End of training
Browse files- README.md +8 -4
- all_results.json +15 -0
- eval_results.json +10 -0
- train_results.json +8 -0
- trainer_state.json +1043 -0
README.md
CHANGED
|
@@ -1,5 +1,9 @@
|
|
| 1 |
---
|
|
|
|
|
|
|
| 2 |
tags:
|
|
|
|
|
|
|
| 3 |
- generated_from_trainer
|
| 4 |
metrics:
|
| 5 |
- wer
|
|
@@ -13,11 +17,11 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 13 |
|
| 14 |
# mms-common_voice_13_0-eo-1
|
| 15 |
|
| 16 |
-
This model is a fine-tuned version of [patrickvonplaten/mms-300m](https://huggingface.co/patrickvonplaten/mms-300m) on the
|
| 17 |
It achieves the following results on the evaluation set:
|
| 18 |
-
- Loss: 0.
|
| 19 |
-
- Cer: 0.
|
| 20 |
-
- Wer: 0.
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
|
|
| 1 |
---
|
| 2 |
+
language:
|
| 3 |
+
- eo
|
| 4 |
tags:
|
| 5 |
+
- automatic-speech-recognition
|
| 6 |
+
- mozilla-foundation/common_voice_13_0
|
| 7 |
- generated_from_trainer
|
| 8 |
metrics:
|
| 9 |
- wer
|
|
|
|
| 17 |
|
| 18 |
# mms-common_voice_13_0-eo-1
|
| 19 |
|
| 20 |
+
This model is a fine-tuned version of [patrickvonplaten/mms-300m](https://huggingface.co/patrickvonplaten/mms-300m) on the MOZILLA-FOUNDATION/COMMON_VOICE_13_0 - EO dataset.
|
| 21 |
It achieves the following results on the evaluation set:
|
| 22 |
+
- Loss: 0.2257
|
| 23 |
+
- Cer: 0.0209
|
| 24 |
+
- Wer: 0.0678
|
| 25 |
|
| 26 |
## Model description
|
| 27 |
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 99.96,
|
| 3 |
+
"eval_cer": 0.020881859054222414,
|
| 4 |
+
"eval_loss": 0.22571781277656555,
|
| 5 |
+
"eval_runtime": 70.8237,
|
| 6 |
+
"eval_samples": 1500,
|
| 7 |
+
"eval_samples_per_second": 21.179,
|
| 8 |
+
"eval_steps_per_second": 2.654,
|
| 9 |
+
"eval_wer": 0.06779661016949153,
|
| 10 |
+
"train_loss": 0.005932311532843826,
|
| 11 |
+
"train_runtime": 23991.1474,
|
| 12 |
+
"train_samples": 15000,
|
| 13 |
+
"train_samples_per_second": 62.523,
|
| 14 |
+
"train_steps_per_second": 1.951
|
| 15 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 99.96,
|
| 3 |
+
"eval_cer": 0.020881859054222414,
|
| 4 |
+
"eval_loss": 0.22571781277656555,
|
| 5 |
+
"eval_runtime": 70.8237,
|
| 6 |
+
"eval_samples": 1500,
|
| 7 |
+
"eval_samples_per_second": 21.179,
|
| 8 |
+
"eval_steps_per_second": 2.654,
|
| 9 |
+
"eval_wer": 0.06779661016949153
|
| 10 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 99.96,
|
| 3 |
+
"train_loss": 0.005932311532843826,
|
| 4 |
+
"train_runtime": 23991.1474,
|
| 5 |
+
"train_samples": 15000,
|
| 6 |
+
"train_samples_per_second": 62.523,
|
| 7 |
+
"train_steps_per_second": 1.951
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1043 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 99.9584,
|
| 5 |
+
"global_step": 46800,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 1.07,
|
| 12 |
+
"learning_rate": 2.9880000000000002e-05,
|
| 13 |
+
"loss": 7.2457,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 2.13,
|
| 18 |
+
"learning_rate": 2.967732181425486e-05,
|
| 19 |
+
"loss": 2.3129,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 2.13,
|
| 24 |
+
"eval_cer": 0.05796002383402849,
|
| 25 |
+
"eval_loss": 0.5042470097541809,
|
| 26 |
+
"eval_runtime": 80.7316,
|
| 27 |
+
"eval_samples_per_second": 18.58,
|
| 28 |
+
"eval_steps_per_second": 2.329,
|
| 29 |
+
"eval_wer": 0.270302835569122,
|
| 30 |
+
"step": 1000
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 3.2,
|
| 34 |
+
"learning_rate": 2.9353347732181423e-05,
|
| 35 |
+
"loss": 0.389,
|
| 36 |
+
"step": 1500
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"epoch": 4.27,
|
| 40 |
+
"learning_rate": 2.9029373650107992e-05,
|
| 41 |
+
"loss": 0.2251,
|
| 42 |
+
"step": 2000
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 4.27,
|
| 46 |
+
"eval_cer": 0.02945398407453551,
|
| 47 |
+
"eval_loss": 0.17824731767177582,
|
| 48 |
+
"eval_runtime": 80.3146,
|
| 49 |
+
"eval_samples_per_second": 18.677,
|
| 50 |
+
"eval_steps_per_second": 2.341,
|
| 51 |
+
"eval_wer": 0.11984898385412483,
|
| 52 |
+
"step": 2000
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 5.33,
|
| 56 |
+
"learning_rate": 2.870539956803456e-05,
|
| 57 |
+
"loss": 0.1727,
|
| 58 |
+
"step": 2500
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 6.4,
|
| 62 |
+
"learning_rate": 2.8381425485961122e-05,
|
| 63 |
+
"loss": 0.1462,
|
| 64 |
+
"step": 3000
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 6.4,
|
| 68 |
+
"eval_cer": 0.02651535669790369,
|
| 69 |
+
"eval_loss": 0.1635013222694397,
|
| 70 |
+
"eval_runtime": 72.2551,
|
| 71 |
+
"eval_samples_per_second": 20.76,
|
| 72 |
+
"eval_steps_per_second": 2.602,
|
| 73 |
+
"eval_wer": 0.10185557072857257,
|
| 74 |
+
"step": 3000
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 7.47,
|
| 78 |
+
"learning_rate": 2.805745140388769e-05,
|
| 79 |
+
"loss": 0.128,
|
| 80 |
+
"step": 3500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 8.53,
|
| 84 |
+
"learning_rate": 2.7733477321814256e-05,
|
| 85 |
+
"loss": 0.1162,
|
| 86 |
+
"step": 4000
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 8.53,
|
| 90 |
+
"eval_cer": 0.02478197280754022,
|
| 91 |
+
"eval_loss": 0.1619153469800949,
|
| 92 |
+
"eval_runtime": 81.2175,
|
| 93 |
+
"eval_samples_per_second": 18.469,
|
| 94 |
+
"eval_steps_per_second": 2.315,
|
| 95 |
+
"eval_wer": 0.09309984737729939,
|
| 96 |
+
"step": 4000
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 9.6,
|
| 100 |
+
"learning_rate": 2.740950323974082e-05,
|
| 101 |
+
"loss": 0.1095,
|
| 102 |
+
"step": 4500
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 10.67,
|
| 106 |
+
"learning_rate": 2.7085529157667386e-05,
|
| 107 |
+
"loss": 0.0988,
|
| 108 |
+
"step": 5000
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 10.67,
|
| 112 |
+
"eval_cer": 0.024890309300687936,
|
| 113 |
+
"eval_loss": 0.1653970628976822,
|
| 114 |
+
"eval_runtime": 72.7324,
|
| 115 |
+
"eval_samples_per_second": 20.624,
|
| 116 |
+
"eval_steps_per_second": 2.585,
|
| 117 |
+
"eval_wer": 0.09398345248614347,
|
| 118 |
+
"step": 5000
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 11.73,
|
| 122 |
+
"learning_rate": 2.676155507559395e-05,
|
| 123 |
+
"loss": 0.0949,
|
| 124 |
+
"step": 5500
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 12.8,
|
| 128 |
+
"learning_rate": 2.643758099352052e-05,
|
| 129 |
+
"loss": 0.0904,
|
| 130 |
+
"step": 6000
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 12.8,
|
| 134 |
+
"eval_cer": 0.024186122095227778,
|
| 135 |
+
"eval_loss": 0.17023487389087677,
|
| 136 |
+
"eval_runtime": 80.5666,
|
| 137 |
+
"eval_samples_per_second": 18.618,
|
| 138 |
+
"eval_steps_per_second": 2.333,
|
| 139 |
+
"eval_wer": 0.08450477950036148,
|
| 140 |
+
"step": 6000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 13.87,
|
| 144 |
+
"learning_rate": 2.6113606911447086e-05,
|
| 145 |
+
"loss": 0.0859,
|
| 146 |
+
"step": 6500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 14.93,
|
| 150 |
+
"learning_rate": 2.578963282937365e-05,
|
| 151 |
+
"loss": 0.0813,
|
| 152 |
+
"step": 7000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 14.93,
|
| 156 |
+
"eval_cer": 0.02392882292400195,
|
| 157 |
+
"eval_loss": 0.16578005254268646,
|
| 158 |
+
"eval_runtime": 72.3723,
|
| 159 |
+
"eval_samples_per_second": 20.726,
|
| 160 |
+
"eval_steps_per_second": 2.598,
|
| 161 |
+
"eval_wer": 0.08458510723752911,
|
| 162 |
+
"step": 7000
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 16.02,
|
| 166 |
+
"learning_rate": 2.5466306695464364e-05,
|
| 167 |
+
"loss": 0.0763,
|
| 168 |
+
"step": 7500
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 17.09,
|
| 172 |
+
"learning_rate": 2.514233261339093e-05,
|
| 173 |
+
"loss": 0.074,
|
| 174 |
+
"step": 8000
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 17.09,
|
| 178 |
+
"eval_cer": 0.02395590704728888,
|
| 179 |
+
"eval_loss": 0.1762518584728241,
|
| 180 |
+
"eval_runtime": 72.6367,
|
| 181 |
+
"eval_samples_per_second": 20.651,
|
| 182 |
+
"eval_steps_per_second": 2.588,
|
| 183 |
+
"eval_wer": 0.07928347658446462,
|
| 184 |
+
"step": 8000
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 18.16,
|
| 188 |
+
"learning_rate": 2.4818358531317497e-05,
|
| 189 |
+
"loss": 0.0714,
|
| 190 |
+
"step": 8500
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 19.22,
|
| 194 |
+
"learning_rate": 2.4495032397408207e-05,
|
| 195 |
+
"loss": 0.0692,
|
| 196 |
+
"step": 9000
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 19.22,
|
| 200 |
+
"eval_cer": 0.0242538324034451,
|
| 201 |
+
"eval_loss": 0.17680838704109192,
|
| 202 |
+
"eval_runtime": 68.8274,
|
| 203 |
+
"eval_samples_per_second": 21.794,
|
| 204 |
+
"eval_steps_per_second": 2.731,
|
| 205 |
+
"eval_wer": 0.0834605189171821,
|
| 206 |
+
"step": 9000
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 20.3,
|
| 210 |
+
"learning_rate": 2.4171058315334776e-05,
|
| 211 |
+
"loss": 0.0658,
|
| 212 |
+
"step": 9500
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 21.36,
|
| 216 |
+
"learning_rate": 2.3847084233261337e-05,
|
| 217 |
+
"loss": 0.0652,
|
| 218 |
+
"step": 10000
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 21.36,
|
| 222 |
+
"eval_cer": 0.02365798169113266,
|
| 223 |
+
"eval_loss": 0.18117107450962067,
|
| 224 |
+
"eval_runtime": 72.0293,
|
| 225 |
+
"eval_samples_per_second": 20.825,
|
| 226 |
+
"eval_steps_per_second": 2.61,
|
| 227 |
+
"eval_wer": 0.07968511527030284,
|
| 228 |
+
"step": 10000
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 22.43,
|
| 232 |
+
"learning_rate": 2.3523110151187906e-05,
|
| 233 |
+
"loss": 0.0625,
|
| 234 |
+
"step": 10500
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 23.5,
|
| 238 |
+
"learning_rate": 2.319913606911447e-05,
|
| 239 |
+
"loss": 0.0593,
|
| 240 |
+
"step": 11000
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 23.5,
|
| 244 |
+
"eval_cer": 0.022127728725421156,
|
| 245 |
+
"eval_loss": 0.18102437257766724,
|
| 246 |
+
"eval_runtime": 69.723,
|
| 247 |
+
"eval_samples_per_second": 21.514,
|
| 248 |
+
"eval_steps_per_second": 2.696,
|
| 249 |
+
"eval_wer": 0.07502610651457949,
|
| 250 |
+
"step": 11000
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 24.56,
|
| 254 |
+
"learning_rate": 2.2875809935205184e-05,
|
| 255 |
+
"loss": 0.0604,
|
| 256 |
+
"step": 11500
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 25.63,
|
| 260 |
+
"learning_rate": 2.255183585313175e-05,
|
| 261 |
+
"loss": 0.0547,
|
| 262 |
+
"step": 12000
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 25.63,
|
| 266 |
+
"eval_cer": 0.023265261903472185,
|
| 267 |
+
"eval_loss": 0.18346643447875977,
|
| 268 |
+
"eval_runtime": 70.9539,
|
| 269 |
+
"eval_samples_per_second": 21.14,
|
| 270 |
+
"eval_steps_per_second": 2.65,
|
| 271 |
+
"eval_wer": 0.07944413205879991,
|
| 272 |
+
"step": 12000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 26.7,
|
| 276 |
+
"learning_rate": 2.2227861771058315e-05,
|
| 277 |
+
"loss": 0.0565,
|
| 278 |
+
"step": 12500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 27.76,
|
| 282 |
+
"learning_rate": 2.1903887688984883e-05,
|
| 283 |
+
"loss": 0.0514,
|
| 284 |
+
"step": 13000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 27.76,
|
| 288 |
+
"eval_cer": 0.022385027896646984,
|
| 289 |
+
"eval_loss": 0.18281148374080658,
|
| 290 |
+
"eval_runtime": 70.1644,
|
| 291 |
+
"eval_samples_per_second": 21.378,
|
| 292 |
+
"eval_steps_per_second": 2.679,
|
| 293 |
+
"eval_wer": 0.07607036709775886,
|
| 294 |
+
"step": 13000
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 28.83,
|
| 298 |
+
"learning_rate": 2.157991360691145e-05,
|
| 299 |
+
"loss": 0.053,
|
| 300 |
+
"step": 13500
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 29.9,
|
| 304 |
+
"learning_rate": 2.1255939524838014e-05,
|
| 305 |
+
"loss": 0.0488,
|
| 306 |
+
"step": 14000
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 29.9,
|
| 310 |
+
"eval_cer": 0.02242565408157738,
|
| 311 |
+
"eval_loss": 0.18435174226760864,
|
| 312 |
+
"eval_runtime": 69.6075,
|
| 313 |
+
"eval_samples_per_second": 21.549,
|
| 314 |
+
"eval_steps_per_second": 2.701,
|
| 315 |
+
"eval_wer": 0.07663266125793236,
|
| 316 |
+
"step": 14000
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 30.96,
|
| 320 |
+
"learning_rate": 2.093196544276458e-05,
|
| 321 |
+
"loss": 0.0514,
|
| 322 |
+
"step": 14500
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 32.03,
|
| 326 |
+
"learning_rate": 2.0607991360691144e-05,
|
| 327 |
+
"loss": 0.0478,
|
| 328 |
+
"step": 15000
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 32.03,
|
| 332 |
+
"eval_cer": 0.022588158821298953,
|
| 333 |
+
"eval_loss": 0.1909506469964981,
|
| 334 |
+
"eval_runtime": 72.4928,
|
| 335 |
+
"eval_samples_per_second": 20.692,
|
| 336 |
+
"eval_steps_per_second": 2.593,
|
| 337 |
+
"eval_wer": 0.0768736444694353,
|
| 338 |
+
"step": 15000
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 33.1,
|
| 342 |
+
"learning_rate": 2.0284017278617713e-05,
|
| 343 |
+
"loss": 0.0489,
|
| 344 |
+
"step": 15500
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 34.16,
|
| 348 |
+
"learning_rate": 1.9960043196544274e-05,
|
| 349 |
+
"loss": 0.0459,
|
| 350 |
+
"step": 16000
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 34.16,
|
| 354 |
+
"eval_cer": 0.02386111261578463,
|
| 355 |
+
"eval_loss": 0.19651177525520325,
|
| 356 |
+
"eval_runtime": 69.572,
|
| 357 |
+
"eval_samples_per_second": 21.56,
|
| 358 |
+
"eval_steps_per_second": 2.702,
|
| 359 |
+
"eval_wer": 0.08313920796851153,
|
| 360 |
+
"step": 16000
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 35.23,
|
| 364 |
+
"learning_rate": 1.9636069114470843e-05,
|
| 365 |
+
"loss": 0.0481,
|
| 366 |
+
"step": 16500
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 36.3,
|
| 370 |
+
"learning_rate": 1.9312095032397408e-05,
|
| 371 |
+
"loss": 0.0429,
|
| 372 |
+
"step": 17000
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 36.3,
|
| 376 |
+
"eval_cer": 0.021951681924056117,
|
| 377 |
+
"eval_loss": 0.20002886652946472,
|
| 378 |
+
"eval_runtime": 69.2113,
|
| 379 |
+
"eval_samples_per_second": 21.673,
|
| 380 |
+
"eval_steps_per_second": 2.716,
|
| 381 |
+
"eval_wer": 0.07599003936059122,
|
| 382 |
+
"step": 17000
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 37.36,
|
| 386 |
+
"learning_rate": 1.898876889848812e-05,
|
| 387 |
+
"loss": 0.0461,
|
| 388 |
+
"step": 17500
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 38.43,
|
| 392 |
+
"learning_rate": 1.8664794816414686e-05,
|
| 393 |
+
"loss": 0.0443,
|
| 394 |
+
"step": 18000
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 38.43,
|
| 398 |
+
"eval_cer": 0.02277774768430746,
|
| 399 |
+
"eval_loss": 0.20385122299194336,
|
| 400 |
+
"eval_runtime": 72.6401,
|
| 401 |
+
"eval_samples_per_second": 20.65,
|
| 402 |
+
"eval_steps_per_second": 2.588,
|
| 403 |
+
"eval_wer": 0.0774359386296088,
|
| 404 |
+
"step": 18000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 39.5,
|
| 408 |
+
"learning_rate": 1.834082073434125e-05,
|
| 409 |
+
"loss": 0.0411,
|
| 410 |
+
"step": 18500
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 40.56,
|
| 414 |
+
"learning_rate": 1.8017494600431965e-05,
|
| 415 |
+
"loss": 0.0398,
|
| 416 |
+
"step": 19000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 40.56,
|
| 420 |
+
"eval_cer": 0.021856887492551866,
|
| 421 |
+
"eval_loss": 0.19806204736232758,
|
| 422 |
+
"eval_runtime": 72.4106,
|
| 423 |
+
"eval_samples_per_second": 20.715,
|
| 424 |
+
"eval_steps_per_second": 2.596,
|
| 425 |
+
"eval_wer": 0.07550807293758535,
|
| 426 |
+
"step": 19000
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 41.63,
|
| 430 |
+
"learning_rate": 1.7693520518358533e-05,
|
| 431 |
+
"loss": 0.0431,
|
| 432 |
+
"step": 19500
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 42.7,
|
| 436 |
+
"learning_rate": 1.7369546436285095e-05,
|
| 437 |
+
"loss": 0.0408,
|
| 438 |
+
"step": 20000
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 42.7,
|
| 442 |
+
"eval_cer": 0.02387465467742809,
|
| 443 |
+
"eval_loss": 0.2053176462650299,
|
| 444 |
+
"eval_runtime": 69.6266,
|
| 445 |
+
"eval_samples_per_second": 21.543,
|
| 446 |
+
"eval_steps_per_second": 2.7,
|
| 447 |
+
"eval_wer": 0.0775965941039441,
|
| 448 |
+
"step": 20000
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 43.76,
|
| 452 |
+
"learning_rate": 1.7045572354211664e-05,
|
| 453 |
+
"loss": 0.0399,
|
| 454 |
+
"step": 20500
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 44.83,
|
| 458 |
+
"learning_rate": 1.6721598272138232e-05,
|
| 459 |
+
"loss": 0.0406,
|
| 460 |
+
"step": 21000
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 44.83,
|
| 464 |
+
"eval_cer": 0.022060018417203835,
|
| 465 |
+
"eval_loss": 0.2049773633480072,
|
| 466 |
+
"eval_runtime": 70.2646,
|
| 467 |
+
"eval_samples_per_second": 21.348,
|
| 468 |
+
"eval_steps_per_second": 2.676,
|
| 469 |
+
"eval_wer": 0.07398184593140011,
|
| 470 |
+
"step": 21000
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 45.9,
|
| 474 |
+
"learning_rate": 1.6397624190064794e-05,
|
| 475 |
+
"loss": 0.0381,
|
| 476 |
+
"step": 21500
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 46.96,
|
| 480 |
+
"learning_rate": 1.6073650107991363e-05,
|
| 481 |
+
"loss": 0.0383,
|
| 482 |
+
"step": 22000
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 46.96,
|
| 486 |
+
"eval_cer": 0.022357943773360055,
|
| 487 |
+
"eval_loss": 0.21280354261398315,
|
| 488 |
+
"eval_runtime": 72.639,
|
| 489 |
+
"eval_samples_per_second": 20.65,
|
| 490 |
+
"eval_steps_per_second": 2.588,
|
| 491 |
+
"eval_wer": 0.07325889629689132,
|
| 492 |
+
"step": 22000
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 48.03,
|
| 496 |
+
"learning_rate": 1.5749676025917928e-05,
|
| 497 |
+
"loss": 0.0365,
|
| 498 |
+
"step": 22500
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 49.1,
|
| 502 |
+
"learning_rate": 1.5425701943844493e-05,
|
| 503 |
+
"loss": 0.0379,
|
| 504 |
+
"step": 23000
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 49.1,
|
| 508 |
+
"eval_cer": 0.021978766047343046,
|
| 509 |
+
"eval_loss": 0.21096549928188324,
|
| 510 |
+
"eval_runtime": 73.1028,
|
| 511 |
+
"eval_samples_per_second": 20.519,
|
| 512 |
+
"eval_steps_per_second": 2.572,
|
| 513 |
+
"eval_wer": 0.07309824082255603,
|
| 514 |
+
"step": 23000
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 50.16,
|
| 518 |
+
"learning_rate": 1.5102375809935208e-05,
|
| 519 |
+
"loss": 0.0361,
|
| 520 |
+
"step": 23500
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 51.23,
|
| 524 |
+
"learning_rate": 1.4778401727861771e-05,
|
| 525 |
+
"loss": 0.0369,
|
| 526 |
+
"step": 24000
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 51.23,
|
| 530 |
+
"eval_cer": 0.021951681924056117,
|
| 531 |
+
"eval_loss": 0.2144922912120819,
|
| 532 |
+
"eval_runtime": 69.8484,
|
| 533 |
+
"eval_samples_per_second": 21.475,
|
| 534 |
+
"eval_steps_per_second": 2.692,
|
| 535 |
+
"eval_wer": 0.07446381235440598,
|
| 536 |
+
"step": 24000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 52.3,
|
| 540 |
+
"learning_rate": 1.4455075593952484e-05,
|
| 541 |
+
"loss": 0.0334,
|
| 542 |
+
"step": 24500
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 53.36,
|
| 546 |
+
"learning_rate": 1.4131101511879051e-05,
|
| 547 |
+
"loss": 0.0341,
|
| 548 |
+
"step": 25000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 53.36,
|
| 552 |
+
"eval_cer": 0.022208981095281945,
|
| 553 |
+
"eval_loss": 0.2145662158727646,
|
| 554 |
+
"eval_runtime": 69.6873,
|
| 555 |
+
"eval_samples_per_second": 21.525,
|
| 556 |
+
"eval_steps_per_second": 2.698,
|
| 557 |
+
"eval_wer": 0.07245561892521488,
|
| 558 |
+
"step": 25000
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 54.43,
|
| 562 |
+
"learning_rate": 1.3807127429805616e-05,
|
| 563 |
+
"loss": 0.0356,
|
| 564 |
+
"step": 25500
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 55.5,
|
| 568 |
+
"learning_rate": 1.3483801295896327e-05,
|
| 569 |
+
"loss": 0.0322,
|
| 570 |
+
"step": 26000
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 55.5,
|
| 574 |
+
"eval_cer": 0.021626672444612968,
|
| 575 |
+
"eval_loss": 0.21297892928123474,
|
| 576 |
+
"eval_runtime": 73.9282,
|
| 577 |
+
"eval_samples_per_second": 20.29,
|
| 578 |
+
"eval_steps_per_second": 2.543,
|
| 579 |
+
"eval_wer": 0.07100971965619729,
|
| 580 |
+
"step": 26000
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 56.56,
|
| 584 |
+
"learning_rate": 1.3159827213822896e-05,
|
| 585 |
+
"loss": 0.0327,
|
| 586 |
+
"step": 26500
|
| 587 |
+
},
|
| 588 |
+
{
|
| 589 |
+
"epoch": 57.63,
|
| 590 |
+
"learning_rate": 1.2835853131749461e-05,
|
| 591 |
+
"loss": 0.0316,
|
| 592 |
+
"step": 27000
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 57.63,
|
| 596 |
+
"eval_cer": 0.02224960728021234,
|
| 597 |
+
"eval_loss": 0.2133886069059372,
|
| 598 |
+
"eval_runtime": 69.9039,
|
| 599 |
+
"eval_samples_per_second": 21.458,
|
| 600 |
+
"eval_steps_per_second": 2.689,
|
| 601 |
+
"eval_wer": 0.0715720138163708,
|
| 602 |
+
"step": 27000
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 58.7,
|
| 606 |
+
"learning_rate": 1.2511879049676026e-05,
|
| 607 |
+
"loss": 0.0319,
|
| 608 |
+
"step": 27500
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 59.76,
|
| 612 |
+
"learning_rate": 1.218855291576674e-05,
|
| 613 |
+
"loss": 0.0324,
|
| 614 |
+
"step": 28000
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 59.76,
|
| 618 |
+
"eval_cer": 0.022236065218568874,
|
| 619 |
+
"eval_loss": 0.21722277998924255,
|
| 620 |
+
"eval_runtime": 69.753,
|
| 621 |
+
"eval_samples_per_second": 21.504,
|
| 622 |
+
"eval_steps_per_second": 2.695,
|
| 623 |
+
"eval_wer": 0.07309824082255603,
|
| 624 |
+
"step": 28000
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 60.83,
|
| 628 |
+
"learning_rate": 1.1864578833693305e-05,
|
| 629 |
+
"loss": 0.0296,
|
| 630 |
+
"step": 28500
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 61.9,
|
| 634 |
+
"learning_rate": 1.1540604751619871e-05,
|
| 635 |
+
"loss": 0.0315,
|
| 636 |
+
"step": 29000
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 61.9,
|
| 640 |
+
"eval_cer": 0.02280483180759439,
|
| 641 |
+
"eval_loss": 0.22069784998893738,
|
| 642 |
+
"eval_runtime": 79.4911,
|
| 643 |
+
"eval_samples_per_second": 18.87,
|
| 644 |
+
"eval_steps_per_second": 2.365,
|
| 645 |
+
"eval_wer": 0.07454414009157362,
|
| 646 |
+
"step": 29000
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 62.96,
|
| 650 |
+
"learning_rate": 1.1216630669546437e-05,
|
| 651 |
+
"loss": 0.0315,
|
| 652 |
+
"step": 29500
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 64.03,
|
| 656 |
+
"learning_rate": 1.0892656587473002e-05,
|
| 657 |
+
"loss": 0.0294,
|
| 658 |
+
"step": 30000
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 64.03,
|
| 662 |
+
"eval_cer": 0.021843345430908403,
|
| 663 |
+
"eval_loss": 0.2183454930782318,
|
| 664 |
+
"eval_runtime": 77.818,
|
| 665 |
+
"eval_samples_per_second": 19.276,
|
| 666 |
+
"eval_steps_per_second": 2.416,
|
| 667 |
+
"eval_wer": 0.07165234155353843,
|
| 668 |
+
"step": 30000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 65.1,
|
| 672 |
+
"learning_rate": 1.0569330453563715e-05,
|
| 673 |
+
"loss": 0.0301,
|
| 674 |
+
"step": 30500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 66.16,
|
| 678 |
+
"learning_rate": 1.024535637149028e-05,
|
| 679 |
+
"loss": 0.028,
|
| 680 |
+
"step": 31000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 66.16,
|
| 684 |
+
"eval_cer": 0.021355831211743677,
|
| 685 |
+
"eval_loss": 0.2184668779373169,
|
| 686 |
+
"eval_runtime": 80.0286,
|
| 687 |
+
"eval_samples_per_second": 18.743,
|
| 688 |
+
"eval_steps_per_second": 2.349,
|
| 689 |
+
"eval_wer": 0.06956382038717969,
|
| 690 |
+
"step": 31000
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 67.23,
|
| 694 |
+
"learning_rate": 9.921382289416847e-06,
|
| 695 |
+
"loss": 0.0283,
|
| 696 |
+
"step": 31500
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 68.3,
|
| 700 |
+
"learning_rate": 9.597408207343414e-06,
|
| 701 |
+
"loss": 0.0263,
|
| 702 |
+
"step": 32000
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 68.3,
|
| 706 |
+
"eval_cer": 0.021504793889821787,
|
| 707 |
+
"eval_loss": 0.21668484807014465,
|
| 708 |
+
"eval_runtime": 70.5125,
|
| 709 |
+
"eval_samples_per_second": 21.273,
|
| 710 |
+
"eval_steps_per_second": 2.666,
|
| 711 |
+
"eval_wer": 0.06964414812434734,
|
| 712 |
+
"step": 32000
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 69.36,
|
| 716 |
+
"learning_rate": 9.273434125269979e-06,
|
| 717 |
+
"loss": 0.0286,
|
| 718 |
+
"step": 32500
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 70.43,
|
| 722 |
+
"learning_rate": 8.949460043196544e-06,
|
| 723 |
+
"loss": 0.0299,
|
| 724 |
+
"step": 33000
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 70.43,
|
| 728 |
+
"eval_cer": 0.021707924814473756,
|
| 729 |
+
"eval_loss": 0.22012607753276825,
|
| 730 |
+
"eval_runtime": 72.8041,
|
| 731 |
+
"eval_samples_per_second": 20.603,
|
| 732 |
+
"eval_steps_per_second": 2.582,
|
| 733 |
+
"eval_wer": 0.07092939191902964,
|
| 734 |
+
"step": 33000
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 71.5,
|
| 738 |
+
"learning_rate": 8.62548596112311e-06,
|
| 739 |
+
"loss": 0.0267,
|
| 740 |
+
"step": 33500
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 72.56,
|
| 744 |
+
"learning_rate": 8.301511879049676e-06,
|
| 745 |
+
"loss": 0.0273,
|
| 746 |
+
"step": 34000
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 72.56,
|
| 750 |
+
"eval_cer": 0.02219543903363848,
|
| 751 |
+
"eval_loss": 0.21641910076141357,
|
| 752 |
+
"eval_runtime": 72.2901,
|
| 753 |
+
"eval_samples_per_second": 20.75,
|
| 754 |
+
"eval_steps_per_second": 2.601,
|
| 755 |
+
"eval_wer": 0.07237529118804724,
|
| 756 |
+
"step": 34000
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 73.63,
|
| 760 |
+
"learning_rate": 7.977537796976242e-06,
|
| 761 |
+
"loss": 0.0267,
|
| 762 |
+
"step": 34500
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 74.7,
|
| 766 |
+
"learning_rate": 7.653563714902809e-06,
|
| 767 |
+
"loss": 0.0269,
|
| 768 |
+
"step": 35000
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 74.7,
|
| 772 |
+
"eval_cer": 0.021951681924056117,
|
| 773 |
+
"eval_loss": 0.22398144006729126,
|
| 774 |
+
"eval_runtime": 69.5423,
|
| 775 |
+
"eval_samples_per_second": 21.57,
|
| 776 |
+
"eval_steps_per_second": 2.703,
|
| 777 |
+
"eval_wer": 0.06932283717567676,
|
| 778 |
+
"step": 35000
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 75.85,
|
| 782 |
+
"learning_rate": 7.329589632829374e-06,
|
| 783 |
+
"loss": 0.0265,
|
| 784 |
+
"step": 35500
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 76.92,
|
| 788 |
+
"learning_rate": 7.00561555075594e-06,
|
| 789 |
+
"loss": 0.0264,
|
| 790 |
+
"step": 36000
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 76.92,
|
| 794 |
+
"eval_cer": 0.021789177184334544,
|
| 795 |
+
"eval_loss": 0.22204121947288513,
|
| 796 |
+
"eval_runtime": 71.976,
|
| 797 |
+
"eval_samples_per_second": 20.84,
|
| 798 |
+
"eval_steps_per_second": 2.612,
|
| 799 |
+
"eval_wer": 0.07036709775885613,
|
| 800 |
+
"step": 36000
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 77.99,
|
| 804 |
+
"learning_rate": 6.682289416846653e-06,
|
| 805 |
+
"loss": 0.0257,
|
| 806 |
+
"step": 36500
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 79.05,
|
| 810 |
+
"learning_rate": 6.358963282937365e-06,
|
| 811 |
+
"loss": 0.0257,
|
| 812 |
+
"step": 37000
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 79.05,
|
| 816 |
+
"eval_cer": 0.02172146687611722,
|
| 817 |
+
"eval_loss": 0.22285164892673492,
|
| 818 |
+
"eval_runtime": 68.7898,
|
| 819 |
+
"eval_samples_per_second": 21.806,
|
| 820 |
+
"eval_steps_per_second": 2.733,
|
| 821 |
+
"eval_wer": 0.0688408707526709,
|
| 822 |
+
"step": 37000
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 80.12,
|
| 826 |
+
"learning_rate": 6.034989200863931e-06,
|
| 827 |
+
"loss": 0.0257,
|
| 828 |
+
"step": 37500
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 81.19,
|
| 832 |
+
"learning_rate": 5.711015118790497e-06,
|
| 833 |
+
"loss": 0.0251,
|
| 834 |
+
"step": 38000
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 81.19,
|
| 838 |
+
"eval_cer": 0.02154542007475218,
|
| 839 |
+
"eval_loss": 0.22632832825183868,
|
| 840 |
+
"eval_runtime": 71.6223,
|
| 841 |
+
"eval_samples_per_second": 20.943,
|
| 842 |
+
"eval_steps_per_second": 2.625,
|
| 843 |
+
"eval_wer": 0.0694031649128444,
|
| 844 |
+
"step": 38000
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 82.25,
|
| 848 |
+
"learning_rate": 5.387041036717062e-06,
|
| 849 |
+
"loss": 0.0235,
|
| 850 |
+
"step": 38500
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 83.32,
|
| 854 |
+
"learning_rate": 5.063066954643629e-06,
|
| 855 |
+
"loss": 0.0245,
|
| 856 |
+
"step": 39000
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 83.32,
|
| 860 |
+
"eval_cer": 0.02101727967065706,
|
| 861 |
+
"eval_loss": 0.22526498138904572,
|
| 862 |
+
"eval_runtime": 70.7229,
|
| 863 |
+
"eval_samples_per_second": 21.21,
|
| 864 |
+
"eval_steps_per_second": 2.658,
|
| 865 |
+
"eval_wer": 0.06731464374648566,
|
| 866 |
+
"step": 39000
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 84.39,
|
| 870 |
+
"learning_rate": 4.739740820734342e-06,
|
| 871 |
+
"loss": 0.0252,
|
| 872 |
+
"step": 39500
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 85.45,
|
| 876 |
+
"learning_rate": 4.415766738660907e-06,
|
| 877 |
+
"loss": 0.0243,
|
| 878 |
+
"step": 40000
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 85.45,
|
| 882 |
+
"eval_cer": 0.02151833595146525,
|
| 883 |
+
"eval_loss": 0.2263830602169037,
|
| 884 |
+
"eval_runtime": 69.4576,
|
| 885 |
+
"eval_samples_per_second": 21.596,
|
| 886 |
+
"eval_steps_per_second": 2.707,
|
| 887 |
+
"eval_wer": 0.06916218170134147,
|
| 888 |
+
"step": 40000
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 86.52,
|
| 892 |
+
"learning_rate": 4.09244060475162e-06,
|
| 893 |
+
"loss": 0.0228,
|
| 894 |
+
"step": 40500
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 87.59,
|
| 898 |
+
"learning_rate": 3.7684665226781857e-06,
|
| 899 |
+
"loss": 0.0236,
|
| 900 |
+
"step": 41000
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 87.59,
|
| 904 |
+
"eval_cer": 0.021667298629543363,
|
| 905 |
+
"eval_loss": 0.22610026597976685,
|
| 906 |
+
"eval_runtime": 69.0446,
|
| 907 |
+
"eval_samples_per_second": 21.725,
|
| 908 |
+
"eval_steps_per_second": 2.723,
|
| 909 |
+
"eval_wer": 0.06892119848983855,
|
| 910 |
+
"step": 41000
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 88.65,
|
| 914 |
+
"learning_rate": 3.4444924406047518e-06,
|
| 915 |
+
"loss": 0.0244,
|
| 916 |
+
"step": 41500
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 89.72,
|
| 920 |
+
"learning_rate": 3.1205183585313174e-06,
|
| 921 |
+
"loss": 0.0225,
|
| 922 |
+
"step": 42000
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 89.72,
|
| 926 |
+
"eval_cer": 0.021179784410378637,
|
| 927 |
+
"eval_loss": 0.22654421627521515,
|
| 928 |
+
"eval_runtime": 72.2716,
|
| 929 |
+
"eval_samples_per_second": 20.755,
|
| 930 |
+
"eval_steps_per_second": 2.601,
|
| 931 |
+
"eval_wer": 0.06803759338099445,
|
| 932 |
+
"step": 42000
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 90.79,
|
| 936 |
+
"learning_rate": 2.7965442764578835e-06,
|
| 937 |
+
"loss": 0.0216,
|
| 938 |
+
"step": 42500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 91.85,
|
| 942 |
+
"learning_rate": 2.472570194384449e-06,
|
| 943 |
+
"loss": 0.023,
|
| 944 |
+
"step": 43000
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 91.85,
|
| 948 |
+
"eval_cer": 0.021030821732300524,
|
| 949 |
+
"eval_loss": 0.22652284801006317,
|
| 950 |
+
"eval_runtime": 72.2083,
|
| 951 |
+
"eval_samples_per_second": 20.773,
|
| 952 |
+
"eval_steps_per_second": 2.604,
|
| 953 |
+
"eval_wer": 0.06739497148365331,
|
| 954 |
+
"step": 43000
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 92.92,
|
| 958 |
+
"learning_rate": 2.149244060475162e-06,
|
| 959 |
+
"loss": 0.0228,
|
| 960 |
+
"step": 43500
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 93.99,
|
| 964 |
+
"learning_rate": 1.825917926565875e-06,
|
| 965 |
+
"loss": 0.0217,
|
| 966 |
+
"step": 44000
|
| 967 |
+
},
|
| 968 |
+
{
|
| 969 |
+
"epoch": 93.99,
|
| 970 |
+
"eval_cer": 0.020936027300796273,
|
| 971 |
+
"eval_loss": 0.22653113305568695,
|
| 972 |
+
"eval_runtime": 69.3882,
|
| 973 |
+
"eval_samples_per_second": 21.618,
|
| 974 |
+
"eval_steps_per_second": 2.709,
|
| 975 |
+
"eval_wer": 0.06771628243232389,
|
| 976 |
+
"step": 44000
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 95.05,
|
| 980 |
+
"learning_rate": 1.5019438444924408e-06,
|
| 981 |
+
"loss": 0.0203,
|
| 982 |
+
"step": 44500
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 96.12,
|
| 986 |
+
"learning_rate": 1.1779697624190064e-06,
|
| 987 |
+
"loss": 0.022,
|
| 988 |
+
"step": 45000
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 96.12,
|
| 992 |
+
"eval_cer": 0.021057905855587453,
|
| 993 |
+
"eval_loss": 0.22544603049755096,
|
| 994 |
+
"eval_runtime": 72.3315,
|
| 995 |
+
"eval_samples_per_second": 20.738,
|
| 996 |
+
"eval_steps_per_second": 2.599,
|
| 997 |
+
"eval_wer": 0.06851955980400032,
|
| 998 |
+
"step": 45000
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 97.19,
|
| 1002 |
+
"learning_rate": 8.539956803455724e-07,
|
| 1003 |
+
"loss": 0.0215,
|
| 1004 |
+
"step": 45500
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 98.25,
|
| 1008 |
+
"learning_rate": 5.300215982721382e-07,
|
| 1009 |
+
"loss": 0.0219,
|
| 1010 |
+
"step": 46000
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 98.25,
|
| 1014 |
+
"eval_cer": 0.020759980499431233,
|
| 1015 |
+
"eval_loss": 0.22618666291236877,
|
| 1016 |
+
"eval_runtime": 73.3776,
|
| 1017 |
+
"eval_samples_per_second": 20.442,
|
| 1018 |
+
"eval_steps_per_second": 2.562,
|
| 1019 |
+
"eval_wer": 0.06715398827215037,
|
| 1020 |
+
"step": 46000
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 99.32,
|
| 1024 |
+
"learning_rate": 2.0604751619870412e-07,
|
| 1025 |
+
"loss": 0.0204,
|
| 1026 |
+
"step": 46500
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 99.96,
|
| 1030 |
+
"step": 46800,
|
| 1031 |
+
"total_flos": 2.6878408769720543e+20,
|
| 1032 |
+
"train_loss": 0.005932311532843826,
|
| 1033 |
+
"train_runtime": 23991.1474,
|
| 1034 |
+
"train_samples_per_second": 62.523,
|
| 1035 |
+
"train_steps_per_second": 1.951
|
| 1036 |
+
}
|
| 1037 |
+
],
|
| 1038 |
+
"max_steps": 46800,
|
| 1039 |
+
"num_train_epochs": 100,
|
| 1040 |
+
"total_flos": 2.6878408769720543e+20,
|
| 1041 |
+
"trial_name": null,
|
| 1042 |
+
"trial_params": null
|
| 1043 |
+
}
|