End of training
Browse files- README.md +6 -3
- adapter.bem.safetensors +3 -0
- all_results.json +15 -0
- eval_results.json +9 -0
- runs/Feb09_10-57-50_srvrocgpu011.uct.ac.za/events.out.tfevents.1739097140.srvrocgpu011.uct.ac.za +3 -0
- train_results.json +9 -0
- trainer_state.json +627 -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-bigcgen-baseline-model
|
| 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 |
+
- bigcgen
|
| 8 |
+
- mms
|
| 9 |
- generated_from_trainer
|
| 10 |
metrics:
|
| 11 |
- wer
|
|
|
|
| 19 |
|
| 20 |
# mms-1b-bigcgen-baseline-model
|
| 21 |
|
| 22 |
+
This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on the BIGCGEN - BEM dataset.
|
| 23 |
It achieves the following results on the evaluation set:
|
| 24 |
+
- Loss: 0.5089
|
| 25 |
+
- Wer: 0.4580
|
| 26 |
|
| 27 |
## Model description
|
| 28 |
|
adapter.bem.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:84acbeac47beef7a996135af07853f6dfc92b10798b8de2a43222d56aa67fb63
|
| 3 |
+
size 8788284
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 11.009174311926605,
|
| 3 |
+
"eval_loss": 0.5089433789253235,
|
| 4 |
+
"eval_runtime": 31.5655,
|
| 5 |
+
"eval_samples": 456,
|
| 6 |
+
"eval_samples_per_second": 14.446,
|
| 7 |
+
"eval_steps_per_second": 3.612,
|
| 8 |
+
"eval_wer": 0.45802790257744147,
|
| 9 |
+
"total_flos": 1.8690129169621533e+19,
|
| 10 |
+
"train_loss": 1.669767551422119,
|
| 11 |
+
"train_runtime": 5274.2639,
|
| 12 |
+
"train_samples": 2615,
|
| 13 |
+
"train_samples_per_second": 14.874,
|
| 14 |
+
"train_steps_per_second": 1.86
|
| 15 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 11.009174311926605,
|
| 3 |
+
"eval_loss": 0.5089433789253235,
|
| 4 |
+
"eval_runtime": 31.5655,
|
| 5 |
+
"eval_samples": 456,
|
| 6 |
+
"eval_samples_per_second": 14.446,
|
| 7 |
+
"eval_steps_per_second": 3.612,
|
| 8 |
+
"eval_wer": 0.45802790257744147
|
| 9 |
+
}
|
runs/Feb09_10-57-50_srvrocgpu011.uct.ac.za/events.out.tfevents.1739097140.srvrocgpu011.uct.ac.za
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0319773aeae5d8969754e8019374f86b3c2653b50d4c8c7def076ca29f5df4b6
|
| 3 |
+
size 40
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 11.009174311926605,
|
| 3 |
+
"total_flos": 1.8690129169621533e+19,
|
| 4 |
+
"train_loss": 1.669767551422119,
|
| 5 |
+
"train_runtime": 5274.2639,
|
| 6 |
+
"train_samples": 2615,
|
| 7 |
+
"train_samples_per_second": 14.874,
|
| 8 |
+
"train_steps_per_second": 1.86
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,627 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.508959949016571,
|
| 3 |
+
"best_model_checkpoint": "/scratch/skscla001/speech/results/mms-1b-bigcgen-baseline-model/checkpoint-3200",
|
| 4 |
+
"epoch": 11.009174311926605,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 3600,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.3058103975535168,
|
| 13 |
+
"grad_norm": 6.544164657592773,
|
| 14 |
+
"learning_rate": 0.000285,
|
| 15 |
+
"loss": 12.3367,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.3058103975535168,
|
| 20 |
+
"eval_loss": 1.3241018056869507,
|
| 21 |
+
"eval_runtime": 33.411,
|
| 22 |
+
"eval_samples_per_second": 13.648,
|
| 23 |
+
"eval_steps_per_second": 3.412,
|
| 24 |
+
"eval_wer": 0.8893355403168598,
|
| 25 |
+
"step": 100
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.6116207951070336,
|
| 29 |
+
"grad_norm": 25.74690818786621,
|
| 30 |
+
"learning_rate": 0.00029706488156539646,
|
| 31 |
+
"loss": 1.8764,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.6116207951070336,
|
| 36 |
+
"eval_loss": 0.689376950263977,
|
| 37 |
+
"eval_runtime": 31.5961,
|
| 38 |
+
"eval_samples_per_second": 14.432,
|
| 39 |
+
"eval_steps_per_second": 3.608,
|
| 40 |
+
"eval_wer": 0.5814613383778671,
|
| 41 |
+
"step": 200
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.9174311926605505,
|
| 45 |
+
"grad_norm": 6.327740669250488,
|
| 46 |
+
"learning_rate": 0.00029397528321318227,
|
| 47 |
+
"loss": 1.6712,
|
| 48 |
+
"step": 300
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 0.9174311926605505,
|
| 52 |
+
"eval_loss": 0.6389656662940979,
|
| 53 |
+
"eval_runtime": 31.4047,
|
| 54 |
+
"eval_samples_per_second": 14.52,
|
| 55 |
+
"eval_steps_per_second": 3.63,
|
| 56 |
+
"eval_wer": 0.5514305982501774,
|
| 57 |
+
"step": 300
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 1.2232415902140672,
|
| 61 |
+
"grad_norm": 1.2617219686508179,
|
| 62 |
+
"learning_rate": 0.00029088568486096807,
|
| 63 |
+
"loss": 1.5044,
|
| 64 |
+
"step": 400
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 1.2232415902140672,
|
| 68 |
+
"eval_loss": 0.6300708651542664,
|
| 69 |
+
"eval_runtime": 31.4218,
|
| 70 |
+
"eval_samples_per_second": 14.512,
|
| 71 |
+
"eval_steps_per_second": 3.628,
|
| 72 |
+
"eval_wer": 0.5351146843225348,
|
| 73 |
+
"step": 400
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 1.529051987767584,
|
| 77 |
+
"grad_norm": 2.4596145153045654,
|
| 78 |
+
"learning_rate": 0.0002877960865087538,
|
| 79 |
+
"loss": 1.6648,
|
| 80 |
+
"step": 500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 1.529051987767584,
|
| 84 |
+
"eval_loss": 0.6076481342315674,
|
| 85 |
+
"eval_runtime": 31.6805,
|
| 86 |
+
"eval_samples_per_second": 14.394,
|
| 87 |
+
"eval_steps_per_second": 3.598,
|
| 88 |
+
"eval_wer": 0.5282572712225112,
|
| 89 |
+
"step": 500
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 1.834862385321101,
|
| 93 |
+
"grad_norm": 3.1689512729644775,
|
| 94 |
+
"learning_rate": 0.0002847064881565396,
|
| 95 |
+
"loss": 1.6411,
|
| 96 |
+
"step": 600
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.834862385321101,
|
| 100 |
+
"eval_loss": 0.6073117256164551,
|
| 101 |
+
"eval_runtime": 31.4984,
|
| 102 |
+
"eval_samples_per_second": 14.477,
|
| 103 |
+
"eval_steps_per_second": 3.619,
|
| 104 |
+
"eval_wer": 0.5282572712225112,
|
| 105 |
+
"step": 600
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 2.140672782874618,
|
| 109 |
+
"grad_norm": 3.601820468902588,
|
| 110 |
+
"learning_rate": 0.00028161688980432543,
|
| 111 |
+
"loss": 1.4016,
|
| 112 |
+
"step": 700
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.140672782874618,
|
| 116 |
+
"eval_loss": 0.5994192361831665,
|
| 117 |
+
"eval_runtime": 31.3579,
|
| 118 |
+
"eval_samples_per_second": 14.542,
|
| 119 |
+
"eval_steps_per_second": 3.635,
|
| 120 |
+
"eval_wer": 0.5124142823362497,
|
| 121 |
+
"step": 700
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 2.4464831804281344,
|
| 125 |
+
"grad_norm": 1.8951566219329834,
|
| 126 |
+
"learning_rate": 0.0002785272914521112,
|
| 127 |
+
"loss": 1.5703,
|
| 128 |
+
"step": 800
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 2.4464831804281344,
|
| 132 |
+
"eval_loss": 0.5996993184089661,
|
| 133 |
+
"eval_runtime": 31.3279,
|
| 134 |
+
"eval_samples_per_second": 14.556,
|
| 135 |
+
"eval_steps_per_second": 3.639,
|
| 136 |
+
"eval_wer": 0.5161976826672973,
|
| 137 |
+
"step": 800
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.7522935779816513,
|
| 141 |
+
"grad_norm": 2.722160577774048,
|
| 142 |
+
"learning_rate": 0.00027546858908341916,
|
| 143 |
+
"loss": 1.4165,
|
| 144 |
+
"step": 900
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.7522935779816513,
|
| 148 |
+
"eval_loss": 0.5849923491477966,
|
| 149 |
+
"eval_runtime": 31.6108,
|
| 150 |
+
"eval_samples_per_second": 14.425,
|
| 151 |
+
"eval_steps_per_second": 3.606,
|
| 152 |
+
"eval_wer": 0.5083944194845117,
|
| 153 |
+
"step": 900
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 3.058103975535168,
|
| 157 |
+
"grad_norm": 6.126498699188232,
|
| 158 |
+
"learning_rate": 0.0002723789907312049,
|
| 159 |
+
"loss": 1.4703,
|
| 160 |
+
"step": 1000
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 3.058103975535168,
|
| 164 |
+
"eval_loss": 0.5912389755249023,
|
| 165 |
+
"eval_runtime": 31.5441,
|
| 166 |
+
"eval_samples_per_second": 14.456,
|
| 167 |
+
"eval_steps_per_second": 3.614,
|
| 168 |
+
"eval_wer": 0.5126507448569402,
|
| 169 |
+
"step": 1000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 3.363914373088685,
|
| 173 |
+
"grad_norm": 2.3694710731506348,
|
| 174 |
+
"learning_rate": 0.0002692893923789907,
|
| 175 |
+
"loss": 1.48,
|
| 176 |
+
"step": 1100
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.363914373088685,
|
| 180 |
+
"eval_loss": 0.570675253868103,
|
| 181 |
+
"eval_runtime": 31.3265,
|
| 182 |
+
"eval_samples_per_second": 14.556,
|
| 183 |
+
"eval_steps_per_second": 3.639,
|
| 184 |
+
"eval_wer": 0.4998817687396548,
|
| 185 |
+
"step": 1100
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 3.669724770642202,
|
| 189 |
+
"grad_norm": 2.2272789478302,
|
| 190 |
+
"learning_rate": 0.0002661997940267765,
|
| 191 |
+
"loss": 1.4769,
|
| 192 |
+
"step": 1200
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 3.669724770642202,
|
| 196 |
+
"eval_loss": 0.5674562454223633,
|
| 197 |
+
"eval_runtime": 31.4068,
|
| 198 |
+
"eval_samples_per_second": 14.519,
|
| 199 |
+
"eval_steps_per_second": 3.63,
|
| 200 |
+
"eval_wer": 0.4949160558051549,
|
| 201 |
+
"step": 1200
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 3.9755351681957185,
|
| 205 |
+
"grad_norm": 1.9602124691009521,
|
| 206 |
+
"learning_rate": 0.0002631101956745623,
|
| 207 |
+
"loss": 1.312,
|
| 208 |
+
"step": 1300
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 3.9755351681957185,
|
| 212 |
+
"eval_loss": 0.585631787776947,
|
| 213 |
+
"eval_runtime": 31.7069,
|
| 214 |
+
"eval_samples_per_second": 14.382,
|
| 215 |
+
"eval_steps_per_second": 3.595,
|
| 216 |
+
"eval_wer": 0.497990068574131,
|
| 217 |
+
"step": 1300
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 4.281345565749236,
|
| 221 |
+
"grad_norm": 2.401339054107666,
|
| 222 |
+
"learning_rate": 0.00026002059732234807,
|
| 223 |
+
"loss": 1.3821,
|
| 224 |
+
"step": 1400
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 4.281345565749236,
|
| 228 |
+
"eval_loss": 0.5641556978225708,
|
| 229 |
+
"eval_runtime": 31.5622,
|
| 230 |
+
"eval_samples_per_second": 14.448,
|
| 231 |
+
"eval_steps_per_second": 3.612,
|
| 232 |
+
"eval_wer": 0.49917238117758334,
|
| 233 |
+
"step": 1400
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 4.587155963302752,
|
| 237 |
+
"grad_norm": 3.3140265941619873,
|
| 238 |
+
"learning_rate": 0.0002569309989701339,
|
| 239 |
+
"loss": 1.457,
|
| 240 |
+
"step": 1500
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 4.587155963302752,
|
| 244 |
+
"eval_loss": 0.558822751045227,
|
| 245 |
+
"eval_runtime": 31.4842,
|
| 246 |
+
"eval_samples_per_second": 14.483,
|
| 247 |
+
"eval_steps_per_second": 3.621,
|
| 248 |
+
"eval_wer": 0.5053204067155356,
|
| 249 |
+
"step": 1500
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 4.892966360856269,
|
| 253 |
+
"grad_norm": 1.205606460571289,
|
| 254 |
+
"learning_rate": 0.00025384140061791963,
|
| 255 |
+
"loss": 1.3606,
|
| 256 |
+
"step": 1600
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 4.892966360856269,
|
| 260 |
+
"eval_loss": 0.5637312531471252,
|
| 261 |
+
"eval_runtime": 31.5743,
|
| 262 |
+
"eval_samples_per_second": 14.442,
|
| 263 |
+
"eval_steps_per_second": 3.611,
|
| 264 |
+
"eval_wer": 0.48663986758098843,
|
| 265 |
+
"step": 1600
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 5.198776758409786,
|
| 269 |
+
"grad_norm": 5.104424953460693,
|
| 270 |
+
"learning_rate": 0.00025075180226570543,
|
| 271 |
+
"loss": 1.3986,
|
| 272 |
+
"step": 1700
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 5.198776758409786,
|
| 276 |
+
"eval_loss": 0.5511406064033508,
|
| 277 |
+
"eval_runtime": 31.6463,
|
| 278 |
+
"eval_samples_per_second": 14.409,
|
| 279 |
+
"eval_steps_per_second": 3.602,
|
| 280 |
+
"eval_wer": 0.48663986758098843,
|
| 281 |
+
"step": 1700
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 5.504587155963303,
|
| 285 |
+
"grad_norm": 2.85310435295105,
|
| 286 |
+
"learning_rate": 0.00024766220391349124,
|
| 287 |
+
"loss": 1.421,
|
| 288 |
+
"step": 1800
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 5.504587155963303,
|
| 292 |
+
"eval_loss": 0.5845613479614258,
|
| 293 |
+
"eval_runtime": 31.614,
|
| 294 |
+
"eval_samples_per_second": 14.424,
|
| 295 |
+
"eval_steps_per_second": 3.606,
|
| 296 |
+
"eval_wer": 0.534641759281154,
|
| 297 |
+
"step": 1800
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 5.81039755351682,
|
| 301 |
+
"grad_norm": 1.2034293413162231,
|
| 302 |
+
"learning_rate": 0.00024457260556127704,
|
| 303 |
+
"loss": 1.3004,
|
| 304 |
+
"step": 1900
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 5.81039755351682,
|
| 308 |
+
"eval_loss": 0.5440376400947571,
|
| 309 |
+
"eval_runtime": 31.6052,
|
| 310 |
+
"eval_samples_per_second": 14.428,
|
| 311 |
+
"eval_steps_per_second": 3.607,
|
| 312 |
+
"eval_wer": 0.47363442894301255,
|
| 313 |
+
"step": 1900
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 6.116207951070336,
|
| 317 |
+
"grad_norm": 5.384260654449463,
|
| 318 |
+
"learning_rate": 0.0002414830072090628,
|
| 319 |
+
"loss": 1.3319,
|
| 320 |
+
"step": 2000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 6.116207951070336,
|
| 324 |
+
"eval_loss": 0.531833291053772,
|
| 325 |
+
"eval_runtime": 31.4837,
|
| 326 |
+
"eval_samples_per_second": 14.484,
|
| 327 |
+
"eval_steps_per_second": 3.621,
|
| 328 |
+
"eval_wer": 0.47860014187751243,
|
| 329 |
+
"step": 2000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 6.422018348623853,
|
| 333 |
+
"grad_norm": 1.9580897092819214,
|
| 334 |
+
"learning_rate": 0.0002383934088568486,
|
| 335 |
+
"loss": 1.2665,
|
| 336 |
+
"step": 2100
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 6.422018348623853,
|
| 340 |
+
"eval_loss": 0.5487645268440247,
|
| 341 |
+
"eval_runtime": 31.7917,
|
| 342 |
+
"eval_samples_per_second": 14.343,
|
| 343 |
+
"eval_steps_per_second": 3.586,
|
| 344 |
+
"eval_wer": 0.5065027193189879,
|
| 345 |
+
"step": 2100
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 6.72782874617737,
|
| 349 |
+
"grad_norm": 2.0100314617156982,
|
| 350 |
+
"learning_rate": 0.00023530381050463438,
|
| 351 |
+
"loss": 1.3703,
|
| 352 |
+
"step": 2200
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 6.72782874617737,
|
| 356 |
+
"eval_loss": 0.53043133020401,
|
| 357 |
+
"eval_runtime": 31.6267,
|
| 358 |
+
"eval_samples_per_second": 14.418,
|
| 359 |
+
"eval_steps_per_second": 3.605,
|
| 360 |
+
"eval_wer": 0.4878221801844408,
|
| 361 |
+
"step": 2200
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 7.033639143730887,
|
| 365 |
+
"grad_norm": 2.1251752376556396,
|
| 366 |
+
"learning_rate": 0.00023221421215242018,
|
| 367 |
+
"loss": 1.1954,
|
| 368 |
+
"step": 2300
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 7.033639143730887,
|
| 372 |
+
"eval_loss": 0.5297590494155884,
|
| 373 |
+
"eval_runtime": 31.5678,
|
| 374 |
+
"eval_samples_per_second": 14.445,
|
| 375 |
+
"eval_steps_per_second": 3.611,
|
| 376 |
+
"eval_wer": 0.48072830456372667,
|
| 377 |
+
"step": 2300
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 7.339449541284404,
|
| 381 |
+
"grad_norm": 3.3306617736816406,
|
| 382 |
+
"learning_rate": 0.00022912461380020596,
|
| 383 |
+
"loss": 1.2973,
|
| 384 |
+
"step": 2400
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 7.339449541284404,
|
| 388 |
+
"eval_loss": 0.5258393883705139,
|
| 389 |
+
"eval_runtime": 31.6969,
|
| 390 |
+
"eval_samples_per_second": 14.386,
|
| 391 |
+
"eval_steps_per_second": 3.597,
|
| 392 |
+
"eval_wer": 0.47056041617403643,
|
| 393 |
+
"step": 2400
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 7.6452599388379205,
|
| 397 |
+
"grad_norm": 1.3215351104736328,
|
| 398 |
+
"learning_rate": 0.00022603501544799176,
|
| 399 |
+
"loss": 1.2086,
|
| 400 |
+
"step": 2500
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 7.6452599388379205,
|
| 404 |
+
"eval_loss": 0.5230885744094849,
|
| 405 |
+
"eval_runtime": 31.7287,
|
| 406 |
+
"eval_samples_per_second": 14.372,
|
| 407 |
+
"eval_steps_per_second": 3.593,
|
| 408 |
+
"eval_wer": 0.48072830456372667,
|
| 409 |
+
"step": 2500
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 7.951070336391437,
|
| 413 |
+
"grad_norm": 7.100344181060791,
|
| 414 |
+
"learning_rate": 0.0002229454170957775,
|
| 415 |
+
"loss": 1.2796,
|
| 416 |
+
"step": 2600
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 7.951070336391437,
|
| 420 |
+
"eval_loss": 0.5403640866279602,
|
| 421 |
+
"eval_runtime": 31.7077,
|
| 422 |
+
"eval_samples_per_second": 14.381,
|
| 423 |
+
"eval_steps_per_second": 3.595,
|
| 424 |
+
"eval_wer": 0.473870891463703,
|
| 425 |
+
"step": 2600
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 8.256880733944953,
|
| 429 |
+
"grad_norm": 2.845608949661255,
|
| 430 |
+
"learning_rate": 0.00021985581874356332,
|
| 431 |
+
"loss": 1.1428,
|
| 432 |
+
"step": 2700
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 8.256880733944953,
|
| 436 |
+
"eval_loss": 0.5328220725059509,
|
| 437 |
+
"eval_runtime": 31.486,
|
| 438 |
+
"eval_samples_per_second": 14.483,
|
| 439 |
+
"eval_steps_per_second": 3.621,
|
| 440 |
+
"eval_wer": 0.48309292977063134,
|
| 441 |
+
"step": 2700
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 8.562691131498472,
|
| 445 |
+
"grad_norm": 2.0753355026245117,
|
| 446 |
+
"learning_rate": 0.0002167662203913491,
|
| 447 |
+
"loss": 1.3118,
|
| 448 |
+
"step": 2800
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 8.562691131498472,
|
| 452 |
+
"eval_loss": 0.5198370814323425,
|
| 453 |
+
"eval_runtime": 31.5982,
|
| 454 |
+
"eval_samples_per_second": 14.431,
|
| 455 |
+
"eval_steps_per_second": 3.608,
|
| 456 |
+
"eval_wer": 0.4769449042326791,
|
| 457 |
+
"step": 2800
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 8.868501529051988,
|
| 461 |
+
"grad_norm": 1.6540861129760742,
|
| 462 |
+
"learning_rate": 0.0002136766220391349,
|
| 463 |
+
"loss": 1.2569,
|
| 464 |
+
"step": 2900
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 8.868501529051988,
|
| 468 |
+
"eval_loss": 0.530635416507721,
|
| 469 |
+
"eval_runtime": 31.7976,
|
| 470 |
+
"eval_samples_per_second": 14.341,
|
| 471 |
+
"eval_steps_per_second": 3.585,
|
| 472 |
+
"eval_wer": 0.48474816741546467,
|
| 473 |
+
"step": 2900
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 9.174311926605505,
|
| 477 |
+
"grad_norm": 3.073190450668335,
|
| 478 |
+
"learning_rate": 0.00021058702368692068,
|
| 479 |
+
"loss": 1.1718,
|
| 480 |
+
"step": 3000
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 9.174311926605505,
|
| 484 |
+
"eval_loss": 0.5160026550292969,
|
| 485 |
+
"eval_runtime": 31.7143,
|
| 486 |
+
"eval_samples_per_second": 14.378,
|
| 487 |
+
"eval_steps_per_second": 3.595,
|
| 488 |
+
"eval_wer": 0.4648853156774651,
|
| 489 |
+
"step": 3000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 9.480122324159021,
|
| 493 |
+
"grad_norm": 5.7398881912231445,
|
| 494 |
+
"learning_rate": 0.00020749742533470648,
|
| 495 |
+
"loss": 1.1354,
|
| 496 |
+
"step": 3100
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 9.480122324159021,
|
| 500 |
+
"eval_loss": 0.5265021920204163,
|
| 501 |
+
"eval_runtime": 31.5861,
|
| 502 |
+
"eval_samples_per_second": 14.437,
|
| 503 |
+
"eval_steps_per_second": 3.609,
|
| 504 |
+
"eval_wer": 0.47765429179475055,
|
| 505 |
+
"step": 3100
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 9.785932721712538,
|
| 509 |
+
"grad_norm": 7.869950294494629,
|
| 510 |
+
"learning_rate": 0.00020440782698249226,
|
| 511 |
+
"loss": 1.2795,
|
| 512 |
+
"step": 3200
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 9.785932721712538,
|
| 516 |
+
"eval_loss": 0.508959949016571,
|
| 517 |
+
"eval_runtime": 31.4657,
|
| 518 |
+
"eval_samples_per_second": 14.492,
|
| 519 |
+
"eval_steps_per_second": 3.623,
|
| 520 |
+
"eval_wer": 0.45897375266020335,
|
| 521 |
+
"step": 3200
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 10.091743119266056,
|
| 525 |
+
"grad_norm": 2.3973193168640137,
|
| 526 |
+
"learning_rate": 0.00020131822863027807,
|
| 527 |
+
"loss": 1.1793,
|
| 528 |
+
"step": 3300
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 10.091743119266056,
|
| 532 |
+
"eval_loss": 0.5265406370162964,
|
| 533 |
+
"eval_runtime": 31.7563,
|
| 534 |
+
"eval_samples_per_second": 14.359,
|
| 535 |
+
"eval_steps_per_second": 3.59,
|
| 536 |
+
"eval_wer": 0.4684322534878222,
|
| 537 |
+
"step": 3300
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 10.397553516819572,
|
| 541 |
+
"grad_norm": 2.35862398147583,
|
| 542 |
+
"learning_rate": 0.00019822863027806382,
|
| 543 |
+
"loss": 1.1647,
|
| 544 |
+
"step": 3400
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"epoch": 10.397553516819572,
|
| 548 |
+
"eval_loss": 0.538512110710144,
|
| 549 |
+
"eval_runtime": 31.6591,
|
| 550 |
+
"eval_samples_per_second": 14.403,
|
| 551 |
+
"eval_steps_per_second": 3.601,
|
| 552 |
+
"eval_wer": 0.4762355166706077,
|
| 553 |
+
"step": 3400
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 10.703363914373089,
|
| 557 |
+
"grad_norm": 1.5017459392547607,
|
| 558 |
+
"learning_rate": 0.00019513903192584962,
|
| 559 |
+
"loss": 1.1978,
|
| 560 |
+
"step": 3500
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 10.703363914373089,
|
| 564 |
+
"eval_loss": 0.5132156610488892,
|
| 565 |
+
"eval_runtime": 31.4578,
|
| 566 |
+
"eval_samples_per_second": 14.496,
|
| 567 |
+
"eval_steps_per_second": 3.624,
|
| 568 |
+
"eval_wer": 0.4715062662567983,
|
| 569 |
+
"step": 3500
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 11.009174311926605,
|
| 573 |
+
"grad_norm": 8.986639022827148,
|
| 574 |
+
"learning_rate": 0.0001920494335736354,
|
| 575 |
+
"loss": 1.1802,
|
| 576 |
+
"step": 3600
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 11.009174311926605,
|
| 580 |
+
"eval_loss": 0.5129914283752441,
|
| 581 |
+
"eval_runtime": 31.4287,
|
| 582 |
+
"eval_samples_per_second": 14.509,
|
| 583 |
+
"eval_steps_per_second": 3.627,
|
| 584 |
+
"eval_wer": 0.45968314022227474,
|
| 585 |
+
"step": 3600
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 11.009174311926605,
|
| 589 |
+
"step": 3600,
|
| 590 |
+
"total_flos": 1.8690129169621533e+19,
|
| 591 |
+
"train_loss": 1.669767551422119,
|
| 592 |
+
"train_runtime": 5274.2639,
|
| 593 |
+
"train_samples_per_second": 14.874,
|
| 594 |
+
"train_steps_per_second": 1.86
|
| 595 |
+
}
|
| 596 |
+
],
|
| 597 |
+
"logging_steps": 100,
|
| 598 |
+
"max_steps": 9810,
|
| 599 |
+
"num_input_tokens_seen": 0,
|
| 600 |
+
"num_train_epochs": 30,
|
| 601 |
+
"save_steps": 400,
|
| 602 |
+
"stateful_callbacks": {
|
| 603 |
+
"EarlyStoppingCallback": {
|
| 604 |
+
"args": {
|
| 605 |
+
"early_stopping_patience": 4,
|
| 606 |
+
"early_stopping_threshold": 0.0
|
| 607 |
+
},
|
| 608 |
+
"attributes": {
|
| 609 |
+
"early_stopping_patience_counter": 4
|
| 610 |
+
}
|
| 611 |
+
},
|
| 612 |
+
"TrainerControl": {
|
| 613 |
+
"args": {
|
| 614 |
+
"should_epoch_stop": false,
|
| 615 |
+
"should_evaluate": false,
|
| 616 |
+
"should_log": false,
|
| 617 |
+
"should_save": true,
|
| 618 |
+
"should_training_stop": true
|
| 619 |
+
},
|
| 620 |
+
"attributes": {}
|
| 621 |
+
}
|
| 622 |
+
},
|
| 623 |
+
"total_flos": 1.8690129169621533e+19,
|
| 624 |
+
"train_batch_size": 4,
|
| 625 |
+
"trial_name": null,
|
| 626 |
+
"trial_params": null
|
| 627 |
+
}
|