime": 77.5785,
|
| 38 |
+
"eval_samples_per_second": 19.065,
|
| 39 |
+
"eval_steps_per_second": 2.385,
|
| 40 |
+
"eval_wer": 0.48825295884119413,
|
| 41 |
+
"step": 1000
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.5334281650071123,
|
| 45 |
+
"grad_norm": 1.561378836631775,
|
| 46 |
+
"learning_rate": 0.00028916002896451845,
|
| 47 |
+
"loss": 0.591,
|
| 48 |
+
"step": 1500
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 0.5334281650071123,
|
| 52 |
+
"eval_loss": 0.25524866580963135,
|
| 53 |
+
"eval_runtime": 77.3789,
|
| 54 |
+
"eval_samples_per_second": 19.114,
|
| 55 |
+
"eval_steps_per_second": 2.391,
|
| 56 |
+
"eval_wer": 0.47200141317788374,
|
| 57 |
+
"step": 1500
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.7112375533428165,
|
| 61 |
+
"grad_norm": 1.4678888320922852,
|
| 62 |
+
"learning_rate": 0.00028372918175235333,
|
| 63 |
+
"loss": 0.5467,
|
| 64 |
+
"step": 2000
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 0.7112375533428165,
|
| 68 |
+
"eval_loss": 0.23701371252536774,
|
| 69 |
+
"eval_runtime": 77.4845,
|
| 70 |
+
"eval_samples_per_second": 19.088,
|
| 71 |
+
"eval_steps_per_second": 2.388,
|
| 72 |
+
"eval_wer": 0.4541600423953365,
|
| 73 |
+
"step": 2000
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.8890469416785206,
|
| 77 |
+
"grad_norm": 2.583188056945801,
|
| 78 |
+
"learning_rate": 0.00027829833454018826,
|
| 79 |
+
"loss": 0.5405,
|
| 80 |
+
"step": 2500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.8890469416785206,
|
| 84 |
+
"eval_loss": 0.23755988478660583,
|
| 85 |
+
"eval_runtime": 78.0759,
|
| 86 |
+
"eval_samples_per_second": 18.943,
|
| 87 |
+
"eval_steps_per_second": 2.369,
|
| 88 |
+
"eval_wer": 0.45557322027910263,
|
| 89 |
+
"step": 2500
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 1.0668563300142249,
|
| 93 |
+
"grad_norm": 1.422250509262085,
|
| 94 |
+
"learning_rate": 0.00027286748732802314,
|
| 95 |
+
"loss": 0.5027,
|
| 96 |
+
"step": 3000
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.0668563300142249,
|
| 100 |
+
"eval_loss": 0.22338581085205078,
|
| 101 |
+
"eval_runtime": 77.6024,
|
| 102 |
+
"eval_samples_per_second": 19.059,
|
| 103 |
+
"eval_steps_per_second": 2.384,
|
| 104 |
+
"eval_wer": 0.4306659600777248,
|
| 105 |
+
"step": 3000
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.2446657183499288,
|
| 109 |
+
"grad_norm": 1.006721019744873,
|
| 110 |
+
"learning_rate": 0.00026743664011585807,
|
| 111 |
+
"loss": 0.5001,
|
| 112 |
+
"step": 3500
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.2446657183499288,
|
| 116 |
+
"eval_loss": 0.21763387322425842,
|
| 117 |
+
"eval_runtime": 77.5866,
|
| 118 |
+
"eval_samples_per_second": 19.063,
|
| 119 |
+
"eval_steps_per_second": 2.384,
|
| 120 |
+
"eval_wer": 0.42130365659777425,
|
| 121 |
+
"step": 3500
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 1.422475106685633,
|
| 125 |
+
"grad_norm": 1.3477824926376343,
|
| 126 |
+
"learning_rate": 0.00026200579290369295,
|
| 127 |
+
"loss": 0.4962,
|
| 128 |
+
"step": 4000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.422475106685633,
|
| 132 |
+
"eval_loss": 0.21994474530220032,
|
| 133 |
+
"eval_runtime": 77.7546,
|
| 134 |
+
"eval_samples_per_second": 19.021,
|
| 135 |
+
"eval_steps_per_second": 2.379,
|
| 136 |
+
"eval_wer": 0.4205087440381558,
|
| 137 |
+
"step": 4000
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 1.600284495021337,
|
| 141 |
+
"grad_norm": 1.08402419090271,
|
| 142 |
+
"learning_rate": 0.00025657494569152783,
|
| 143 |
+
"loss": 0.486,
|
| 144 |
+
"step": 4500
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 1.600284495021337,
|
| 148 |
+
"eval_loss": 0.21454988420009613,
|
| 149 |
+
"eval_runtime": 77.6412,
|
| 150 |
+
"eval_samples_per_second": 19.049,
|
| 151 |
+
"eval_steps_per_second": 2.383,
|
| 152 |
+
"eval_wer": 0.41671082847553437,
|
| 153 |
+
"step": 4500
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 1.7780938833570412,
|
| 157 |
+
"grad_norm": 0.5935032963752747,
|
| 158 |
+
"learning_rate": 0.00025114409847936276,
|
| 159 |
+
"loss": 0.47,
|
| 160 |
+
"step": 5000
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 1.7780938833570412,
|
| 164 |
+
"eval_loss": 0.21590569615364075,
|
| 165 |
+
"eval_runtime": 77.5792,
|
| 166 |
+
"eval_samples_per_second": 19.064,
|
| 167 |
+
"eval_steps_per_second": 2.385,
|
| 168 |
+
"eval_wer": 0.4168874757110051,
|
| 169 |
+
"step": 5000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 1.9559032716927454,
|
| 173 |
+
"grad_norm": 1.4301828145980835,
|
| 174 |
+
"learning_rate": 0.00024571325126719764,
|
| 175 |
+
"loss": 0.4557,
|
| 176 |
+
"step": 5500
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 1.9559032716927454,
|
| 180 |
+
"eval_loss": 0.2098698765039444,
|
| 181 |
+
"eval_runtime": 78.1464,
|
| 182 |
+
"eval_samples_per_second": 18.926,
|
| 183 |
+
"eval_steps_per_second": 2.367,
|
| 184 |
+
"eval_wer": 0.4135311782370606,
|
| 185 |
+
"step": 5500
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 2.1337126600284497,
|
| 189 |
+
"grad_norm": 1.2725244760513306,
|
| 190 |
+
"learning_rate": 0.00024028240405503257,
|
| 191 |
+
"loss": 0.4514,
|
| 192 |
+
"step": 6000
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 2.1337126600284497,
|
| 196 |
+
"eval_loss": 0.20907503366470337,
|
| 197 |
+
"eval_runtime": 78.1315,
|
| 198 |
+
"eval_samples_per_second": 18.93,
|
| 199 |
+
"eval_steps_per_second": 2.368,
|
| 200 |
+
"eval_wer": 0.4099982335276453,
|
| 201 |
+
"step": 6000
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 2.3115220483641536,
|
| 205 |
+
"grad_norm": 1.0271036624908447,
|
| 206 |
+
"learning_rate": 0.00023485155684286748,
|
| 207 |
+
"loss": 0.4539,
|
| 208 |
+
"step": 6500
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 2.3115220483641536,
|
| 212 |
+
"eval_loss": 0.2038286179304123,
|
| 213 |
+
"eval_runtime": 77.8532,
|
| 214 |
+
"eval_samples_per_second": 18.997,
|
| 215 |
+
"eval_steps_per_second": 2.376,
|
| 216 |
+
"eval_wer": 0.40160748984278394,
|
| 217 |
+
"step": 6500
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 2.4893314366998576,
|
| 221 |
+
"grad_norm": 1.0436575412750244,
|
| 222 |
+
"learning_rate": 0.0002294315713251267,
|
| 223 |
+
"loss": 0.439,
|
| 224 |
+
"step": 7000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 2.4893314366998576,
|
| 228 |
+
"eval_loss": 0.20407754182815552,
|
| 229 |
+
"eval_runtime": 78.101,
|
| 230 |
+
"eval_samples_per_second": 18.937,
|
| 231 |
+
"eval_steps_per_second": 2.369,
|
| 232 |
+
"eval_wer": 0.4024907260201378,
|
| 233 |
+
"step": 7000
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 2.667140825035562,
|
| 237 |
+
"grad_norm": 1.0521398782730103,
|
| 238 |
+
"learning_rate": 0.00022400072411296162,
|
| 239 |
+
"loss": 0.4378,
|
| 240 |
+
"step": 7500
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 2.667140825035562,
|
| 244 |
+
"eval_loss": 0.20021408796310425,
|
| 245 |
+
"eval_runtime": 78.1969,
|
| 246 |
+
"eval_samples_per_second": 18.914,
|
| 247 |
+
"eval_steps_per_second": 2.366,
|
| 248 |
+
"eval_wer": 0.39162692103868574,
|
| 249 |
+
"step": 7500
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 2.844950213371266,
|
| 253 |
+
"grad_norm": 0.716386616230011,
|
| 254 |
+
"learning_rate": 0.0002185698769007965,
|
| 255 |
+
"loss": 0.4347,
|
| 256 |
+
"step": 8000
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 2.844950213371266,
|
| 260 |
+
"eval_loss": 0.19606797397136688,
|
| 261 |
+
"eval_runtime": 78.0563,
|
| 262 |
+
"eval_samples_per_second": 18.948,
|
| 263 |
+
"eval_steps_per_second": 2.37,
|
| 264 |
+
"eval_wer": 0.39109697933227344,
|
| 265 |
+
"step": 8000
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 3.0227596017069702,
|
| 269 |
+
"grad_norm": 0.6210708022117615,
|
| 270 |
+
"learning_rate": 0.0002131390296886314,
|
| 271 |
+
"loss": 0.4278,
|
| 272 |
+
"step": 8500
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 3.0227596017069702,
|
| 276 |
+
"eval_loss": 0.1994515061378479,
|
| 277 |
+
"eval_runtime": 78.0396,
|
| 278 |
+
"eval_samples_per_second": 18.952,
|
| 279 |
+
"eval_steps_per_second": 2.371,
|
| 280 |
+
"eval_wer": 0.3923335099805688,
|
| 281 |
+
"step": 8500
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 3.200568990042674,
|
| 285 |
+
"grad_norm": 0.4629976153373718,
|
| 286 |
+
"learning_rate": 0.0002077081824764663,
|
| 287 |
+
"loss": 0.4117,
|
| 288 |
+
"step": 9000
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 3.200568990042674,
|
| 292 |
+
"eval_loss": 0.19594013690948486,
|
| 293 |
+
"eval_runtime": 77.8452,
|
| 294 |
+
"eval_samples_per_second": 18.999,
|
| 295 |
+
"eval_steps_per_second": 2.377,
|
| 296 |
+
"eval_wer": 0.38915385974209504,
|
| 297 |
+
"step": 9000
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 3.3783783783783785,
|
| 301 |
+
"grad_norm": 0.9285233020782471,
|
| 302 |
+
"learning_rate": 0.0002022773352643012,
|
| 303 |
+
"loss": 0.4149,
|
| 304 |
+
"step": 9500
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 3.3783783783783785,
|
| 308 |
+
"eval_loss": 0.1925920844078064,
|
| 309 |
+
"eval_runtime": 77.8068,
|
| 310 |
+
"eval_samples_per_second": 19.009,
|
| 311 |
+
"eval_steps_per_second": 2.378,
|
| 312 |
+
"eval_wer": 0.3858858858858859,
|
| 313 |
+
"step": 9500
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 3.5561877667140824,
|
| 317 |
+
"grad_norm": 1.0625221729278564,
|
| 318 |
+
"learning_rate": 0.00019684648805213612,
|
| 319 |
+
"loss": 0.4148,
|
| 320 |
+
"step": 10000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 3.5561877667140824,
|
| 324 |
+
"eval_loss": 0.1958448737859726,
|
| 325 |
+
"eval_runtime": 77.9401,
|
| 326 |
+
"eval_samples_per_second": 18.976,
|
| 327 |
+
"eval_steps_per_second": 2.374,
|
| 328 |
+
"eval_wer": 0.38040982158629216,
|
| 329 |
+
"step": 10000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 3.733997155049787,
|
| 333 |
+
"grad_norm": 0.7951219081878662,
|
| 334 |
+
"learning_rate": 0.00019142650253439536,
|
| 335 |
+
"loss": 0.4009,
|
| 336 |
+
"step": 10500
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 3.733997155049787,
|
| 340 |
+
"eval_loss": 0.19297942519187927,
|
| 341 |
+
"eval_runtime": 77.4746,
|
| 342 |
+
"eval_samples_per_second": 19.09,
|
| 343 |
+
"eval_steps_per_second": 2.388,
|
| 344 |
+
"eval_wer": 0.37899664370252606,
|
| 345 |
+
"step": 10500
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 3.9118065433854907,
|
| 349 |
+
"grad_norm": 0.689900279045105,
|
| 350 |
+
"learning_rate": 0.00018599565532223026,
|
| 351 |
+
"loss": 0.4174,
|
| 352 |
+
"step": 11000
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 3.9118065433854907,
|
| 356 |
+
"eval_loss": 0.19552023708820343,
|
| 357 |
+
"eval_runtime": 77.7104,
|
| 358 |
+
"eval_samples_per_second": 19.032,
|
| 359 |
+
"eval_steps_per_second": 2.381,
|
| 360 |
+
"eval_wer": 0.3822646175587352,
|
| 361 |
+
"step": 11000
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 4.089615931721195,
|
| 365 |
+
"grad_norm": 0.5333609580993652,
|
| 366 |
+
"learning_rate": 0.00018056480811006514,
|
| 367 |
+
"loss": 0.4012,
|
| 368 |
+
"step": 11500
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 4.089615931721195,
|
| 372 |
+
"eval_loss": 0.19501054286956787,
|
| 373 |
+
"eval_runtime": 77.674,
|
| 374 |
+
"eval_samples_per_second": 19.041,
|
| 375 |
+
"eval_steps_per_second": 2.382,
|
| 376 |
+
"eval_wer": 0.3812047341459106,
|
| 377 |
+
"step": 11500
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 4.2674253200568995,
|
| 381 |
+
"grad_norm": 1.5428721904754639,
|
| 382 |
+
"learning_rate": 0.00017513396089790005,
|
| 383 |
+
"loss": 0.3974,
|
| 384 |
+
"step": 12000
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 4.2674253200568995,
|
| 388 |
+
"eval_loss": 0.19340351223945618,
|
| 389 |
+
"eval_runtime": 78.0603,
|
| 390 |
+
"eval_samples_per_second": 18.947,
|
| 391 |
+
"eval_steps_per_second": 2.37,
|
| 392 |
+
"eval_wer": 0.3773184949655538,
|
| 393 |
+
"step": 12000
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 4.445234708392603,
|
| 397 |
+
"grad_norm": 0.44063669443130493,
|
| 398 |
+
"learning_rate": 0.00016970311368573495,
|
| 399 |
+
"loss": 0.3943,
|
| 400 |
+
"step": 12500
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 4.445234708392603,
|
| 404 |
+
"eval_loss": 0.18450064957141876,
|
| 405 |
+
"eval_runtime": 78.009,
|
| 406 |
+
"eval_samples_per_second": 18.959,
|
| 407 |
+
"eval_steps_per_second": 2.372,
|
| 408 |
+
"eval_wer": 0.37201907790143085,
|
| 409 |
+
"step": 12500
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 4.623044096728307,
|
| 413 |
+
"grad_norm": 0.9376386404037476,
|
| 414 |
+
"learning_rate": 0.0001642831281679942,
|
| 415 |
+
"loss": 0.4071,
|
| 416 |
+
"step": 13000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 4.623044096728307,
|
| 420 |
+
"eval_loss": 0.1920066624879837,
|
| 421 |
+
"eval_runtime": 77.3116,
|
| 422 |
+
"eval_samples_per_second": 19.13,
|
| 423 |
+
"eval_steps_per_second": 2.393,
|
| 424 |
+
"eval_wer": 0.3838544426779721,
|
| 425 |
+
"step": 13000
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 4.800853485064011,
|
| 429 |
+
"grad_norm": 0.6557429432868958,
|
| 430 |
+
"learning_rate": 0.0001588522809558291,
|
| 431 |
+
"loss": 0.3968,
|
| 432 |
+
"step": 13500
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 4.800853485064011,
|
| 436 |
+
"eval_loss": 0.1866944283246994,
|
| 437 |
+
"eval_runtime": 77.5081,
|
| 438 |
+
"eval_samples_per_second": 19.082,
|
| 439 |
+
"eval_steps_per_second": 2.387,
|
| 440 |
+
"eval_wer": 0.37431549196255076,
|
| 441 |
+
"step": 13500
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 4.978662873399715,
|
| 445 |
+
"grad_norm": 0.5748176574707031,
|
| 446 |
+
"learning_rate": 0.000153421433743664,
|
| 447 |
+
"loss": 0.3795,
|
| 448 |
+
"step": 14000
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 4.978662873399715,
|
| 452 |
+
"eval_loss": 0.18717192113399506,
|
| 453 |
+
"eval_runtime": 77.4836,
|
| 454 |
+
"eval_samples_per_second": 19.088,
|
| 455 |
+
"eval_steps_per_second": 2.388,
|
| 456 |
+
"eval_wer": 0.37131248895954777,
|
| 457 |
+
"step": 14000
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 5.15647226173542,
|
| 461 |
+
"grad_norm": 0.8843936920166016,
|
| 462 |
+
"learning_rate": 0.0001479905865314989,
|
| 463 |
+
"loss": 0.3856,
|
| 464 |
+
"step": 14500
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 5.15647226173542,
|
| 468 |
+
"eval_loss": 0.18692350387573242,
|
| 469 |
+
"eval_runtime": 77.4161,
|
| 470 |
+
"eval_samples_per_second": 19.105,
|
| 471 |
+
"eval_steps_per_second": 2.39,
|
| 472 |
+
"eval_wer": 0.3736972266384031,
|
| 473 |
+
"step": 14500
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 5.334281650071124,
|
| 477 |
+
"grad_norm": 1.5647237300872803,
|
| 478 |
+
"learning_rate": 0.0001425597393193338,
|
| 479 |
+
"loss": 0.3706,
|
| 480 |
+
"step": 15000
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 5.334281650071124,
|
| 484 |
+
"eval_loss": 0.19033658504486084,
|
| 485 |
+
"eval_runtime": 77.561,
|
| 486 |
+
"eval_samples_per_second": 19.069,
|
| 487 |
+
"eval_steps_per_second": 2.385,
|
| 488 |
+
"eval_wer": 0.37661190602367073,
|
| 489 |
+
"step": 15000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 5.512091038406828,
|
| 493 |
+
"grad_norm": 2.1427793502807617,
|
| 494 |
+
"learning_rate": 0.00013713975380159305,
|
| 495 |
+
"loss": 0.3784,
|
| 496 |
+
"step": 15500
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 5.512091038406828,
|
| 500 |
+
"eval_loss": 0.18607346713542938,
|
| 501 |
+
"eval_runtime": 77.668,
|
| 502 |
+
"eval_samples_per_second": 19.043,
|
| 503 |
+
"eval_steps_per_second": 2.382,
|
| 504 |
+
"eval_wer": 0.3683094859565448,
|
| 505 |
+
"step": 15500
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 5.689900426742532,
|
| 509 |
+
"grad_norm": 2.697434186935425,
|
| 510 |
+
"learning_rate": 0.00013170890658942793,
|
| 511 |
+
"loss": 0.3777,
|
| 512 |
+
"step": 16000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 5.689900426742532,
|
| 516 |
+
"eval_loss": 0.1866033524274826,
|
| 517 |
+
"eval_runtime": 77.9873,
|
| 518 |
+
"eval_samples_per_second": 18.965,
|
| 519 |
+
"eval_steps_per_second": 2.372,
|
| 520 |
+
"eval_wer": 0.37131248895954777,
|
| 521 |
+
"step": 16000
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 5.867709815078236,
|
| 525 |
+
"grad_norm": 1.5033948421478271,
|
| 526 |
+
"learning_rate": 0.00012627805937726286,
|
| 527 |
+
"loss": 0.3861,
|
| 528 |
+
"step": 16500
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 5.867709815078236,
|
| 532 |
+
"eval_loss": 0.18121445178985596,
|
| 533 |
+
"eval_runtime": 77.6152,
|
| 534 |
+
"eval_samples_per_second": 19.056,
|
| 535 |
+
"eval_steps_per_second": 2.384,
|
| 536 |
+
"eval_wer": 0.3637166578343049,
|
| 537 |
+
"step": 16500
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 6.0455192034139404,
|
| 541 |
+
"grad_norm": 0.5671353936195374,
|
| 542 |
+
"learning_rate": 0.00012084721216509774,
|
| 543 |
+
"loss": 0.3711,
|
| 544 |
+
"step": 17000
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"epoch": 6.0455192034139404,
|
| 548 |
+
"eval_loss": 0.18417000770568848,
|
| 549 |
+
"eval_runtime": 77.4756,
|
| 550 |
+
"eval_samples_per_second": 19.09,
|
| 551 |
+
"eval_steps_per_second": 2.388,
|
| 552 |
+
"eval_wer": 0.3667196608373079,
|
| 553 |
+
"step": 17000
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 6.223328591749644,
|
| 557 |
+
"grad_norm": 1.0996285676956177,
|
| 558 |
+
"learning_rate": 0.00011542722664735697,
|
| 559 |
+
"loss": 0.374,
|
| 560 |
+
"step": 17500
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 6.223328591749644,
|
| 564 |
+
"eval_loss": 0.18148785829544067,
|
| 565 |
+
"eval_runtime": 77.7259,
|
| 566 |
+
"eval_samples_per_second": 19.028,
|
| 567 |
+
"eval_steps_per_second": 2.38,
|
| 568 |
+
"eval_wer": 0.3617735382441265,
|
| 569 |
+
"step": 17500
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 6.401137980085348,
|
| 573 |
+
"grad_norm": 0.6861454844474792,
|
| 574 |
+
"learning_rate": 0.00010999637943519187,
|
| 575 |
+
"loss": 0.3539,
|
| 576 |
+
"step": 18000
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 6.401137980085348,
|
| 580 |
+
"eval_loss": 0.18153779208660126,
|
| 581 |
+
"eval_runtime": 77.7594,
|
| 582 |
+
"eval_samples_per_second": 19.02,
|
| 583 |
+
"eval_steps_per_second": 2.379,
|
| 584 |
+
"eval_wer": 0.3646882176293941,
|
| 585 |
+
"step": 18000
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 6.578947368421053,
|
| 589 |
+
"grad_norm": 1.423963189125061,
|
| 590 |
+
"learning_rate": 0.00010456553222302678,
|
| 591 |
+
"loss": 0.3625,
|
| 592 |
+
"step": 18500
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 6.578947368421053,
|
| 596 |
+
"eval_loss": 0.17849859595298767,
|
| 597 |
+
"eval_runtime": 77.8982,
|
| 598 |
+
"eval_samples_per_second": 18.986,
|
| 599 |
+
"eval_steps_per_second": 2.375,
|
| 600 |
+
"eval_wer": 0.3588588588588589,
|
| 601 |
+
"step": 18500
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 6.756756756756757,
|
| 605 |
+
"grad_norm": 0.9670858383178711,
|
| 606 |
+
"learning_rate": 9.913468501086169e-05,
|
| 607 |
+
"loss": 0.3599,
|
| 608 |
+
"step": 19000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 6.756756756756757,
|
| 612 |
+
"eval_loss": 0.17952215671539307,
|
| 613 |
+
"eval_runtime": 77.3124,
|
| 614 |
+
"eval_samples_per_second": 19.13,
|
| 615 |
+
"eval_steps_per_second": 2.393,
|
| 616 |
+
"eval_wer": 0.362126832715068,
|
| 617 |
+
"step": 19000
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 6.934566145092461,
|
| 621 |
+
"grad_norm": 0.523705005645752,
|
| 622 |
+
"learning_rate": 9.370383779869658e-05,
|
| 623 |
+
"loss": 0.3654,
|
| 624 |
+
"step": 19500
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 6.934566145092461,
|
| 628 |
+
"eval_loss": 0.1822131723165512,
|
| 629 |
+
"eval_runtime": 77.744,
|
| 630 |
+
"eval_samples_per_second": 19.024,
|
| 631 |
+
"eval_steps_per_second": 2.38,
|
| 632 |
+
"eval_wer": 0.36239180356827416,
|
| 633 |
+
"step": 19500
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 7.112375533428165,
|
| 637 |
+
"grad_norm": 1.0047301054000854,
|
| 638 |
+
"learning_rate": 8.82729905865315e-05,
|
| 639 |
+
"loss": 0.3693,
|
| 640 |
+
"step": 20000
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 7.112375533428165,
|
| 644 |
+
"eval_loss": 0.17921391129493713,
|
| 645 |
+
"eval_runtime": 77.3373,
|
| 646 |
+
"eval_samples_per_second": 19.124,
|
| 647 |
+
"eval_steps_per_second": 2.392,
|
| 648 |
+
"eval_wer": 0.3611552729199788,
|
| 649 |
+
"step": 20000
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 7.290184921763869,
|
| 653 |
+
"grad_norm": 0.6277859807014465,
|
| 654 |
+
"learning_rate": 8.285300506879071e-05,
|
| 655 |
+
"loss": 0.3519,
|
| 656 |
+
"step": 20500
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 7.290184921763869,
|
| 660 |
+
"eval_loss": 0.18002206087112427,
|
| 661 |
+
"eval_runtime": 77.6246,
|
| 662 |
+
"eval_samples_per_second": 19.053,
|
| 663 |
+
"eval_steps_per_second": 2.383,
|
| 664 |
+
"eval_wer": 0.36751457339692634,
|
| 665 |
+
"step": 20500
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 7.467994310099574,
|
| 669 |
+
"grad_norm": 0.8712663650512695,
|
| 670 |
+
"learning_rate": 7.743301955104996e-05,
|
| 671 |
+
"loss": 0.3553,
|
| 672 |
+
"step": 21000
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 7.467994310099574,
|
| 676 |
+
"eval_loss": 0.1808168739080429,
|
| 677 |
+
"eval_runtime": 77.4403,
|
| 678 |
+
"eval_samples_per_second": 19.099,
|
| 679 |
+
"eval_steps_per_second": 2.389,
|
| 680 |
+
"eval_wer": 0.36398162868751105,
|
| 681 |
+
"step": 21000
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 7.6458036984352775,
|
| 685 |
+
"grad_norm": 1.000291109085083,
|
| 686 |
+
"learning_rate": 7.200217233888485e-05,
|
| 687 |
+
"loss": 0.3451,
|
| 688 |
+
"step": 21500
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 7.6458036984352775,
|
| 692 |
+
"eval_loss": 0.18079166114330292,
|
| 693 |
+
"eval_runtime": 77.717,
|
| 694 |
+
"eval_samples_per_second": 19.031,
|
| 695 |
+
"eval_steps_per_second": 2.38,
|
| 696 |
+
"eval_wer": 0.36195018547959723,
|
| 697 |
+
"step": 21500
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 7.823613086770981,
|
| 701 |
+
"grad_norm": 0.9150896668434143,
|
| 702 |
+
"learning_rate": 6.657132512671976e-05,
|
| 703 |
+
"loss": 0.3558,
|
| 704 |
+
"step": 22000
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 7.823613086770981,
|
| 708 |
+
"eval_loss": 0.17938227951526642,
|
| 709 |
+
"eval_runtime": 77.6629,
|
| 710 |
+
"eval_samples_per_second": 19.044,
|
| 711 |
+
"eval_steps_per_second": 2.382,
|
| 712 |
+
"eval_wer": 0.36097862568450806,
|
| 713 |
+
"step": 22000
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 8.001422475106686,
|
| 717 |
+
"grad_norm": 0.4225611686706543,
|
| 718 |
+
"learning_rate": 6.114047791455467e-05,
|
| 719 |
+
"loss": 0.3595,
|
| 720 |
+
"step": 22500
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 8.001422475106686,
|
| 724 |
+
"eval_loss": 0.17718034982681274,
|
| 725 |
+
"eval_runtime": 78.0056,
|
| 726 |
+
"eval_samples_per_second": 18.96,
|
| 727 |
+
"eval_steps_per_second": 2.372,
|
| 728 |
+
"eval_wer": 0.3576223282105635,
|
| 729 |
+
"step": 22500
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 8.17923186344239,
|
| 733 |
+
"grad_norm": 0.7559336423873901,
|
| 734 |
+
"learning_rate": 5.570963070238957e-05,
|
| 735 |
+
"loss": 0.3404,
|
| 736 |
+
"step": 23000
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 8.17923186344239,
|
| 740 |
+
"eval_loss": 0.17881204187870026,
|
| 741 |
+
"eval_runtime": 78.2489,
|
| 742 |
+
"eval_samples_per_second": 18.901,
|
| 743 |
+
"eval_steps_per_second": 2.364,
|
| 744 |
+
"eval_wer": 0.35806394629924043,
|
| 745 |
+
"step": 23000
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 8.357041251778094,
|
| 749 |
+
"grad_norm": 1.2298369407653809,
|
| 750 |
+
"learning_rate": 5.027878349022447e-05,
|
| 751 |
+
"loss": 0.3593,
|
| 752 |
+
"step": 23500
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 8.357041251778094,
|
| 756 |
+
"eval_loss": 0.1782107651233673,
|
| 757 |
+
"eval_runtime": 77.8969,
|
| 758 |
+
"eval_samples_per_second": 18.987,
|
| 759 |
+
"eval_steps_per_second": 2.375,
|
| 760 |
+
"eval_wer": 0.357975622681505,
|
| 761 |
+
"step": 23500
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 8.534850640113799,
|
| 765 |
+
"grad_norm": 0.5965376496315002,
|
| 766 |
+
"learning_rate": 4.4847936278059375e-05,
|
| 767 |
+
"loss": 0.3471,
|
| 768 |
+
"step": 24000
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 8.534850640113799,
|
| 772 |
+
"eval_loss": 0.17967215180397034,
|
| 773 |
+
"eval_runtime": 77.691,
|
| 774 |
+
"eval_samples_per_second": 19.037,
|
| 775 |
+
"eval_steps_per_second": 2.381,
|
| 776 |
+
"eval_wer": 0.3606253312135665,
|
| 777 |
+
"step": 24000
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 8.712660028449502,
|
| 781 |
+
"grad_norm": 0.3367222249507904,
|
| 782 |
+
"learning_rate": 3.941708906589427e-05,
|
| 783 |
+
"loss": 0.3497,
|
| 784 |
+
"step": 24500
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 8.712660028449502,
|
| 788 |
+
"eval_loss": 0.17775095999240875,
|
| 789 |
+
"eval_runtime": 78.061,
|
| 790 |
+
"eval_samples_per_second": 18.947,
|
| 791 |
+
"eval_steps_per_second": 2.37,
|
| 792 |
+
"eval_wer": 0.35877053524112346,
|
| 793 |
+
"step": 24500
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 8.890469416785207,
|
| 797 |
+
"grad_norm": 1.0665998458862305,
|
| 798 |
+
"learning_rate": 3.399710354815351e-05,
|
| 799 |
+
"loss": 0.3398,
|
| 800 |
+
"step": 25000
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 8.890469416785207,
|
| 804 |
+
"eval_loss": 0.1774686574935913,
|
| 805 |
+
"eval_runtime": 77.6438,
|
| 806 |
+
"eval_samples_per_second": 19.049,
|
| 807 |
+
"eval_steps_per_second": 2.383,
|
| 808 |
+
"eval_wer": 0.3583289171524466,
|
| 809 |
+
"step": 25000
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 9.06827880512091,
|
| 813 |
+
"grad_norm": 1.8358111381530762,
|
| 814 |
+
"learning_rate": 2.8566256335988413e-05,
|
| 815 |
+
"loss": 0.3444,
|
| 816 |
+
"step": 25500
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 9.06827880512091,
|
| 820 |
+
"eval_loss": 0.1796201765537262,
|
| 821 |
+
"eval_runtime": 77.7985,
|
| 822 |
+
"eval_samples_per_second": 19.011,
|
| 823 |
+
"eval_steps_per_second": 2.378,
|
| 824 |
+
"eval_wer": 0.35859388800565273,
|
| 825 |
+
"step": 25500
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 9.246088193456615,
|
| 829 |
+
"grad_norm": 1.32257878780365,
|
| 830 |
+
"learning_rate": 2.3135409123823315e-05,
|
| 831 |
+
"loss": 0.3366,
|
| 832 |
+
"step": 26000
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 9.246088193456615,
|
| 836 |
+
"eval_loss": 0.1784891039133072,
|
| 837 |
+
"eval_runtime": 77.5756,
|
| 838 |
+
"eval_samples_per_second": 19.065,
|
| 839 |
+
"eval_steps_per_second": 2.385,
|
| 840 |
+
"eval_wer": 0.35735735735735735,
|
| 841 |
+
"step": 26000
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 9.42389758179232,
|
| 845 |
+
"grad_norm": 1.8535629510879517,
|
| 846 |
+
"learning_rate": 1.7704561911658217e-05,
|
| 847 |
+
"loss": 0.3434,
|
| 848 |
+
"step": 26500
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 9.42389758179232,
|
| 852 |
+
"eval_loss": 0.17805208265781403,
|
| 853 |
+
"eval_runtime": 77.8845,
|
| 854 |
+
"eval_samples_per_second": 18.99,
|
| 855 |
+
"eval_steps_per_second": 2.375,
|
| 856 |
+
"eval_wer": 0.3592121533298004,
|
| 857 |
+
"step": 26500
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 9.601706970128022,
|
| 861 |
+
"grad_norm": 1.9332554340362549,
|
| 862 |
+
"learning_rate": 1.228457639391745e-05,
|
| 863 |
+
"loss": 0.3426,
|
| 864 |
+
"step": 27000
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 9.601706970128022,
|
| 868 |
+
"eval_loss": 0.17857009172439575,
|
| 869 |
+
"eval_runtime": 77.3945,
|
| 870 |
+
"eval_samples_per_second": 19.11,
|
| 871 |
+
"eval_steps_per_second": 2.39,
|
| 872 |
+
"eval_wer": 0.35930047694753575,
|
| 873 |
+
"step": 27000
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 9.779516358463727,
|
| 877 |
+
"grad_norm": 1.1602191925048828,
|
| 878 |
+
"learning_rate": 6.8537291817523524e-06,
|
| 879 |
+
"loss": 0.3496,
|
| 880 |
+
"step": 27500
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 9.779516358463727,
|
| 884 |
+
"eval_loss": 0.17868547141551971,
|
| 885 |
+
"eval_runtime": 77.5791,
|
| 886 |
+
"eval_samples_per_second": 19.064,
|
| 887 |
+
"eval_steps_per_second": 2.385,
|
| 888 |
+
"eval_wer": 0.3590355060943296,
|
| 889 |
+
"step": 27500
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 9.95732574679943,
|
| 893 |
+
"grad_norm": 1.2770110368728638,
|
| 894 |
+
"learning_rate": 1.4228819695872554e-06,
|
| 895 |
+
"loss": 0.334,
|
| 896 |
+
"step": 28000
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 9.95732574679943,
|
| 900 |
+
"eval_loss": 0.17876511812210083,
|
| 901 |
+
"eval_runtime": 78.0274,
|
| 902 |
+
"eval_samples_per_second": 18.955,
|
| 903 |
+
"eval_steps_per_second": 2.371,
|
| 904 |
+
"eval_wer": 0.35877053524112346,
|
| 905 |
+
"step": 28000
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 10.0,
|
| 909 |
+
"step": 28120,
|
| 910 |
+
"total_flos": 2.1329860467760157e+19,
|
| 911 |
+
"train_loss": 0.48163046043254915,
|
| 912 |
+
"train_runtime": 15300.4436,
|
| 913 |
+
"train_samples_per_second": 7.349,
|
| 914 |
+
"train_steps_per_second": 1.838
|
| 915 |
+
}
|
| 916 |
+
],
|
| 917 |
+
"logging_steps": 500,
|
| 918 |
+
"max_steps": 28120,
|
| 919 |
+
"num_input_tokens_seen": 0,
|
| 920 |
+
"num_train_epochs": 10,
|
| 921 |
+
"save_steps": 500,
|
| 922 |
+
"stateful_callbacks": {
|
| 923 |
+
"TrainerControl": {
|
| 924 |
+
"args": {
|
| 925 |
+
"should_epoch_stop": false,
|
| 926 |
+
"should_evaluate": false,
|
| 927 |
+
"should_log": false,
|
| 928 |
+
"should_save": true,
|
| 929 |
+
"should_training_stop": true
|
| 930 |
+
},
|
| 931 |
+
"attributes": {}
|
| 932 |
+
}
|
| 933 |
+
},
|
| 934 |
+
"total_flos": 2.1329860467760157e+19,
|
| 935 |
+
"train_batch_size": 4,
|
| 936 |
+
"trial_name": null,
|
| 937 |
+
"trial_params": null
|
| 938 |
+
}
|