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 +883 -0
README.md
CHANGED
|
@@ -3,6 +3,9 @@ library_name: transformers
|
|
| 3 |
license: cc-by-nc-4.0
|
| 4 |
base_model: facebook/mms-1b-all
|
| 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-1b-all-bem-genbed-combined-adapter-test
|
| 18 |
|
| 19 |
-
This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) 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: facebook/mms-1b-all
|
| 5 |
tags:
|
| 6 |
+
- automatic-speech-recognition
|
| 7 |
+
- genbed
|
| 8 |
+
- mms
|
| 9 |
- generated_from_trainer
|
| 10 |
metrics:
|
| 11 |
- wer
|
|
|
|
| 19 |
|
| 20 |
# mms-1b-all-bem-genbed-combined-adapter-test
|
| 21 |
|
| 22 |
+
This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on the GENBED - BEM dataset.
|
| 23 |
It achieves the following results on the evaluation set:
|
| 24 |
+
- Loss: 0.2293
|
| 25 |
+
- Wer: 0.3752
|
| 26 |
|
| 27 |
## Model description
|
| 28 |
|
adapter.bem.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:96eafc43ef00af7d94701f7e9f824e572d894b48630b41fa45a081dd412253a9
|
| 3 |
+
size 8798532
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 7.142857142857143,
|
| 3 |
+
"eval_loss": 0.22929497063159943,
|
| 4 |
+
"eval_runtime": 134.2212,
|
| 5 |
+
"eval_samples": 1939,
|
| 6 |
+
"eval_samples_per_second": 14.446,
|
| 7 |
+
"eval_steps_per_second": 1.81,
|
| 8 |
+
"eval_wer": 0.3751891891891892,
|
| 9 |
+
"total_flos": 2.4864489394936607e+19,
|
| 10 |
+
"train_loss": 0.46319753646850587,
|
| 11 |
+
"train_runtime": 13134.8195,
|
| 12 |
+
"train_samples": 5817,
|
| 13 |
+
"train_samples_per_second": 13.286,
|
| 14 |
+
"train_steps_per_second": 1.663
|
| 15 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 7.142857142857143,
|
| 3 |
+
"eval_loss": 0.22929497063159943,
|
| 4 |
+
"eval_runtime": 134.2212,
|
| 5 |
+
"eval_samples": 1939,
|
| 6 |
+
"eval_samples_per_second": 14.446,
|
| 7 |
+
"eval_steps_per_second": 1.81,
|
| 8 |
+
"eval_wer": 0.3751891891891892
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 7.142857142857143,
|
| 3 |
+
"total_flos": 2.4864489394936607e+19,
|
| 4 |
+
"train_loss": 0.46319753646850587,
|
| 5 |
+
"train_runtime": 13134.8195,
|
| 6 |
+
"train_samples": 5817,
|
| 7 |
+
"train_samples_per_second": 13.286,
|
| 8 |
+
"train_steps_per_second": 1.663
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,883 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.22929497063159943,
|
| 3 |
+
"best_model_checkpoint": "/scratch/skscla001/results/mms-1b-all-bem-genbed-combined-adapter-test/checkpoint-4800",
|
| 4 |
+
"epoch": 7.142857142857143,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 5200,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.13736263736263737,
|
| 13 |
+
"grad_norm": 3.0789577960968018,
|
| 14 |
+
"learning_rate": 0.00028799999999999995,
|
| 15 |
+
"loss": 6.4783,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.13736263736263737,
|
| 20 |
+
"eval_loss": 0.6402517557144165,
|
| 21 |
+
"eval_runtime": 132.5796,
|
| 22 |
+
"eval_samples_per_second": 14.625,
|
| 23 |
+
"eval_steps_per_second": 1.833,
|
| 24 |
+
"eval_wer": 0.69,
|
| 25 |
+
"step": 100
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.27472527472527475,
|
| 29 |
+
"grad_norm": 1.5154528617858887,
|
| 30 |
+
"learning_rate": 0.0002986752529898804,
|
| 31 |
+
"loss": 0.5413,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.27472527472527475,
|
| 36 |
+
"eval_loss": 0.32597342133522034,
|
| 37 |
+
"eval_runtime": 131.6167,
|
| 38 |
+
"eval_samples_per_second": 14.732,
|
| 39 |
+
"eval_steps_per_second": 1.846,
|
| 40 |
+
"eval_wer": 0.4931351351351351,
|
| 41 |
+
"step": 200
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.41208791208791207,
|
| 45 |
+
"grad_norm": 1.3768789768218994,
|
| 46 |
+
"learning_rate": 0.0002972953081876725,
|
| 47 |
+
"loss": 0.437,
|
| 48 |
+
"step": 300
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 0.41208791208791207,
|
| 52 |
+
"eval_loss": 0.30144402384757996,
|
| 53 |
+
"eval_runtime": 132.6277,
|
| 54 |
+
"eval_samples_per_second": 14.62,
|
| 55 |
+
"eval_steps_per_second": 1.832,
|
| 56 |
+
"eval_wer": 0.46778378378378377,
|
| 57 |
+
"step": 300
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.5494505494505495,
|
| 61 |
+
"grad_norm": 2.0690324306488037,
|
| 62 |
+
"learning_rate": 0.00029591536338546457,
|
| 63 |
+
"loss": 0.4402,
|
| 64 |
+
"step": 400
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 0.5494505494505495,
|
| 68 |
+
"eval_loss": 0.29820021986961365,
|
| 69 |
+
"eval_runtime": 132.486,
|
| 70 |
+
"eval_samples_per_second": 14.636,
|
| 71 |
+
"eval_steps_per_second": 1.834,
|
| 72 |
+
"eval_wer": 0.4817837837837838,
|
| 73 |
+
"step": 400
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.6868131868131868,
|
| 77 |
+
"grad_norm": 1.9618090391159058,
|
| 78 |
+
"learning_rate": 0.0002945354185832566,
|
| 79 |
+
"loss": 0.4153,
|
| 80 |
+
"step": 500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.6868131868131868,
|
| 84 |
+
"eval_loss": 0.2935521602630615,
|
| 85 |
+
"eval_runtime": 133.0965,
|
| 86 |
+
"eval_samples_per_second": 14.568,
|
| 87 |
+
"eval_steps_per_second": 1.826,
|
| 88 |
+
"eval_wer": 0.47021621621621623,
|
| 89 |
+
"step": 500
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.8241758241758241,
|
| 93 |
+
"grad_norm": 1.3698606491088867,
|
| 94 |
+
"learning_rate": 0.0002931554737810487,
|
| 95 |
+
"loss": 0.4154,
|
| 96 |
+
"step": 600
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 0.8241758241758241,
|
| 100 |
+
"eval_loss": 0.2883508503437042,
|
| 101 |
+
"eval_runtime": 131.9274,
|
| 102 |
+
"eval_samples_per_second": 14.697,
|
| 103 |
+
"eval_steps_per_second": 1.842,
|
| 104 |
+
"eval_wer": 0.4492972972972973,
|
| 105 |
+
"step": 600
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.9615384615384616,
|
| 109 |
+
"grad_norm": 1.5283702611923218,
|
| 110 |
+
"learning_rate": 0.0002917755289788408,
|
| 111 |
+
"loss": 0.3789,
|
| 112 |
+
"step": 700
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 0.9615384615384616,
|
| 116 |
+
"eval_loss": 0.2806306481361389,
|
| 117 |
+
"eval_runtime": 132.4493,
|
| 118 |
+
"eval_samples_per_second": 14.64,
|
| 119 |
+
"eval_steps_per_second": 1.835,
|
| 120 |
+
"eval_wer": 0.4521081081081081,
|
| 121 |
+
"step": 700
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 1.098901098901099,
|
| 125 |
+
"grad_norm": 0.4610430598258972,
|
| 126 |
+
"learning_rate": 0.0002903955841766329,
|
| 127 |
+
"loss": 0.3667,
|
| 128 |
+
"step": 800
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.098901098901099,
|
| 132 |
+
"eval_loss": 0.2763822078704834,
|
| 133 |
+
"eval_runtime": 133.2255,
|
| 134 |
+
"eval_samples_per_second": 14.554,
|
| 135 |
+
"eval_steps_per_second": 1.824,
|
| 136 |
+
"eval_wer": 0.4351891891891892,
|
| 137 |
+
"step": 800
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 1.2362637362637363,
|
| 141 |
+
"grad_norm": 0.7730151414871216,
|
| 142 |
+
"learning_rate": 0.000289015639374425,
|
| 143 |
+
"loss": 0.3929,
|
| 144 |
+
"step": 900
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 1.2362637362637363,
|
| 148 |
+
"eval_loss": 0.27625423669815063,
|
| 149 |
+
"eval_runtime": 133.7811,
|
| 150 |
+
"eval_samples_per_second": 14.494,
|
| 151 |
+
"eval_steps_per_second": 1.816,
|
| 152 |
+
"eval_wer": 0.46535135135135136,
|
| 153 |
+
"step": 900
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 1.3736263736263736,
|
| 157 |
+
"grad_norm": 0.9868985414505005,
|
| 158 |
+
"learning_rate": 0.0002876356945722171,
|
| 159 |
+
"loss": 0.3847,
|
| 160 |
+
"step": 1000
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 1.3736263736263736,
|
| 164 |
+
"eval_loss": 0.27053016424179077,
|
| 165 |
+
"eval_runtime": 133.1254,
|
| 166 |
+
"eval_samples_per_second": 14.565,
|
| 167 |
+
"eval_steps_per_second": 1.825,
|
| 168 |
+
"eval_wer": 0.44064864864864867,
|
| 169 |
+
"step": 1000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 1.510989010989011,
|
| 173 |
+
"grad_norm": 0.7302649021148682,
|
| 174 |
+
"learning_rate": 0.00028625574977000917,
|
| 175 |
+
"loss": 0.3833,
|
| 176 |
+
"step": 1100
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 1.510989010989011,
|
| 180 |
+
"eval_loss": 0.2696720063686371,
|
| 181 |
+
"eval_runtime": 134.0459,
|
| 182 |
+
"eval_samples_per_second": 14.465,
|
| 183 |
+
"eval_steps_per_second": 1.813,
|
| 184 |
+
"eval_wer": 0.42464864864864865,
|
| 185 |
+
"step": 1100
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 1.6483516483516483,
|
| 189 |
+
"grad_norm": 0.8602890372276306,
|
| 190 |
+
"learning_rate": 0.00028487580496780126,
|
| 191 |
+
"loss": 0.3742,
|
| 192 |
+
"step": 1200
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 1.6483516483516483,
|
| 196 |
+
"eval_loss": 0.2667711675167084,
|
| 197 |
+
"eval_runtime": 133.3019,
|
| 198 |
+
"eval_samples_per_second": 14.546,
|
| 199 |
+
"eval_steps_per_second": 1.823,
|
| 200 |
+
"eval_wer": 0.42502702702702705,
|
| 201 |
+
"step": 1200
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 1.7857142857142856,
|
| 205 |
+
"grad_norm": 1.5730996131896973,
|
| 206 |
+
"learning_rate": 0.00028349586016559336,
|
| 207 |
+
"loss": 0.3694,
|
| 208 |
+
"step": 1300
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 1.7857142857142856,
|
| 212 |
+
"eval_loss": 0.2689613401889801,
|
| 213 |
+
"eval_runtime": 133.41,
|
| 214 |
+
"eval_samples_per_second": 14.534,
|
| 215 |
+
"eval_steps_per_second": 1.821,
|
| 216 |
+
"eval_wer": 0.4189189189189189,
|
| 217 |
+
"step": 1300
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 1.9230769230769231,
|
| 221 |
+
"grad_norm": 1.534987449645996,
|
| 222 |
+
"learning_rate": 0.00028211591536338545,
|
| 223 |
+
"loss": 0.3494,
|
| 224 |
+
"step": 1400
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.9230769230769231,
|
| 228 |
+
"eval_loss": 0.26353445649147034,
|
| 229 |
+
"eval_runtime": 132.3792,
|
| 230 |
+
"eval_samples_per_second": 14.647,
|
| 231 |
+
"eval_steps_per_second": 1.836,
|
| 232 |
+
"eval_wer": 0.416,
|
| 233 |
+
"step": 1400
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 2.0604395604395602,
|
| 237 |
+
"grad_norm": 1.566632866859436,
|
| 238 |
+
"learning_rate": 0.00028073597056117754,
|
| 239 |
+
"loss": 0.3724,
|
| 240 |
+
"step": 1500
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 2.0604395604395602,
|
| 244 |
+
"eval_loss": 0.2625681757926941,
|
| 245 |
+
"eval_runtime": 134.0736,
|
| 246 |
+
"eval_samples_per_second": 14.462,
|
| 247 |
+
"eval_steps_per_second": 1.812,
|
| 248 |
+
"eval_wer": 0.4323243243243243,
|
| 249 |
+
"step": 1500
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 2.197802197802198,
|
| 253 |
+
"grad_norm": 1.1042734384536743,
|
| 254 |
+
"learning_rate": 0.00027935602575896964,
|
| 255 |
+
"loss": 0.3723,
|
| 256 |
+
"step": 1600
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 2.197802197802198,
|
| 260 |
+
"eval_loss": 0.2598220705986023,
|
| 261 |
+
"eval_runtime": 132.9896,
|
| 262 |
+
"eval_samples_per_second": 14.58,
|
| 263 |
+
"eval_steps_per_second": 1.827,
|
| 264 |
+
"eval_wer": 0.4247027027027027,
|
| 265 |
+
"step": 1600
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 2.3351648351648353,
|
| 269 |
+
"grad_norm": 1.444578766822815,
|
| 270 |
+
"learning_rate": 0.00027797608095676173,
|
| 271 |
+
"loss": 0.3505,
|
| 272 |
+
"step": 1700
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 2.3351648351648353,
|
| 276 |
+
"eval_loss": 0.25828292965888977,
|
| 277 |
+
"eval_runtime": 133.6765,
|
| 278 |
+
"eval_samples_per_second": 14.505,
|
| 279 |
+
"eval_steps_per_second": 1.818,
|
| 280 |
+
"eval_wer": 0.412,
|
| 281 |
+
"step": 1700
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 2.4725274725274726,
|
| 285 |
+
"grad_norm": 2.0142714977264404,
|
| 286 |
+
"learning_rate": 0.00027659613615455377,
|
| 287 |
+
"loss": 0.3393,
|
| 288 |
+
"step": 1800
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 2.4725274725274726,
|
| 292 |
+
"eval_loss": 0.2563156187534332,
|
| 293 |
+
"eval_runtime": 132.7739,
|
| 294 |
+
"eval_samples_per_second": 14.604,
|
| 295 |
+
"eval_steps_per_second": 1.83,
|
| 296 |
+
"eval_wer": 0.41281081081081084,
|
| 297 |
+
"step": 1800
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 2.60989010989011,
|
| 301 |
+
"grad_norm": 2.562643527984619,
|
| 302 |
+
"learning_rate": 0.00027521619135234586,
|
| 303 |
+
"loss": 0.3352,
|
| 304 |
+
"step": 1900
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 2.60989010989011,
|
| 308 |
+
"eval_loss": 0.25453999638557434,
|
| 309 |
+
"eval_runtime": 133.7423,
|
| 310 |
+
"eval_samples_per_second": 14.498,
|
| 311 |
+
"eval_steps_per_second": 1.817,
|
| 312 |
+
"eval_wer": 0.4155675675675676,
|
| 313 |
+
"step": 1900
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 2.7472527472527473,
|
| 317 |
+
"grad_norm": 3.292581796646118,
|
| 318 |
+
"learning_rate": 0.00027383624655013796,
|
| 319 |
+
"loss": 0.3516,
|
| 320 |
+
"step": 2000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 2.7472527472527473,
|
| 324 |
+
"eval_loss": 0.25513678789138794,
|
| 325 |
+
"eval_runtime": 133.1058,
|
| 326 |
+
"eval_samples_per_second": 14.567,
|
| 327 |
+
"eval_steps_per_second": 1.826,
|
| 328 |
+
"eval_wer": 0.4315135135135135,
|
| 329 |
+
"step": 2000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 2.8846153846153846,
|
| 333 |
+
"grad_norm": 0.9819747805595398,
|
| 334 |
+
"learning_rate": 0.00027245630174793005,
|
| 335 |
+
"loss": 0.3489,
|
| 336 |
+
"step": 2100
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 2.8846153846153846,
|
| 340 |
+
"eval_loss": 0.256041020154953,
|
| 341 |
+
"eval_runtime": 134.2008,
|
| 342 |
+
"eval_samples_per_second": 14.448,
|
| 343 |
+
"eval_steps_per_second": 1.811,
|
| 344 |
+
"eval_wer": 0.426972972972973,
|
| 345 |
+
"step": 2100
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 3.021978021978022,
|
| 349 |
+
"grad_norm": 0.77928626537323,
|
| 350 |
+
"learning_rate": 0.00027107635694572214,
|
| 351 |
+
"loss": 0.3512,
|
| 352 |
+
"step": 2200
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 3.021978021978022,
|
| 356 |
+
"eval_loss": 0.25362125039100647,
|
| 357 |
+
"eval_runtime": 132.6388,
|
| 358 |
+
"eval_samples_per_second": 14.619,
|
| 359 |
+
"eval_steps_per_second": 1.832,
|
| 360 |
+
"eval_wer": 0.4038918918918919,
|
| 361 |
+
"step": 2200
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 3.159340659340659,
|
| 365 |
+
"grad_norm": 1.5085543394088745,
|
| 366 |
+
"learning_rate": 0.00026971021159153633,
|
| 367 |
+
"loss": 0.339,
|
| 368 |
+
"step": 2300
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 3.159340659340659,
|
| 372 |
+
"eval_loss": 0.24898956716060638,
|
| 373 |
+
"eval_runtime": 133.1617,
|
| 374 |
+
"eval_samples_per_second": 14.561,
|
| 375 |
+
"eval_steps_per_second": 1.825,
|
| 376 |
+
"eval_wer": 0.3988648648648649,
|
| 377 |
+
"step": 2300
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 3.2967032967032965,
|
| 381 |
+
"grad_norm": 0.6166655421257019,
|
| 382 |
+
"learning_rate": 0.0002683302667893284,
|
| 383 |
+
"loss": 0.3374,
|
| 384 |
+
"step": 2400
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 3.2967032967032965,
|
| 388 |
+
"eval_loss": 0.24945656955242157,
|
| 389 |
+
"eval_runtime": 132.9861,
|
| 390 |
+
"eval_samples_per_second": 14.58,
|
| 391 |
+
"eval_steps_per_second": 1.827,
|
| 392 |
+
"eval_wer": 0.39637837837837836,
|
| 393 |
+
"step": 2400
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 3.4340659340659343,
|
| 397 |
+
"grad_norm": 0.9863089919090271,
|
| 398 |
+
"learning_rate": 0.00026695032198712046,
|
| 399 |
+
"loss": 0.3295,
|
| 400 |
+
"step": 2500
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 3.4340659340659343,
|
| 404 |
+
"eval_loss": 0.2518012523651123,
|
| 405 |
+
"eval_runtime": 133.4804,
|
| 406 |
+
"eval_samples_per_second": 14.526,
|
| 407 |
+
"eval_steps_per_second": 1.82,
|
| 408 |
+
"eval_wer": 0.4037297297297297,
|
| 409 |
+
"step": 2500
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 3.571428571428571,
|
| 413 |
+
"grad_norm": 0.5508905649185181,
|
| 414 |
+
"learning_rate": 0.00026557037718491256,
|
| 415 |
+
"loss": 0.3391,
|
| 416 |
+
"step": 2600
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 3.571428571428571,
|
| 420 |
+
"eval_loss": 0.24910376965999603,
|
| 421 |
+
"eval_runtime": 133.409,
|
| 422 |
+
"eval_samples_per_second": 14.534,
|
| 423 |
+
"eval_steps_per_second": 1.821,
|
| 424 |
+
"eval_wer": 0.4077297297297297,
|
| 425 |
+
"step": 2600
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 3.708791208791209,
|
| 429 |
+
"grad_norm": 0.5989585518836975,
|
| 430 |
+
"learning_rate": 0.00026419043238270465,
|
| 431 |
+
"loss": 0.3373,
|
| 432 |
+
"step": 2700
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 3.708791208791209,
|
| 436 |
+
"eval_loss": 0.24448369443416595,
|
| 437 |
+
"eval_runtime": 133.5414,
|
| 438 |
+
"eval_samples_per_second": 14.52,
|
| 439 |
+
"eval_steps_per_second": 1.82,
|
| 440 |
+
"eval_wer": 0.3989189189189189,
|
| 441 |
+
"step": 2700
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 3.8461538461538463,
|
| 445 |
+
"grad_norm": 0.696869969367981,
|
| 446 |
+
"learning_rate": 0.00026281048758049674,
|
| 447 |
+
"loss": 0.3097,
|
| 448 |
+
"step": 2800
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 3.8461538461538463,
|
| 452 |
+
"eval_loss": 0.24623404443264008,
|
| 453 |
+
"eval_runtime": 134.7186,
|
| 454 |
+
"eval_samples_per_second": 14.393,
|
| 455 |
+
"eval_steps_per_second": 1.804,
|
| 456 |
+
"eval_wer": 0.4117837837837838,
|
| 457 |
+
"step": 2800
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 3.9835164835164836,
|
| 461 |
+
"grad_norm": 0.8349074721336365,
|
| 462 |
+
"learning_rate": 0.00026143054277828884,
|
| 463 |
+
"loss": 0.3458,
|
| 464 |
+
"step": 2900
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 3.9835164835164836,
|
| 468 |
+
"eval_loss": 0.24432998895645142,
|
| 469 |
+
"eval_runtime": 133.9009,
|
| 470 |
+
"eval_samples_per_second": 14.481,
|
| 471 |
+
"eval_steps_per_second": 1.815,
|
| 472 |
+
"eval_wer": 0.4034054054054054,
|
| 473 |
+
"step": 2900
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 4.1208791208791204,
|
| 477 |
+
"grad_norm": 0.9415439963340759,
|
| 478 |
+
"learning_rate": 0.00026005059797608093,
|
| 479 |
+
"loss": 0.313,
|
| 480 |
+
"step": 3000
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 4.1208791208791204,
|
| 484 |
+
"eval_loss": 0.24325776100158691,
|
| 485 |
+
"eval_runtime": 133.6886,
|
| 486 |
+
"eval_samples_per_second": 14.504,
|
| 487 |
+
"eval_steps_per_second": 1.818,
|
| 488 |
+
"eval_wer": 0.3882162162162162,
|
| 489 |
+
"step": 3000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 4.258241758241758,
|
| 493 |
+
"grad_norm": 1.5970408916473389,
|
| 494 |
+
"learning_rate": 0.000258670653173873,
|
| 495 |
+
"loss": 0.3171,
|
| 496 |
+
"step": 3100
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 4.258241758241758,
|
| 500 |
+
"eval_loss": 0.24258361756801605,
|
| 501 |
+
"eval_runtime": 135.0574,
|
| 502 |
+
"eval_samples_per_second": 14.357,
|
| 503 |
+
"eval_steps_per_second": 1.799,
|
| 504 |
+
"eval_wer": 0.3968108108108108,
|
| 505 |
+
"step": 3100
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 4.395604395604396,
|
| 509 |
+
"grad_norm": 1.8427647352218628,
|
| 510 |
+
"learning_rate": 0.0002572907083716651,
|
| 511 |
+
"loss": 0.3122,
|
| 512 |
+
"step": 3200
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 4.395604395604396,
|
| 516 |
+
"eval_loss": 0.2430083155632019,
|
| 517 |
+
"eval_runtime": 133.0857,
|
| 518 |
+
"eval_samples_per_second": 14.57,
|
| 519 |
+
"eval_steps_per_second": 1.826,
|
| 520 |
+
"eval_wer": 0.39356756756756756,
|
| 521 |
+
"step": 3200
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 4.532967032967033,
|
| 525 |
+
"grad_norm": 0.6071587800979614,
|
| 526 |
+
"learning_rate": 0.0002559107635694572,
|
| 527 |
+
"loss": 0.3255,
|
| 528 |
+
"step": 3300
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 4.532967032967033,
|
| 532 |
+
"eval_loss": 0.24039362370967865,
|
| 533 |
+
"eval_runtime": 133.7381,
|
| 534 |
+
"eval_samples_per_second": 14.498,
|
| 535 |
+
"eval_steps_per_second": 1.817,
|
| 536 |
+
"eval_wer": 0.38216216216216214,
|
| 537 |
+
"step": 3300
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 4.670329670329671,
|
| 541 |
+
"grad_norm": 0.9580681324005127,
|
| 542 |
+
"learning_rate": 0.0002545308187672493,
|
| 543 |
+
"loss": 0.3253,
|
| 544 |
+
"step": 3400
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"epoch": 4.670329670329671,
|
| 548 |
+
"eval_loss": 0.2355840504169464,
|
| 549 |
+
"eval_runtime": 132.9989,
|
| 550 |
+
"eval_samples_per_second": 14.579,
|
| 551 |
+
"eval_steps_per_second": 1.827,
|
| 552 |
+
"eval_wer": 0.3945945945945946,
|
| 553 |
+
"step": 3400
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 4.8076923076923075,
|
| 557 |
+
"grad_norm": 1.4110567569732666,
|
| 558 |
+
"learning_rate": 0.0002531508739650414,
|
| 559 |
+
"loss": 0.3341,
|
| 560 |
+
"step": 3500
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 4.8076923076923075,
|
| 564 |
+
"eval_loss": 0.23690326511859894,
|
| 565 |
+
"eval_runtime": 133.2507,
|
| 566 |
+
"eval_samples_per_second": 14.552,
|
| 567 |
+
"eval_steps_per_second": 1.824,
|
| 568 |
+
"eval_wer": 0.3872432432432432,
|
| 569 |
+
"step": 3500
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 4.945054945054945,
|
| 573 |
+
"grad_norm": 1.9591115713119507,
|
| 574 |
+
"learning_rate": 0.0002517709291628335,
|
| 575 |
+
"loss": 0.3183,
|
| 576 |
+
"step": 3600
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 4.945054945054945,
|
| 580 |
+
"eval_loss": 0.2345106154680252,
|
| 581 |
+
"eval_runtime": 133.4011,
|
| 582 |
+
"eval_samples_per_second": 14.535,
|
| 583 |
+
"eval_steps_per_second": 1.822,
|
| 584 |
+
"eval_wer": 0.38535135135135135,
|
| 585 |
+
"step": 3600
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 5.082417582417582,
|
| 589 |
+
"grad_norm": 0.616919219493866,
|
| 590 |
+
"learning_rate": 0.0002503909843606256,
|
| 591 |
+
"loss": 0.3461,
|
| 592 |
+
"step": 3700
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 5.082417582417582,
|
| 596 |
+
"eval_loss": 0.23949113488197327,
|
| 597 |
+
"eval_runtime": 133.8985,
|
| 598 |
+
"eval_samples_per_second": 14.481,
|
| 599 |
+
"eval_steps_per_second": 1.815,
|
| 600 |
+
"eval_wer": 0.3828108108108108,
|
| 601 |
+
"step": 3700
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 5.21978021978022,
|
| 605 |
+
"grad_norm": 0.7555241584777832,
|
| 606 |
+
"learning_rate": 0.0002490110395584176,
|
| 607 |
+
"loss": 0.3147,
|
| 608 |
+
"step": 3800
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 5.21978021978022,
|
| 612 |
+
"eval_loss": 0.23591694235801697,
|
| 613 |
+
"eval_runtime": 133.3437,
|
| 614 |
+
"eval_samples_per_second": 14.541,
|
| 615 |
+
"eval_steps_per_second": 1.822,
|
| 616 |
+
"eval_wer": 0.37751351351351353,
|
| 617 |
+
"step": 3800
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 5.357142857142857,
|
| 621 |
+
"grad_norm": 0.6169973015785217,
|
| 622 |
+
"learning_rate": 0.0002476310947562097,
|
| 623 |
+
"loss": 0.317,
|
| 624 |
+
"step": 3900
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 5.357142857142857,
|
| 628 |
+
"eval_loss": 0.2320287972688675,
|
| 629 |
+
"eval_runtime": 134.295,
|
| 630 |
+
"eval_samples_per_second": 14.438,
|
| 631 |
+
"eval_steps_per_second": 1.809,
|
| 632 |
+
"eval_wer": 0.38075675675675674,
|
| 633 |
+
"step": 3900
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 5.4945054945054945,
|
| 637 |
+
"grad_norm": 0.44300732016563416,
|
| 638 |
+
"learning_rate": 0.0002462511499540018,
|
| 639 |
+
"loss": 0.3094,
|
| 640 |
+
"step": 4000
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 5.4945054945054945,
|
| 644 |
+
"eval_loss": 0.23655937612056732,
|
| 645 |
+
"eval_runtime": 134.695,
|
| 646 |
+
"eval_samples_per_second": 14.395,
|
| 647 |
+
"eval_steps_per_second": 1.804,
|
| 648 |
+
"eval_wer": 0.37972972972972974,
|
| 649 |
+
"step": 4000
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 5.631868131868131,
|
| 653 |
+
"grad_norm": 1.5346554517745972,
|
| 654 |
+
"learning_rate": 0.0002448712051517939,
|
| 655 |
+
"loss": 0.2913,
|
| 656 |
+
"step": 4100
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 5.631868131868131,
|
| 660 |
+
"eval_loss": 0.2357349544763565,
|
| 661 |
+
"eval_runtime": 134.0434,
|
| 662 |
+
"eval_samples_per_second": 14.465,
|
| 663 |
+
"eval_steps_per_second": 1.813,
|
| 664 |
+
"eval_wer": 0.37491891891891893,
|
| 665 |
+
"step": 4100
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 5.769230769230769,
|
| 669 |
+
"grad_norm": 0.7017680406570435,
|
| 670 |
+
"learning_rate": 0.000243491260349586,
|
| 671 |
+
"loss": 0.3195,
|
| 672 |
+
"step": 4200
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 5.769230769230769,
|
| 676 |
+
"eval_loss": 0.23318681120872498,
|
| 677 |
+
"eval_runtime": 133.1069,
|
| 678 |
+
"eval_samples_per_second": 14.567,
|
| 679 |
+
"eval_steps_per_second": 1.826,
|
| 680 |
+
"eval_wer": 0.3694054054054054,
|
| 681 |
+
"step": 4200
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 5.906593406593407,
|
| 685 |
+
"grad_norm": 0.46099045872688293,
|
| 686 |
+
"learning_rate": 0.00024212511499540016,
|
| 687 |
+
"loss": 0.3189,
|
| 688 |
+
"step": 4300
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 5.906593406593407,
|
| 692 |
+
"eval_loss": 0.23127850890159607,
|
| 693 |
+
"eval_runtime": 135.2118,
|
| 694 |
+
"eval_samples_per_second": 14.34,
|
| 695 |
+
"eval_steps_per_second": 1.797,
|
| 696 |
+
"eval_wer": 0.38697297297297295,
|
| 697 |
+
"step": 4300
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 6.043956043956044,
|
| 701 |
+
"grad_norm": 1.010475516319275,
|
| 702 |
+
"learning_rate": 0.00024074517019319225,
|
| 703 |
+
"loss": 0.3105,
|
| 704 |
+
"step": 4400
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 6.043956043956044,
|
| 708 |
+
"eval_loss": 0.23260627686977386,
|
| 709 |
+
"eval_runtime": 134.6368,
|
| 710 |
+
"eval_samples_per_second": 14.402,
|
| 711 |
+
"eval_steps_per_second": 1.805,
|
| 712 |
+
"eval_wer": 0.38064864864864867,
|
| 713 |
+
"step": 4400
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 6.181318681318682,
|
| 717 |
+
"grad_norm": 0.7957504391670227,
|
| 718 |
+
"learning_rate": 0.00023936522539098434,
|
| 719 |
+
"loss": 0.2937,
|
| 720 |
+
"step": 4500
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 6.181318681318682,
|
| 724 |
+
"eval_loss": 0.23464266955852509,
|
| 725 |
+
"eval_runtime": 134.508,
|
| 726 |
+
"eval_samples_per_second": 14.415,
|
| 727 |
+
"eval_steps_per_second": 1.807,
|
| 728 |
+
"eval_wer": 0.3783783783783784,
|
| 729 |
+
"step": 4500
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 6.318681318681318,
|
| 733 |
+
"grad_norm": 0.73511803150177,
|
| 734 |
+
"learning_rate": 0.00023798528058877644,
|
| 735 |
+
"loss": 0.3088,
|
| 736 |
+
"step": 4600
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 6.318681318681318,
|
| 740 |
+
"eval_loss": 0.2313350886106491,
|
| 741 |
+
"eval_runtime": 135.48,
|
| 742 |
+
"eval_samples_per_second": 14.312,
|
| 743 |
+
"eval_steps_per_second": 1.794,
|
| 744 |
+
"eval_wer": 0.3725945945945946,
|
| 745 |
+
"step": 4600
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 6.456043956043956,
|
| 749 |
+
"grad_norm": 0.587340772151947,
|
| 750 |
+
"learning_rate": 0.00023660533578656853,
|
| 751 |
+
"loss": 0.2852,
|
| 752 |
+
"step": 4700
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 6.456043956043956,
|
| 756 |
+
"eval_loss": 0.2307436466217041,
|
| 757 |
+
"eval_runtime": 135.4353,
|
| 758 |
+
"eval_samples_per_second": 14.317,
|
| 759 |
+
"eval_steps_per_second": 1.794,
|
| 760 |
+
"eval_wer": 0.37086486486486486,
|
| 761 |
+
"step": 4700
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 6.593406593406593,
|
| 765 |
+
"grad_norm": 0.5956315994262695,
|
| 766 |
+
"learning_rate": 0.0002352253909843606,
|
| 767 |
+
"loss": 0.3083,
|
| 768 |
+
"step": 4800
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 6.593406593406593,
|
| 772 |
+
"eval_loss": 0.22929497063159943,
|
| 773 |
+
"eval_runtime": 134.2045,
|
| 774 |
+
"eval_samples_per_second": 14.448,
|
| 775 |
+
"eval_steps_per_second": 1.811,
|
| 776 |
+
"eval_wer": 0.3751891891891892,
|
| 777 |
+
"step": 4800
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 6.730769230769231,
|
| 781 |
+
"grad_norm": 1.3765567541122437,
|
| 782 |
+
"learning_rate": 0.0002338454461821527,
|
| 783 |
+
"loss": 0.3194,
|
| 784 |
+
"step": 4900
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 6.730769230769231,
|
| 788 |
+
"eval_loss": 0.22924266755580902,
|
| 789 |
+
"eval_runtime": 135.6964,
|
| 790 |
+
"eval_samples_per_second": 14.289,
|
| 791 |
+
"eval_steps_per_second": 1.791,
|
| 792 |
+
"eval_wer": 0.37108108108108107,
|
| 793 |
+
"step": 4900
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 6.868131868131869,
|
| 797 |
+
"grad_norm": 0.7330417037010193,
|
| 798 |
+
"learning_rate": 0.00023246550137994478,
|
| 799 |
+
"loss": 0.297,
|
| 800 |
+
"step": 5000
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 6.868131868131869,
|
| 804 |
+
"eval_loss": 0.2303166538476944,
|
| 805 |
+
"eval_runtime": 134.3817,
|
| 806 |
+
"eval_samples_per_second": 14.429,
|
| 807 |
+
"eval_steps_per_second": 1.808,
|
| 808 |
+
"eval_wer": 0.3715135135135135,
|
| 809 |
+
"step": 5000
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 7.0054945054945055,
|
| 813 |
+
"grad_norm": 0.9262694716453552,
|
| 814 |
+
"learning_rate": 0.00023108555657773688,
|
| 815 |
+
"loss": 0.3086,
|
| 816 |
+
"step": 5100
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 7.0054945054945055,
|
| 820 |
+
"eval_loss": 0.23395021259784698,
|
| 821 |
+
"eval_runtime": 134.6683,
|
| 822 |
+
"eval_samples_per_second": 14.398,
|
| 823 |
+
"eval_steps_per_second": 1.804,
|
| 824 |
+
"eval_wer": 0.4027027027027027,
|
| 825 |
+
"step": 5100
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 7.142857142857143,
|
| 829 |
+
"grad_norm": 0.5672425627708435,
|
| 830 |
+
"learning_rate": 0.00022970561177552897,
|
| 831 |
+
"loss": 0.3058,
|
| 832 |
+
"step": 5200
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 7.142857142857143,
|
| 836 |
+
"eval_loss": 0.22942779958248138,
|
| 837 |
+
"eval_runtime": 133.8761,
|
| 838 |
+
"eval_samples_per_second": 14.484,
|
| 839 |
+
"eval_steps_per_second": 1.815,
|
| 840 |
+
"eval_wer": 0.36648648648648646,
|
| 841 |
+
"step": 5200
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 7.142857142857143,
|
| 845 |
+
"step": 5200,
|
| 846 |
+
"total_flos": 2.4864489394936607e+19,
|
| 847 |
+
"train_loss": 0.46319753646850587,
|
| 848 |
+
"train_runtime": 13134.8195,
|
| 849 |
+
"train_samples_per_second": 13.286,
|
| 850 |
+
"train_steps_per_second": 1.663
|
| 851 |
+
}
|
| 852 |
+
],
|
| 853 |
+
"logging_steps": 100,
|
| 854 |
+
"max_steps": 21840,
|
| 855 |
+
"num_input_tokens_seen": 0,
|
| 856 |
+
"num_train_epochs": 30,
|
| 857 |
+
"save_steps": 200,
|
| 858 |
+
"stateful_callbacks": {
|
| 859 |
+
"EarlyStoppingCallback": {
|
| 860 |
+
"args": {
|
| 861 |
+
"early_stopping_patience": 3,
|
| 862 |
+
"early_stopping_threshold": 0.0
|
| 863 |
+
},
|
| 864 |
+
"attributes": {
|
| 865 |
+
"early_stopping_patience_counter": 3
|
| 866 |
+
}
|
| 867 |
+
},
|
| 868 |
+
"TrainerControl": {
|
| 869 |
+
"args": {
|
| 870 |
+
"should_epoch_stop": false,
|
| 871 |
+
"should_evaluate": false,
|
| 872 |
+
"should_log": false,
|
| 873 |
+
"should_save": true,
|
| 874 |
+
"should_training_stop": true
|
| 875 |
+
},
|
| 876 |
+
"attributes": {}
|
| 877 |
+
}
|
| 878 |
+
},
|
| 879 |
+
"total_flos": 2.4864489394936607e+19,
|
| 880 |
+
"train_batch_size": 8,
|
| 881 |
+
"trial_name": null,
|
| 882 |
+
"trial_params": null
|
| 883 |
+
}
|