End of training
Browse files- README.md +4 -2
- all_results.json +16 -0
- eval_results.json +10 -0
- train_results.json +9 -0
- trainer_state.json +897 -0
README.md
CHANGED
|
@@ -3,6 +3,8 @@ library_name: transformers
|
|
| 3 |
license: apache-2.0
|
| 4 |
base_model: rinna/japanese-hubert-base
|
| 5 |
tags:
|
|
|
|
|
|
|
| 6 |
- generated_from_trainer
|
| 7 |
metrics:
|
| 8 |
- wer
|
|
@@ -16,9 +18,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
|
| 17 |
# Hubert-kakeiken-W-closed
|
| 18 |
|
| 19 |
-
This model is a fine-tuned version of [rinna/japanese-hubert-base](https://huggingface.co/rinna/japanese-hubert-base) on the
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
-
- Loss: 0.
|
| 22 |
- Wer: 0.9988
|
| 23 |
- Cer: 1.0164
|
| 24 |
|
|
|
|
| 3 |
license: apache-2.0
|
| 4 |
base_model: rinna/japanese-hubert-base
|
| 5 |
tags:
|
| 6 |
+
- automatic-speech-recognition
|
| 7 |
+
- original_kakeiken_W_closed
|
| 8 |
- generated_from_trainer
|
| 9 |
metrics:
|
| 10 |
- wer
|
|
|
|
| 18 |
|
| 19 |
# Hubert-kakeiken-W-closed
|
| 20 |
|
| 21 |
+
This model is a fine-tuned version of [rinna/japanese-hubert-base](https://huggingface.co/rinna/japanese-hubert-base) on the ORIGINAL_KAKEIKEN_W_CLOSED - JA dataset.
|
| 22 |
It achieves the following results on the evaluation set:
|
| 23 |
+
- Loss: 0.0273
|
| 24 |
- Wer: 0.9988
|
| 25 |
- Cer: 1.0164
|
| 26 |
|
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 39.95203400121433,
|
| 3 |
+
"eval_cer": 1.0164177335229967,
|
| 4 |
+
"eval_loss": 0.027329571545124054,
|
| 5 |
+
"eval_runtime": 62.1595,
|
| 6 |
+
"eval_samples": 6840,
|
| 7 |
+
"eval_samples_per_second": 110.039,
|
| 8 |
+
"eval_steps_per_second": 13.755,
|
| 9 |
+
"eval_wer": 0.9988304093567252,
|
| 10 |
+
"total_flos": 1.7147211678918107e+19,
|
| 11 |
+
"train_loss": 1.2561371190290116,
|
| 12 |
+
"train_runtime": 30821.1471,
|
| 13 |
+
"train_samples": 52680,
|
| 14 |
+
"train_samples_per_second": 68.369,
|
| 15 |
+
"train_steps_per_second": 1.068
|
| 16 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 39.95203400121433,
|
| 3 |
+
"eval_cer": 1.0164177335229967,
|
| 4 |
+
"eval_loss": 0.027329571545124054,
|
| 5 |
+
"eval_runtime": 62.1595,
|
| 6 |
+
"eval_samples": 6840,
|
| 7 |
+
"eval_samples_per_second": 110.039,
|
| 8 |
+
"eval_steps_per_second": 13.755,
|
| 9 |
+
"eval_wer": 0.9988304093567252
|
| 10 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 39.95203400121433,
|
| 3 |
+
"total_flos": 1.7147211678918107e+19,
|
| 4 |
+
"train_loss": 1.2561371190290116,
|
| 5 |
+
"train_runtime": 30821.1471,
|
| 6 |
+
"train_samples": 52680,
|
| 7 |
+
"train_samples_per_second": 68.369,
|
| 8 |
+
"train_steps_per_second": 1.068
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,897 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 39.95203400121433,
|
| 5 |
+
"eval_steps": 100.0,
|
| 6 |
+
"global_step": 32920,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.607164541590771,
|
| 13 |
+
"grad_norm": 57.35422134399414,
|
| 14 |
+
"learning_rate": 1.188e-06,
|
| 15 |
+
"loss": 28.44,
|
| 16 |
+
"step": 500
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 1.0,
|
| 20 |
+
"eval_cer": 1.1284080132764343,
|
| 21 |
+
"eval_loss": 10.8607177734375,
|
| 22 |
+
"eval_runtime": 145.137,
|
| 23 |
+
"eval_samples_per_second": 47.128,
|
| 24 |
+
"eval_steps_per_second": 5.891,
|
| 25 |
+
"eval_wer": 1.0,
|
| 26 |
+
"step": 824
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 1.2137219186399515,
|
| 30 |
+
"grad_norm": 44.843685150146484,
|
| 31 |
+
"learning_rate": 2.3880000000000003e-06,
|
| 32 |
+
"loss": 11.3,
|
| 33 |
+
"step": 1000
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 1.8208864602307226,
|
| 37 |
+
"grad_norm": 36.54290008544922,
|
| 38 |
+
"learning_rate": 3.588e-06,
|
| 39 |
+
"loss": 9.2184,
|
| 40 |
+
"step": 1500
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 2.0,
|
| 44 |
+
"eval_cer": 1.1284376481744902,
|
| 45 |
+
"eval_loss": 7.510649681091309,
|
| 46 |
+
"eval_runtime": 131.1044,
|
| 47 |
+
"eval_samples_per_second": 52.172,
|
| 48 |
+
"eval_steps_per_second": 6.522,
|
| 49 |
+
"eval_wer": 1.0,
|
| 50 |
+
"step": 1648
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 2.427443837279903,
|
| 54 |
+
"grad_norm": 22.59714126586914,
|
| 55 |
+
"learning_rate": 4.788e-06,
|
| 56 |
+
"loss": 7.0297,
|
| 57 |
+
"step": 2000
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 3.0,
|
| 61 |
+
"eval_cer": 1.1283783783783783,
|
| 62 |
+
"eval_loss": 4.190179347991943,
|
| 63 |
+
"eval_runtime": 131.4162,
|
| 64 |
+
"eval_samples_per_second": 52.048,
|
| 65 |
+
"eval_steps_per_second": 6.506,
|
| 66 |
+
"eval_wer": 1.0,
|
| 67 |
+
"step": 2472
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 3.0340012143290833,
|
| 71 |
+
"grad_norm": 10.585729598999023,
|
| 72 |
+
"learning_rate": 5.988e-06,
|
| 73 |
+
"loss": 4.8903,
|
| 74 |
+
"step": 2500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 3.6411657559198543,
|
| 78 |
+
"grad_norm": 2.5189406871795654,
|
| 79 |
+
"learning_rate": 7.1880000000000005e-06,
|
| 80 |
+
"loss": 3.6874,
|
| 81 |
+
"step": 3000
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 4.0,
|
| 85 |
+
"eval_cer": 1.1283783783783783,
|
| 86 |
+
"eval_loss": 3.1195626258850098,
|
| 87 |
+
"eval_runtime": 129.8504,
|
| 88 |
+
"eval_samples_per_second": 52.676,
|
| 89 |
+
"eval_steps_per_second": 6.584,
|
| 90 |
+
"eval_wer": 1.0,
|
| 91 |
+
"step": 3296
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 4.247723132969035,
|
| 95 |
+
"grad_norm": 1.423240303993225,
|
| 96 |
+
"learning_rate": 8.388e-06,
|
| 97 |
+
"loss": 3.1803,
|
| 98 |
+
"step": 3500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 4.854887674559806,
|
| 102 |
+
"grad_norm": 1.8786451816558838,
|
| 103 |
+
"learning_rate": 9.588e-06,
|
| 104 |
+
"loss": 2.7259,
|
| 105 |
+
"step": 4000
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 5.0,
|
| 109 |
+
"eval_cer": 1.1283783783783783,
|
| 110 |
+
"eval_loss": 2.2997305393218994,
|
| 111 |
+
"eval_runtime": 128.0868,
|
| 112 |
+
"eval_samples_per_second": 53.401,
|
| 113 |
+
"eval_steps_per_second": 6.675,
|
| 114 |
+
"eval_wer": 1.0,
|
| 115 |
+
"step": 4120
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 5.461445051608986,
|
| 119 |
+
"grad_norm": 3.4784176349639893,
|
| 120 |
+
"learning_rate": 1.0787999999999999e-05,
|
| 121 |
+
"loss": 2.1431,
|
| 122 |
+
"step": 4500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 6.0,
|
| 126 |
+
"eval_cer": 1.0313537221431959,
|
| 127 |
+
"eval_loss": 1.0496456623077393,
|
| 128 |
+
"eval_runtime": 138.552,
|
| 129 |
+
"eval_samples_per_second": 49.368,
|
| 130 |
+
"eval_steps_per_second": 6.171,
|
| 131 |
+
"eval_wer": 0.9997076023391813,
|
| 132 |
+
"step": 4944
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 6.068002428658167,
|
| 136 |
+
"grad_norm": 2.7191321849823,
|
| 137 |
+
"learning_rate": 1.1988000000000001e-05,
|
| 138 |
+
"loss": 1.3807,
|
| 139 |
+
"step": 5000
|
| 140 |
+
},
|
| 141 |
+
{
|
| 142 |
+
"epoch": 6.675166970248937,
|
| 143 |
+
"grad_norm": 2.5248982906341553,
|
| 144 |
+
"learning_rate": 1.3188e-05,
|
| 145 |
+
"loss": 0.8891,
|
| 146 |
+
"step": 5500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 7.0,
|
| 150 |
+
"eval_cer": 1.017425320056899,
|
| 151 |
+
"eval_loss": 0.6379755735397339,
|
| 152 |
+
"eval_runtime": 135.841,
|
| 153 |
+
"eval_samples_per_second": 50.353,
|
| 154 |
+
"eval_steps_per_second": 6.294,
|
| 155 |
+
"eval_wer": 0.9998538011695907,
|
| 156 |
+
"step": 5768
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 7.281724347298118,
|
| 160 |
+
"grad_norm": 7.416316509246826,
|
| 161 |
+
"learning_rate": 1.4388000000000002e-05,
|
| 162 |
+
"loss": 0.6223,
|
| 163 |
+
"step": 6000
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 7.888888888888889,
|
| 167 |
+
"grad_norm": 6.211233615875244,
|
| 168 |
+
"learning_rate": 1.5588e-05,
|
| 169 |
+
"loss": 0.4891,
|
| 170 |
+
"step": 6500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 8.0,
|
| 174 |
+
"eval_cer": 1.0305239449976291,
|
| 175 |
+
"eval_loss": 0.25890296697616577,
|
| 176 |
+
"eval_runtime": 76.2932,
|
| 177 |
+
"eval_samples_per_second": 89.654,
|
| 178 |
+
"eval_steps_per_second": 11.207,
|
| 179 |
+
"eval_wer": 0.9991228070175439,
|
| 180 |
+
"step": 6592
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 8.49544626593807,
|
| 184 |
+
"grad_norm": 2.76125431060791,
|
| 185 |
+
"learning_rate": 1.6788e-05,
|
| 186 |
+
"loss": 0.3675,
|
| 187 |
+
"step": 7000
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 9.0,
|
| 191 |
+
"eval_cer": 1.0932313892840209,
|
| 192 |
+
"eval_loss": 0.6099491119384766,
|
| 193 |
+
"eval_runtime": 74.5927,
|
| 194 |
+
"eval_samples_per_second": 91.698,
|
| 195 |
+
"eval_steps_per_second": 11.462,
|
| 196 |
+
"eval_wer": 0.9994152046783625,
|
| 197 |
+
"step": 7416
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 9.102003642987249,
|
| 201 |
+
"grad_norm": 8.747713088989258,
|
| 202 |
+
"learning_rate": 1.7988e-05,
|
| 203 |
+
"loss": 0.3164,
|
| 204 |
+
"step": 7500
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 9.70916818457802,
|
| 208 |
+
"grad_norm": 7.222991943359375,
|
| 209 |
+
"learning_rate": 1.9188e-05,
|
| 210 |
+
"loss": 0.2744,
|
| 211 |
+
"step": 8000
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 10.0,
|
| 215 |
+
"eval_cer": 1.023944997629208,
|
| 216 |
+
"eval_loss": 0.1262398213148117,
|
| 217 |
+
"eval_runtime": 62.8915,
|
| 218 |
+
"eval_samples_per_second": 108.759,
|
| 219 |
+
"eval_steps_per_second": 13.595,
|
| 220 |
+
"eval_wer": 0.9989766081871345,
|
| 221 |
+
"step": 8240
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 10.3157255616272,
|
| 225 |
+
"grad_norm": 7.39332389831543,
|
| 226 |
+
"learning_rate": 2.0388e-05,
|
| 227 |
+
"loss": 0.2525,
|
| 228 |
+
"step": 8500
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 10.922890103217972,
|
| 232 |
+
"grad_norm": 3.6922662258148193,
|
| 233 |
+
"learning_rate": 2.1588e-05,
|
| 234 |
+
"loss": 0.2278,
|
| 235 |
+
"step": 9000
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 11.0,
|
| 239 |
+
"eval_cer": 1.023411569464201,
|
| 240 |
+
"eval_loss": 0.11068873107433319,
|
| 241 |
+
"eval_runtime": 67.9091,
|
| 242 |
+
"eval_samples_per_second": 100.723,
|
| 243 |
+
"eval_steps_per_second": 12.59,
|
| 244 |
+
"eval_wer": 0.9989766081871345,
|
| 245 |
+
"step": 9064
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 11.529447480267152,
|
| 249 |
+
"grad_norm": 5.623218536376953,
|
| 250 |
+
"learning_rate": 2.2788000000000003e-05,
|
| 251 |
+
"loss": 0.2148,
|
| 252 |
+
"step": 9500
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 12.0,
|
| 256 |
+
"eval_cer": 1.0260787102892366,
|
| 257 |
+
"eval_loss": 0.06716426461935043,
|
| 258 |
+
"eval_runtime": 89.5293,
|
| 259 |
+
"eval_samples_per_second": 76.4,
|
| 260 |
+
"eval_steps_per_second": 9.55,
|
| 261 |
+
"eval_wer": 0.9989766081871345,
|
| 262 |
+
"step": 9888
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 12.136004857316333,
|
| 266 |
+
"grad_norm": 4.379143714904785,
|
| 267 |
+
"learning_rate": 2.3988e-05,
|
| 268 |
+
"loss": 0.2055,
|
| 269 |
+
"step": 10000
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 12.743169398907105,
|
| 273 |
+
"grad_norm": 5.908865451812744,
|
| 274 |
+
"learning_rate": 2.5188e-05,
|
| 275 |
+
"loss": 0.1927,
|
| 276 |
+
"step": 10500
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 13.0,
|
| 280 |
+
"eval_cer": 1.0219298245614035,
|
| 281 |
+
"eval_loss": 0.059190813452005386,
|
| 282 |
+
"eval_runtime": 80.5762,
|
| 283 |
+
"eval_samples_per_second": 84.889,
|
| 284 |
+
"eval_steps_per_second": 10.611,
|
| 285 |
+
"eval_wer": 0.9989766081871345,
|
| 286 |
+
"step": 10712
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 13.349726775956285,
|
| 290 |
+
"grad_norm": 3.240206003189087,
|
| 291 |
+
"learning_rate": 2.63856e-05,
|
| 292 |
+
"loss": 0.1926,
|
| 293 |
+
"step": 11000
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 13.956891317547056,
|
| 297 |
+
"grad_norm": 5.876375675201416,
|
| 298 |
+
"learning_rate": 2.7585600000000002e-05,
|
| 299 |
+
"loss": 0.1723,
|
| 300 |
+
"step": 11500
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 14.0,
|
| 304 |
+
"eval_cer": 1.022848506401138,
|
| 305 |
+
"eval_loss": 0.08003176748752594,
|
| 306 |
+
"eval_runtime": 62.2149,
|
| 307 |
+
"eval_samples_per_second": 109.941,
|
| 308 |
+
"eval_steps_per_second": 13.743,
|
| 309 |
+
"eval_wer": 0.9988304093567252,
|
| 310 |
+
"step": 11536
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 14.563448694596236,
|
| 314 |
+
"grad_norm": 3.714435577392578,
|
| 315 |
+
"learning_rate": 2.87856e-05,
|
| 316 |
+
"loss": 0.1725,
|
| 317 |
+
"step": 12000
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 15.0,
|
| 321 |
+
"eval_cer": 1.021100047415837,
|
| 322 |
+
"eval_loss": 0.051837269216775894,
|
| 323 |
+
"eval_runtime": 62.5224,
|
| 324 |
+
"eval_samples_per_second": 109.401,
|
| 325 |
+
"eval_steps_per_second": 13.675,
|
| 326 |
+
"eval_wer": 0.9989766081871345,
|
| 327 |
+
"step": 12360
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 15.170006071645416,
|
| 331 |
+
"grad_norm": 9.124956130981445,
|
| 332 |
+
"learning_rate": 2.99856e-05,
|
| 333 |
+
"loss": 0.1695,
|
| 334 |
+
"step": 12500
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 15.777170613236187,
|
| 338 |
+
"grad_norm": 3.6504106521606445,
|
| 339 |
+
"learning_rate": 2.9956874399074467e-05,
|
| 340 |
+
"loss": 0.1628,
|
| 341 |
+
"step": 13000
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 16.0,
|
| 345 |
+
"eval_cer": 1.0142247510668563,
|
| 346 |
+
"eval_loss": 0.13594070076942444,
|
| 347 |
+
"eval_runtime": 75.4469,
|
| 348 |
+
"eval_samples_per_second": 90.66,
|
| 349 |
+
"eval_steps_per_second": 11.332,
|
| 350 |
+
"eval_wer": 0.9988304093567252,
|
| 351 |
+
"step": 13184
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 16.38372799028537,
|
| 355 |
+
"grad_norm": 3.7022886276245117,
|
| 356 |
+
"learning_rate": 2.9825295862461663e-05,
|
| 357 |
+
"loss": 0.1626,
|
| 358 |
+
"step": 13500
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"epoch": 16.99089253187614,
|
| 362 |
+
"grad_norm": 1.3335272073745728,
|
| 363 |
+
"learning_rate": 2.9606033905859603e-05,
|
| 364 |
+
"loss": 0.1567,
|
| 365 |
+
"step": 14000
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 17.0,
|
| 369 |
+
"eval_cer": 1.0195886676149835,
|
| 370 |
+
"eval_loss": 0.04436279088258743,
|
| 371 |
+
"eval_runtime": 64.8228,
|
| 372 |
+
"eval_samples_per_second": 105.518,
|
| 373 |
+
"eval_steps_per_second": 13.19,
|
| 374 |
+
"eval_wer": 0.9989766081871345,
|
| 375 |
+
"step": 14008
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 17.59744990892532,
|
| 379 |
+
"grad_norm": 2.4595658779144287,
|
| 380 |
+
"learning_rate": 2.9300385342391396e-05,
|
| 381 |
+
"loss": 0.1436,
|
| 382 |
+
"step": 14500
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 18.0,
|
| 386 |
+
"eval_cer": 1.0192626837363679,
|
| 387 |
+
"eval_loss": 0.04214438423514366,
|
| 388 |
+
"eval_runtime": 63.292,
|
| 389 |
+
"eval_samples_per_second": 108.07,
|
| 390 |
+
"eval_steps_per_second": 13.509,
|
| 391 |
+
"eval_wer": 0.9988304093567252,
|
| 392 |
+
"step": 14832
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 18.204007285974498,
|
| 396 |
+
"grad_norm": 0.9890690445899963,
|
| 397 |
+
"learning_rate": 2.891015791414923e-05,
|
| 398 |
+
"loss": 0.1495,
|
| 399 |
+
"step": 15000
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 18.81117182756527,
|
| 403 |
+
"grad_norm": 8.133474349975586,
|
| 404 |
+
"learning_rate": 2.843765960040039e-05,
|
| 405 |
+
"loss": 0.1351,
|
| 406 |
+
"step": 15500
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 19.0,
|
| 410 |
+
"eval_cer": 1.0173067804646752,
|
| 411 |
+
"eval_loss": 0.03748102858662605,
|
| 412 |
+
"eval_runtime": 63.0874,
|
| 413 |
+
"eval_samples_per_second": 108.421,
|
| 414 |
+
"eval_steps_per_second": 13.553,
|
| 415 |
+
"eval_wer": 0.9988304093567252,
|
| 416 |
+
"step": 15656
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 19.41772920461445,
|
| 420 |
+
"grad_norm": 5.602595806121826,
|
| 421 |
+
"learning_rate": 2.7885684967167233e-05,
|
| 422 |
+
"loss": 0.1454,
|
| 423 |
+
"step": 16000
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 20.0,
|
| 427 |
+
"eval_cer": 1.018640350877193,
|
| 428 |
+
"eval_loss": 0.03039967454969883,
|
| 429 |
+
"eval_runtime": 77.538,
|
| 430 |
+
"eval_samples_per_second": 88.215,
|
| 431 |
+
"eval_steps_per_second": 11.027,
|
| 432 |
+
"eval_wer": 0.9988304093567252,
|
| 433 |
+
"step": 16480
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 20.02428658166363,
|
| 437 |
+
"grad_norm": 2.98942494392395,
|
| 438 |
+
"learning_rate": 2.7257498638915816e-05,
|
| 439 |
+
"loss": 0.1353,
|
| 440 |
+
"step": 16500
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 20.6314511232544,
|
| 444 |
+
"grad_norm": 6.270744800567627,
|
| 445 |
+
"learning_rate": 2.6558287021276313e-05,
|
| 446 |
+
"loss": 0.1252,
|
| 447 |
+
"step": 17000
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 21.0,
|
| 451 |
+
"eval_cer": 1.0237079184447606,
|
| 452 |
+
"eval_loss": 0.05677889287471771,
|
| 453 |
+
"eval_runtime": 62.9546,
|
| 454 |
+
"eval_samples_per_second": 108.65,
|
| 455 |
+
"eval_steps_per_second": 13.581,
|
| 456 |
+
"eval_wer": 0.9988304093567252,
|
| 457 |
+
"step": 17304
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 21.238008500303582,
|
| 461 |
+
"grad_norm": 3.5701606273651123,
|
| 462 |
+
"learning_rate": 2.578938449744228e-05,
|
| 463 |
+
"loss": 0.1249,
|
| 464 |
+
"step": 17500
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 21.845173041894352,
|
| 468 |
+
"grad_norm": 6.276644706726074,
|
| 469 |
+
"learning_rate": 2.4956668735674143e-05,
|
| 470 |
+
"loss": 0.1233,
|
| 471 |
+
"step": 18000
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 22.0,
|
| 475 |
+
"eval_cer": 1.0176031294452348,
|
| 476 |
+
"eval_loss": 0.029143376275897026,
|
| 477 |
+
"eval_runtime": 63.804,
|
| 478 |
+
"eval_samples_per_second": 107.203,
|
| 479 |
+
"eval_steps_per_second": 13.4,
|
| 480 |
+
"eval_wer": 0.9988304093567252,
|
| 481 |
+
"step": 18128
|
| 482 |
+
},
|
| 483 |
+
{
|
| 484 |
+
"epoch": 22.451730418943534,
|
| 485 |
+
"grad_norm": 3.3094053268432617,
|
| 486 |
+
"learning_rate": 2.40650647888375e-05,
|
| 487 |
+
"loss": 0.1179,
|
| 488 |
+
"step": 18500
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 23.0,
|
| 492 |
+
"eval_cer": 1.016743717401612,
|
| 493 |
+
"eval_loss": 0.02712642401456833,
|
| 494 |
+
"eval_runtime": 63.365,
|
| 495 |
+
"eval_samples_per_second": 107.946,
|
| 496 |
+
"eval_steps_per_second": 13.493,
|
| 497 |
+
"eval_wer": 0.9988304093567252,
|
| 498 |
+
"step": 18952
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 23.058287795992715,
|
| 502 |
+
"grad_norm": 4.888893127441406,
|
| 503 |
+
"learning_rate": 2.3123726366487132e-05,
|
| 504 |
+
"loss": 0.1141,
|
| 505 |
+
"step": 19000
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 23.665452337583485,
|
| 509 |
+
"grad_norm": 1.941120982170105,
|
| 510 |
+
"learning_rate": 2.2130663756909194e-05,
|
| 511 |
+
"loss": 0.1108,
|
| 512 |
+
"step": 19500
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 24.0,
|
| 516 |
+
"eval_cer": 1.0178402086296823,
|
| 517 |
+
"eval_loss": 0.027879294008016586,
|
| 518 |
+
"eval_runtime": 65.3857,
|
| 519 |
+
"eval_samples_per_second": 104.61,
|
| 520 |
+
"eval_steps_per_second": 13.076,
|
| 521 |
+
"eval_wer": 0.9988304093567252,
|
| 522 |
+
"step": 19776
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 24.272009714632667,
|
| 526 |
+
"grad_norm": 1.846130132675171,
|
| 527 |
+
"learning_rate": 2.1095427217664034e-05,
|
| 528 |
+
"loss": 0.1089,
|
| 529 |
+
"step": 20000
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 24.879174256223436,
|
| 533 |
+
"grad_norm": 3.097496271133423,
|
| 534 |
+
"learning_rate": 2.002413959993121e-05,
|
| 535 |
+
"loss": 0.1031,
|
| 536 |
+
"step": 20500
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 25.0,
|
| 540 |
+
"eval_cer": 1.0182847321005215,
|
| 541 |
+
"eval_loss": 0.03511003032326698,
|
| 542 |
+
"eval_runtime": 65.1106,
|
| 543 |
+
"eval_samples_per_second": 105.052,
|
| 544 |
+
"eval_steps_per_second": 13.132,
|
| 545 |
+
"eval_wer": 0.9989766081871345,
|
| 546 |
+
"step": 20600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 25.485731633272618,
|
| 550 |
+
"grad_norm": 5.548646926879883,
|
| 551 |
+
"learning_rate": 1.8923136977067138e-05,
|
| 552 |
+
"loss": 0.1006,
|
| 553 |
+
"step": 21000
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 26.0,
|
| 557 |
+
"eval_cer": 1.0173364153627311,
|
| 558 |
+
"eval_loss": 0.04413418844342232,
|
| 559 |
+
"eval_runtime": 62.3139,
|
| 560 |
+
"eval_samples_per_second": 109.767,
|
| 561 |
+
"eval_steps_per_second": 13.721,
|
| 562 |
+
"eval_wer": 0.9989766081871345,
|
| 563 |
+
"step": 21424
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 26.092289010321796,
|
| 567 |
+
"grad_norm": 1.0187814235687256,
|
| 568 |
+
"learning_rate": 1.779893117023784e-05,
|
| 569 |
+
"loss": 0.1032,
|
| 570 |
+
"step": 21500
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 26.69945355191257,
|
| 574 |
+
"grad_norm": 1.9439737796783447,
|
| 575 |
+
"learning_rate": 1.665817123460074e-05,
|
| 576 |
+
"loss": 0.0946,
|
| 577 |
+
"step": 22000
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 27.0,
|
| 581 |
+
"eval_cer": 1.0169511616880038,
|
| 582 |
+
"eval_loss": 0.03058658167719841,
|
| 583 |
+
"eval_runtime": 64.6664,
|
| 584 |
+
"eval_samples_per_second": 105.774,
|
| 585 |
+
"eval_steps_per_second": 13.222,
|
| 586 |
+
"eval_wer": 0.9988304093567252,
|
| 587 |
+
"step": 22248
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 27.306010928961747,
|
| 591 |
+
"grad_norm": 1.2627675533294678,
|
| 592 |
+
"learning_rate": 1.55076041338233e-05,
|
| 593 |
+
"loss": 0.0936,
|
| 594 |
+
"step": 22500
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 27.91317547055252,
|
| 598 |
+
"grad_norm": 0.8128781318664551,
|
| 599 |
+
"learning_rate": 1.4354034835527018e-05,
|
| 600 |
+
"loss": 0.09,
|
| 601 |
+
"step": 23000
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 28.0,
|
| 605 |
+
"eval_cer": 1.0176920341394025,
|
| 606 |
+
"eval_loss": 0.030218515545129776,
|
| 607 |
+
"eval_runtime": 64.0242,
|
| 608 |
+
"eval_samples_per_second": 106.835,
|
| 609 |
+
"eval_steps_per_second": 13.354,
|
| 610 |
+
"eval_wer": 0.9988304093567252,
|
| 611 |
+
"step": 23072
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 28.5197328476017,
|
| 615 |
+
"grad_norm": 4.884032249450684,
|
| 616 |
+
"learning_rate": 1.3206577220714804e-05,
|
| 617 |
+
"loss": 0.0813,
|
| 618 |
+
"step": 23500
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 29.0,
|
| 622 |
+
"eval_cer": 1.0178402086296823,
|
| 623 |
+
"eval_loss": 0.03485483676195145,
|
| 624 |
+
"eval_runtime": 62.0604,
|
| 625 |
+
"eval_samples_per_second": 110.215,
|
| 626 |
+
"eval_steps_per_second": 13.777,
|
| 627 |
+
"eval_wer": 0.9988304093567252,
|
| 628 |
+
"step": 23896
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 29.12629022465088,
|
| 632 |
+
"grad_norm": 0.34174931049346924,
|
| 633 |
+
"learning_rate": 1.2067421110204709e-05,
|
| 634 |
+
"loss": 0.0844,
|
| 635 |
+
"step": 24000
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 29.73345476624165,
|
| 639 |
+
"grad_norm": 0.053012751042842865,
|
| 640 |
+
"learning_rate": 1.0945609580796467e-05,
|
| 641 |
+
"loss": 0.0806,
|
| 642 |
+
"step": 24500
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 30.0,
|
| 646 |
+
"eval_cer": 1.0178698435277382,
|
| 647 |
+
"eval_loss": 0.03685862198472023,
|
| 648 |
+
"eval_runtime": 66.7781,
|
| 649 |
+
"eval_samples_per_second": 102.429,
|
| 650 |
+
"eval_steps_per_second": 12.804,
|
| 651 |
+
"eval_wer": 0.9988304093567252,
|
| 652 |
+
"step": 24720
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 30.34001214329083,
|
| 656 |
+
"grad_norm": 3.6521289348602295,
|
| 657 |
+
"learning_rate": 9.847777526821669e-06,
|
| 658 |
+
"loss": 0.0758,
|
| 659 |
+
"step": 25000
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 30.947176684881605,
|
| 663 |
+
"grad_norm": 2.7911853790283203,
|
| 664 |
+
"learning_rate": 8.780418017286117e-06,
|
| 665 |
+
"loss": 0.0763,
|
| 666 |
+
"step": 25500
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 31.0,
|
| 670 |
+
"eval_cer": 1.0164770033191086,
|
| 671 |
+
"eval_loss": 0.04343733936548233,
|
| 672 |
+
"eval_runtime": 63.0385,
|
| 673 |
+
"eval_samples_per_second": 108.505,
|
| 674 |
+
"eval_steps_per_second": 13.563,
|
| 675 |
+
"eval_wer": 0.9989766081871345,
|
| 676 |
+
"step": 25544
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 31.553734061930783,
|
| 680 |
+
"grad_norm": 1.7392168045043945,
|
| 681 |
+
"learning_rate": 7.749843892960228e-06,
|
| 682 |
+
"loss": 0.075,
|
| 683 |
+
"step": 26000
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 32.0,
|
| 687 |
+
"eval_cer": 1.0160028449502134,
|
| 688 |
+
"eval_loss": 0.030797116458415985,
|
| 689 |
+
"eval_runtime": 63.3897,
|
| 690 |
+
"eval_samples_per_second": 107.904,
|
| 691 |
+
"eval_steps_per_second": 13.488,
|
| 692 |
+
"eval_wer": 0.9988304093567252,
|
| 693 |
+
"step": 26368
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 32.16029143897996,
|
| 697 |
+
"grad_norm": 3.0110056400299072,
|
| 698 |
+
"learning_rate": 6.764079092952775e-06,
|
| 699 |
+
"loss": 0.0703,
|
| 700 |
+
"step": 26500
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 32.76745598057074,
|
| 704 |
+
"grad_norm": 1.4404278993606567,
|
| 705 |
+
"learning_rate": 5.8250048617236015e-06,
|
| 706 |
+
"loss": 0.0708,
|
| 707 |
+
"step": 27000
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 33.0,
|
| 711 |
+
"eval_cer": 1.0166844476055001,
|
| 712 |
+
"eval_loss": 0.030816324055194855,
|
| 713 |
+
"eval_runtime": 62.6517,
|
| 714 |
+
"eval_samples_per_second": 109.175,
|
| 715 |
+
"eval_steps_per_second": 13.647,
|
| 716 |
+
"eval_wer": 0.9988304093567252,
|
| 717 |
+
"step": 27192
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 33.374013357619916,
|
| 721 |
+
"grad_norm": 0.1344936639070511,
|
| 722 |
+
"learning_rate": 4.940195648850366e-06,
|
| 723 |
+
"loss": 0.0684,
|
| 724 |
+
"step": 27500
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 33.981177899210685,
|
| 728 |
+
"grad_norm": 4.214531421661377,
|
| 729 |
+
"learning_rate": 4.114884611130932e-06,
|
| 730 |
+
"loss": 0.0668,
|
| 731 |
+
"step": 28000
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 34.0,
|
| 735 |
+
"eval_cer": 1.0165659080132765,
|
| 736 |
+
"eval_loss": 0.02984553575515747,
|
| 737 |
+
"eval_runtime": 63.6993,
|
| 738 |
+
"eval_samples_per_second": 107.38,
|
| 739 |
+
"eval_steps_per_second": 13.422,
|
| 740 |
+
"eval_wer": 0.9988304093567252,
|
| 741 |
+
"step": 28016
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 34.58773527625986,
|
| 745 |
+
"grad_norm": 4.261690139770508,
|
| 746 |
+
"learning_rate": 3.353953006586277e-06,
|
| 747 |
+
"loss": 0.0639,
|
| 748 |
+
"step": 28500
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 35.0,
|
| 752 |
+
"eval_cer": 1.0163880986249407,
|
| 753 |
+
"eval_loss": 0.02722448669373989,
|
| 754 |
+
"eval_runtime": 64.9453,
|
| 755 |
+
"eval_samples_per_second": 105.319,
|
| 756 |
+
"eval_steps_per_second": 13.165,
|
| 757 |
+
"eval_wer": 0.9988304093567252,
|
| 758 |
+
"step": 28840
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 35.19429265330905,
|
| 762 |
+
"grad_norm": 0.7786476016044617,
|
| 763 |
+
"learning_rate": 2.6619013245208524e-06,
|
| 764 |
+
"loss": 0.0622,
|
| 765 |
+
"step": 29000
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 35.80145719489982,
|
| 769 |
+
"grad_norm": 0.006692953407764435,
|
| 770 |
+
"learning_rate": 2.0439854900570527e-06,
|
| 771 |
+
"loss": 0.0628,
|
| 772 |
+
"step": 29500
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 36.0,
|
| 776 |
+
"eval_cer": 1.0160621147463254,
|
| 777 |
+
"eval_loss": 0.02650887332856655,
|
| 778 |
+
"eval_runtime": 64.4273,
|
| 779 |
+
"eval_samples_per_second": 106.166,
|
| 780 |
+
"eval_steps_per_second": 13.271,
|
| 781 |
+
"eval_wer": 0.9988304093567252,
|
| 782 |
+
"step": 29664
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 36.408014571948996,
|
| 786 |
+
"grad_norm": 2.9724912643432617,
|
| 787 |
+
"learning_rate": 1.501384740615621e-06,
|
| 788 |
+
"loss": 0.0628,
|
| 789 |
+
"step": 30000
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 37.0,
|
| 793 |
+
"eval_cer": 1.0163288288288288,
|
| 794 |
+
"eval_loss": 0.026661457493901253,
|
| 795 |
+
"eval_runtime": 63.7838,
|
| 796 |
+
"eval_samples_per_second": 107.237,
|
| 797 |
+
"eval_steps_per_second": 13.405,
|
| 798 |
+
"eval_wer": 0.9988304093567252,
|
| 799 |
+
"step": 30488
|
| 800 |
+
},
|
| 801 |
+
{
|
| 802 |
+
"epoch": 37.01457194899818,
|
| 803 |
+
"grad_norm": 5.568892478942871,
|
| 804 |
+
"learning_rate": 1.0386208296455812e-06,
|
| 805 |
+
"loss": 0.0618,
|
| 806 |
+
"step": 30500
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 37.62173649058895,
|
| 810 |
+
"grad_norm": 2.4032373428344727,
|
| 811 |
+
"learning_rate": 6.584307495643449e-07,
|
| 812 |
+
"loss": 0.0586,
|
| 813 |
+
"step": 31000
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 38.0,
|
| 817 |
+
"eval_cer": 1.016151019440493,
|
| 818 |
+
"eval_loss": 0.02632048726081848,
|
| 819 |
+
"eval_runtime": 63.4705,
|
| 820 |
+
"eval_samples_per_second": 107.767,
|
| 821 |
+
"eval_steps_per_second": 13.471,
|
| 822 |
+
"eval_wer": 0.9988304093567252,
|
| 823 |
+
"step": 31312
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 38.22829386763813,
|
| 827 |
+
"grad_norm": 0.8156293630599976,
|
| 828 |
+
"learning_rate": 3.6306311427998064e-07,
|
| 829 |
+
"loss": 0.0599,
|
| 830 |
+
"step": 31500
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 38.8354584092289,
|
| 834 |
+
"grad_norm": 6.638393878936768,
|
| 835 |
+
"learning_rate": 1.5426485988442763e-07,
|
| 836 |
+
"loss": 0.058,
|
| 837 |
+
"step": 32000
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 39.0,
|
| 841 |
+
"eval_cer": 1.0164473684210527,
|
| 842 |
+
"eval_loss": 0.028048371896147728,
|
| 843 |
+
"eval_runtime": 62.7293,
|
| 844 |
+
"eval_samples_per_second": 109.04,
|
| 845 |
+
"eval_steps_per_second": 13.63,
|
| 846 |
+
"eval_wer": 0.9988304093567252,
|
| 847 |
+
"step": 32136
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"epoch": 39.442015786278084,
|
| 851 |
+
"grad_norm": 0.12612353265285492,
|
| 852 |
+
"learning_rate": 3.327091249336667e-08,
|
| 853 |
+
"loss": 0.0588,
|
| 854 |
+
"step": 32500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 39.95203400121433,
|
| 858 |
+
"eval_cer": 1.0163584637268848,
|
| 859 |
+
"eval_loss": 0.02687981352210045,
|
| 860 |
+
"eval_runtime": 63.6993,
|
| 861 |
+
"eval_samples_per_second": 107.38,
|
| 862 |
+
"eval_steps_per_second": 13.422,
|
| 863 |
+
"eval_wer": 0.9988304093567252,
|
| 864 |
+
"step": 32920
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 39.95203400121433,
|
| 868 |
+
"step": 32920,
|
| 869 |
+
"total_flos": 1.7147211678918107e+19,
|
| 870 |
+
"train_loss": 1.2561371190290116,
|
| 871 |
+
"train_runtime": 30821.1471,
|
| 872 |
+
"train_samples_per_second": 68.369,
|
| 873 |
+
"train_steps_per_second": 1.068
|
| 874 |
+
}
|
| 875 |
+
],
|
| 876 |
+
"logging_steps": 500,
|
| 877 |
+
"max_steps": 32920,
|
| 878 |
+
"num_input_tokens_seen": 0,
|
| 879 |
+
"num_train_epochs": 40,
|
| 880 |
+
"save_steps": 400,
|
| 881 |
+
"stateful_callbacks": {
|
| 882 |
+
"TrainerControl": {
|
| 883 |
+
"args": {
|
| 884 |
+
"should_epoch_stop": false,
|
| 885 |
+
"should_evaluate": false,
|
| 886 |
+
"should_log": false,
|
| 887 |
+
"should_save": true,
|
| 888 |
+
"should_training_stop": true
|
| 889 |
+
},
|
| 890 |
+
"attributes": {}
|
| 891 |
+
}
|
| 892 |
+
},
|
| 893 |
+
"total_flos": 1.7147211678918107e+19,
|
| 894 |
+
"train_batch_size": 32,
|
| 895 |
+
"trial_name": null,
|
| 896 |
+
"trial_params": null
|
| 897 |
+
}
|