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 +579 -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-natbed-n-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 |
+
- natbed
|
| 8 |
+
- mms
|
| 9 |
- generated_from_trainer
|
| 10 |
metrics:
|
| 11 |
- wer
|
|
|
|
| 19 |
|
| 20 |
# mms-1b-all-bem-natbed-n-model
|
| 21 |
|
| 22 |
+
This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on the NATBED - BEM dataset.
|
| 23 |
It achieves the following results on the evaluation set:
|
| 24 |
+
- Loss: 0.5309
|
| 25 |
+
- Wer: 0.4631
|
| 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:a10e903fa91ac0aa9615d18a159b5ba57d7fc491b0c38dc0e2f0bfc86bb935d7
|
| 3 |
+
size 8798532
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.269662921348315,
|
| 3 |
+
"eval_loss": 0.5309441685676575,
|
| 4 |
+
"eval_runtime": 51.9671,
|
| 5 |
+
"eval_samples": 650,
|
| 6 |
+
"eval_samples_per_second": 12.508,
|
| 7 |
+
"eval_steps_per_second": 1.578,
|
| 8 |
+
"eval_wer": 0.46307132690178715,
|
| 9 |
+
"total_flos": 1.8863577270010868e+19,
|
| 10 |
+
"train_loss": 0.8547844314575195,
|
| 11 |
+
"train_runtime": 5284.1307,
|
| 12 |
+
"train_samples": 2845,
|
| 13 |
+
"train_samples_per_second": 16.152,
|
| 14 |
+
"train_steps_per_second": 2.021
|
| 15 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.269662921348315,
|
| 3 |
+
"eval_loss": 0.5309441685676575,
|
| 4 |
+
"eval_runtime": 51.9671,
|
| 5 |
+
"eval_samples": 650,
|
| 6 |
+
"eval_samples_per_second": 12.508,
|
| 7 |
+
"eval_steps_per_second": 1.578,
|
| 8 |
+
"eval_wer": 0.46307132690178715
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.269662921348315,
|
| 3 |
+
"total_flos": 1.8863577270010868e+19,
|
| 4 |
+
"train_loss": 0.8547844314575195,
|
| 5 |
+
"train_runtime": 5284.1307,
|
| 6 |
+
"train_samples": 2845,
|
| 7 |
+
"train_samples_per_second": 16.152,
|
| 8 |
+
"train_steps_per_second": 2.021
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,579 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.5309441685676575,
|
| 3 |
+
"best_model_checkpoint": "/scratch/skscla001/results/mms-1b-all-bem-natbed-n-model/checkpoint-3000",
|
| 4 |
+
"epoch": 9.269662921348315,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 3300,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.2808988764044944,
|
| 13 |
+
"grad_norm": 6.5729265213012695,
|
| 14 |
+
"learning_rate": 0.00028799999999999995,
|
| 15 |
+
"loss": 7.3623,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.2808988764044944,
|
| 20 |
+
"eval_loss": 0.9286547899246216,
|
| 21 |
+
"eval_runtime": 51.6813,
|
| 22 |
+
"eval_samples_per_second": 12.577,
|
| 23 |
+
"eval_steps_per_second": 1.587,
|
| 24 |
+
"eval_wer": 0.7282935315514787,
|
| 25 |
+
"step": 100
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.5617977528089888,
|
| 29 |
+
"grad_norm": 1.9608774185180664,
|
| 30 |
+
"learning_rate": 0.00029727788279773154,
|
| 31 |
+
"loss": 0.9213,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.5617977528089888,
|
| 36 |
+
"eval_loss": 0.651074230670929,
|
| 37 |
+
"eval_runtime": 51.0267,
|
| 38 |
+
"eval_samples_per_second": 12.738,
|
| 39 |
+
"eval_steps_per_second": 1.607,
|
| 40 |
+
"eval_wer": 0.593072908429543,
|
| 41 |
+
"step": 200
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.8426966292134831,
|
| 45 |
+
"grad_norm": 3.582282066345215,
|
| 46 |
+
"learning_rate": 0.0002944423440453686,
|
| 47 |
+
"loss": 0.7224,
|
| 48 |
+
"step": 300
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 0.8426966292134831,
|
| 52 |
+
"eval_loss": 0.6386900544166565,
|
| 53 |
+
"eval_runtime": 51.4556,
|
| 54 |
+
"eval_samples_per_second": 12.632,
|
| 55 |
+
"eval_steps_per_second": 1.594,
|
| 56 |
+
"eval_wer": 0.5434129368970425,
|
| 57 |
+
"step": 300
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 1.1235955056179776,
|
| 61 |
+
"grad_norm": 1.1070303916931152,
|
| 62 |
+
"learning_rate": 0.00029160680529300565,
|
| 63 |
+
"loss": 0.7132,
|
| 64 |
+
"step": 400
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 1.1235955056179776,
|
| 68 |
+
"eval_loss": 0.6140071749687195,
|
| 69 |
+
"eval_runtime": 51.4993,
|
| 70 |
+
"eval_samples_per_second": 12.622,
|
| 71 |
+
"eval_steps_per_second": 1.592,
|
| 72 |
+
"eval_wer": 0.5212715483156729,
|
| 73 |
+
"step": 400
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 1.404494382022472,
|
| 77 |
+
"grad_norm": 2.090729236602783,
|
| 78 |
+
"learning_rate": 0.0002887712665406427,
|
| 79 |
+
"loss": 0.7195,
|
| 80 |
+
"step": 500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 1.404494382022472,
|
| 84 |
+
"eval_loss": 0.6097270250320435,
|
| 85 |
+
"eval_runtime": 51.1603,
|
| 86 |
+
"eval_samples_per_second": 12.705,
|
| 87 |
+
"eval_steps_per_second": 1.603,
|
| 88 |
+
"eval_wer": 0.514629131741262,
|
| 89 |
+
"step": 500
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 1.6853932584269664,
|
| 93 |
+
"grad_norm": 1.059830904006958,
|
| 94 |
+
"learning_rate": 0.00028593572778827975,
|
| 95 |
+
"loss": 0.7054,
|
| 96 |
+
"step": 600
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.6853932584269664,
|
| 100 |
+
"eval_loss": 0.6144998669624329,
|
| 101 |
+
"eval_runtime": 51.4827,
|
| 102 |
+
"eval_samples_per_second": 12.626,
|
| 103 |
+
"eval_steps_per_second": 1.593,
|
| 104 |
+
"eval_wer": 0.5125731456587063,
|
| 105 |
+
"step": 600
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9662921348314608,
|
| 109 |
+
"grad_norm": 1.5953254699707031,
|
| 110 |
+
"learning_rate": 0.0002831285444234404,
|
| 111 |
+
"loss": 0.7417,
|
| 112 |
+
"step": 700
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9662921348314608,
|
| 116 |
+
"eval_loss": 0.606216311454773,
|
| 117 |
+
"eval_runtime": 51.7647,
|
| 118 |
+
"eval_samples_per_second": 12.557,
|
| 119 |
+
"eval_steps_per_second": 1.584,
|
| 120 |
+
"eval_wer": 0.5256998260319469,
|
| 121 |
+
"step": 700
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 2.247191011235955,
|
| 125 |
+
"grad_norm": 0.7664604187011719,
|
| 126 |
+
"learning_rate": 0.00028029300567107747,
|
| 127 |
+
"loss": 0.7029,
|
| 128 |
+
"step": 800
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 2.247191011235955,
|
| 132 |
+
"eval_loss": 0.6021894812583923,
|
| 133 |
+
"eval_runtime": 52.1384,
|
| 134 |
+
"eval_samples_per_second": 12.467,
|
| 135 |
+
"eval_steps_per_second": 1.573,
|
| 136 |
+
"eval_wer": 0.4947018820180294,
|
| 137 |
+
"step": 800
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.5280898876404496,
|
| 141 |
+
"grad_norm": 1.113645076751709,
|
| 142 |
+
"learning_rate": 0.0002774574669187145,
|
| 143 |
+
"loss": 0.6845,
|
| 144 |
+
"step": 900
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.5280898876404496,
|
| 148 |
+
"eval_loss": 0.5886064767837524,
|
| 149 |
+
"eval_runtime": 51.3267,
|
| 150 |
+
"eval_samples_per_second": 12.664,
|
| 151 |
+
"eval_steps_per_second": 1.598,
|
| 152 |
+
"eval_wer": 0.5022932152459275,
|
| 153 |
+
"step": 900
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 2.808988764044944,
|
| 157 |
+
"grad_norm": 1.8644826412200928,
|
| 158 |
+
"learning_rate": 0.00027462192816635157,
|
| 159 |
+
"loss": 0.663,
|
| 160 |
+
"step": 1000
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 2.808988764044944,
|
| 164 |
+
"eval_loss": 0.5914585590362549,
|
| 165 |
+
"eval_runtime": 51.3494,
|
| 166 |
+
"eval_samples_per_second": 12.658,
|
| 167 |
+
"eval_steps_per_second": 1.597,
|
| 168 |
+
"eval_wer": 0.49264589593547364,
|
| 169 |
+
"step": 1000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 3.0898876404494384,
|
| 173 |
+
"grad_norm": 0.9567739963531494,
|
| 174 |
+
"learning_rate": 0.0002717863894139887,
|
| 175 |
+
"loss": 0.7129,
|
| 176 |
+
"step": 1100
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.0898876404494384,
|
| 180 |
+
"eval_loss": 0.5832971930503845,
|
| 181 |
+
"eval_runtime": 51.7503,
|
| 182 |
+
"eval_samples_per_second": 12.56,
|
| 183 |
+
"eval_steps_per_second": 1.585,
|
| 184 |
+
"eval_wer": 0.4920132848331488,
|
| 185 |
+
"step": 1100
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 3.370786516853933,
|
| 189 |
+
"grad_norm": 2.428119659423828,
|
| 190 |
+
"learning_rate": 0.00026895085066162567,
|
| 191 |
+
"loss": 0.6735,
|
| 192 |
+
"step": 1200
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 3.370786516853933,
|
| 196 |
+
"eval_loss": 0.5876715779304504,
|
| 197 |
+
"eval_runtime": 51.7607,
|
| 198 |
+
"eval_samples_per_second": 12.558,
|
| 199 |
+
"eval_steps_per_second": 1.584,
|
| 200 |
+
"eval_wer": 0.483156729400601,
|
| 201 |
+
"step": 1200
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 3.6516853932584272,
|
| 205 |
+
"grad_norm": 10.181241989135742,
|
| 206 |
+
"learning_rate": 0.0002661153119092627,
|
| 207 |
+
"loss": 0.672,
|
| 208 |
+
"step": 1300
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 3.6516853932584272,
|
| 212 |
+
"eval_loss": 0.5862510204315186,
|
| 213 |
+
"eval_runtime": 51.4111,
|
| 214 |
+
"eval_samples_per_second": 12.643,
|
| 215 |
+
"eval_steps_per_second": 1.595,
|
| 216 |
+
"eval_wer": 0.5151035900680057,
|
| 217 |
+
"step": 1300
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 3.932584269662921,
|
| 221 |
+
"grad_norm": 2.6570065021514893,
|
| 222 |
+
"learning_rate": 0.0002632797731568998,
|
| 223 |
+
"loss": 0.6494,
|
| 224 |
+
"step": 1400
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 3.932584269662921,
|
| 228 |
+
"eval_loss": 0.579518735408783,
|
| 229 |
+
"eval_runtime": 51.3192,
|
| 230 |
+
"eval_samples_per_second": 12.666,
|
| 231 |
+
"eval_steps_per_second": 1.598,
|
| 232 |
+
"eval_wer": 0.48442195160525064,
|
| 233 |
+
"step": 1400
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 4.213483146067416,
|
| 237 |
+
"grad_norm": 2.2995524406433105,
|
| 238 |
+
"learning_rate": 0.0002604442344045368,
|
| 239 |
+
"loss": 0.7049,
|
| 240 |
+
"step": 1500
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 4.213483146067416,
|
| 244 |
+
"eval_loss": 0.5723974704742432,
|
| 245 |
+
"eval_runtime": 51.7136,
|
| 246 |
+
"eval_samples_per_second": 12.569,
|
| 247 |
+
"eval_steps_per_second": 1.586,
|
| 248 |
+
"eval_wer": 0.47161157678317256,
|
| 249 |
+
"step": 1500
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 4.49438202247191,
|
| 253 |
+
"grad_norm": 1.1362881660461426,
|
| 254 |
+
"learning_rate": 0.0002576086956521739,
|
| 255 |
+
"loss": 0.5898,
|
| 256 |
+
"step": 1600
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 4.49438202247191,
|
| 260 |
+
"eval_loss": 0.5640456676483154,
|
| 261 |
+
"eval_runtime": 51.8126,
|
| 262 |
+
"eval_samples_per_second": 12.545,
|
| 263 |
+
"eval_steps_per_second": 1.583,
|
| 264 |
+
"eval_wer": 0.4761980072750277,
|
| 265 |
+
"step": 1600
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 4.775280898876405,
|
| 269 |
+
"grad_norm": 1.0056232213974,
|
| 270 |
+
"learning_rate": 0.00025477315689981093,
|
| 271 |
+
"loss": 0.6581,
|
| 272 |
+
"step": 1700
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 4.775280898876405,
|
| 276 |
+
"eval_loss": 0.5581757426261902,
|
| 277 |
+
"eval_runtime": 51.7242,
|
| 278 |
+
"eval_samples_per_second": 12.567,
|
| 279 |
+
"eval_steps_per_second": 1.585,
|
| 280 |
+
"eval_wer": 0.4724023406610786,
|
| 281 |
+
"step": 1700
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 5.056179775280899,
|
| 285 |
+
"grad_norm": 0.7103342413902283,
|
| 286 |
+
"learning_rate": 0.000251937618147448,
|
| 287 |
+
"loss": 0.6262,
|
| 288 |
+
"step": 1800
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 5.056179775280899,
|
| 292 |
+
"eval_loss": 0.5446608066558838,
|
| 293 |
+
"eval_runtime": 51.6418,
|
| 294 |
+
"eval_samples_per_second": 12.587,
|
| 295 |
+
"eval_steps_per_second": 1.588,
|
| 296 |
+
"eval_wer": 0.4750909378459592,
|
| 297 |
+
"step": 1800
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 5.337078651685394,
|
| 301 |
+
"grad_norm": 1.2600806951522827,
|
| 302 |
+
"learning_rate": 0.00024910207939508503,
|
| 303 |
+
"loss": 0.6179,
|
| 304 |
+
"step": 1900
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 5.337078651685394,
|
| 308 |
+
"eval_loss": 0.549724280834198,
|
| 309 |
+
"eval_runtime": 51.7411,
|
| 310 |
+
"eval_samples_per_second": 12.563,
|
| 311 |
+
"eval_steps_per_second": 1.585,
|
| 312 |
+
"eval_wer": 0.4656017713110865,
|
| 313 |
+
"step": 1900
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 5.617977528089888,
|
| 317 |
+
"grad_norm": 14.24419116973877,
|
| 318 |
+
"learning_rate": 0.0002462665406427221,
|
| 319 |
+
"loss": 0.5896,
|
| 320 |
+
"step": 2000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.617977528089888,
|
| 324 |
+
"eval_loss": 0.5444263815879822,
|
| 325 |
+
"eval_runtime": 51.8052,
|
| 326 |
+
"eval_samples_per_second": 12.547,
|
| 327 |
+
"eval_steps_per_second": 1.583,
|
| 328 |
+
"eval_wer": 0.477937687806421,
|
| 329 |
+
"step": 2000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 5.898876404494382,
|
| 333 |
+
"grad_norm": 0.867696225643158,
|
| 334 |
+
"learning_rate": 0.00024343100189035916,
|
| 335 |
+
"loss": 0.6438,
|
| 336 |
+
"step": 2100
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 5.898876404494382,
|
| 340 |
+
"eval_loss": 0.5399273037910461,
|
| 341 |
+
"eval_runtime": 51.4899,
|
| 342 |
+
"eval_samples_per_second": 12.624,
|
| 343 |
+
"eval_steps_per_second": 1.593,
|
| 344 |
+
"eval_wer": 0.47003004902736045,
|
| 345 |
+
"step": 2100
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 6.179775280898877,
|
| 349 |
+
"grad_norm": 1.4865626096725464,
|
| 350 |
+
"learning_rate": 0.0002405954631379962,
|
| 351 |
+
"loss": 0.6086,
|
| 352 |
+
"step": 2200
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 6.179775280898877,
|
| 356 |
+
"eval_loss": 0.5520233511924744,
|
| 357 |
+
"eval_runtime": 51.3367,
|
| 358 |
+
"eval_samples_per_second": 12.662,
|
| 359 |
+
"eval_steps_per_second": 1.597,
|
| 360 |
+
"eval_wer": 0.4597501186145817,
|
| 361 |
+
"step": 2200
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 6.460674157303371,
|
| 365 |
+
"grad_norm": 1.7858107089996338,
|
| 366 |
+
"learning_rate": 0.00023775992438563324,
|
| 367 |
+
"loss": 0.6226,
|
| 368 |
+
"step": 2300
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 6.460674157303371,
|
| 372 |
+
"eval_loss": 0.5385509133338928,
|
| 373 |
+
"eval_runtime": 51.8468,
|
| 374 |
+
"eval_samples_per_second": 12.537,
|
| 375 |
+
"eval_steps_per_second": 1.582,
|
| 376 |
+
"eval_wer": 0.4796773683378143,
|
| 377 |
+
"step": 2300
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.741573033707866,
|
| 381 |
+
"grad_norm": 3.9033384323120117,
|
| 382 |
+
"learning_rate": 0.0002349243856332703,
|
| 383 |
+
"loss": 0.6148,
|
| 384 |
+
"step": 2400
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.741573033707866,
|
| 388 |
+
"eval_loss": 0.5573983788490295,
|
| 389 |
+
"eval_runtime": 51.5238,
|
| 390 |
+
"eval_samples_per_second": 12.616,
|
| 391 |
+
"eval_steps_per_second": 1.591,
|
| 392 |
+
"eval_wer": 0.4679740629448047,
|
| 393 |
+
"step": 2400
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 7.022471910112359,
|
| 397 |
+
"grad_norm": 1.0407049655914307,
|
| 398 |
+
"learning_rate": 0.00023208884688090737,
|
| 399 |
+
"loss": 0.5838,
|
| 400 |
+
"step": 2500
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 7.022471910112359,
|
| 404 |
+
"eval_loss": 0.5497230887413025,
|
| 405 |
+
"eval_runtime": 51.2462,
|
| 406 |
+
"eval_samples_per_second": 12.684,
|
| 407 |
+
"eval_steps_per_second": 1.6,
|
| 408 |
+
"eval_wer": 0.4638620907796932,
|
| 409 |
+
"step": 2500
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 7.303370786516854,
|
| 413 |
+
"grad_norm": 0.836614191532135,
|
| 414 |
+
"learning_rate": 0.00022925330812854442,
|
| 415 |
+
"loss": 0.5407,
|
| 416 |
+
"step": 2600
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 7.303370786516854,
|
| 420 |
+
"eval_loss": 0.5377057790756226,
|
| 421 |
+
"eval_runtime": 51.5353,
|
| 422 |
+
"eval_samples_per_second": 12.613,
|
| 423 |
+
"eval_steps_per_second": 1.591,
|
| 424 |
+
"eval_wer": 0.46307132690178715,
|
| 425 |
+
"step": 2600
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 7.584269662921348,
|
| 429 |
+
"grad_norm": 2.6512720584869385,
|
| 430 |
+
"learning_rate": 0.00022641776937618147,
|
| 431 |
+
"loss": 0.6186,
|
| 432 |
+
"step": 2700
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 7.584269662921348,
|
| 436 |
+
"eval_loss": 0.5403843522071838,
|
| 437 |
+
"eval_runtime": 51.536,
|
| 438 |
+
"eval_samples_per_second": 12.613,
|
| 439 |
+
"eval_steps_per_second": 1.591,
|
| 440 |
+
"eval_wer": 0.4714534240075913,
|
| 441 |
+
"step": 2700
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 7.865168539325842,
|
| 445 |
+
"grad_norm": 2.884155035018921,
|
| 446 |
+
"learning_rate": 0.0002235822306238185,
|
| 447 |
+
"loss": 0.5922,
|
| 448 |
+
"step": 2800
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 7.865168539325842,
|
| 452 |
+
"eval_loss": 0.5381007790565491,
|
| 453 |
+
"eval_runtime": 51.5402,
|
| 454 |
+
"eval_samples_per_second": 12.612,
|
| 455 |
+
"eval_steps_per_second": 1.591,
|
| 456 |
+
"eval_wer": 0.46085718804365017,
|
| 457 |
+
"step": 2800
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 8.146067415730338,
|
| 461 |
+
"grad_norm": 1.4212076663970947,
|
| 462 |
+
"learning_rate": 0.00022074669187145554,
|
| 463 |
+
"loss": 0.5799,
|
| 464 |
+
"step": 2900
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 8.146067415730338,
|
| 468 |
+
"eval_loss": 0.5311689376831055,
|
| 469 |
+
"eval_runtime": 51.4207,
|
| 470 |
+
"eval_samples_per_second": 12.641,
|
| 471 |
+
"eval_steps_per_second": 1.595,
|
| 472 |
+
"eval_wer": 0.46196425747271863,
|
| 473 |
+
"step": 2900
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 8.426966292134832,
|
| 477 |
+
"grad_norm": 1.6221522092819214,
|
| 478 |
+
"learning_rate": 0.0002179111531190926,
|
| 479 |
+
"loss": 0.5914,
|
| 480 |
+
"step": 3000
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 8.426966292134832,
|
| 484 |
+
"eval_loss": 0.5309441685676575,
|
| 485 |
+
"eval_runtime": 51.2813,
|
| 486 |
+
"eval_samples_per_second": 12.675,
|
| 487 |
+
"eval_steps_per_second": 1.599,
|
| 488 |
+
"eval_wer": 0.46307132690178715,
|
| 489 |
+
"step": 3000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 8.707865168539326,
|
| 493 |
+
"grad_norm": 0.46160590648651123,
|
| 494 |
+
"learning_rate": 0.00021507561436672967,
|
| 495 |
+
"loss": 0.6194,
|
| 496 |
+
"step": 3100
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 8.707865168539326,
|
| 500 |
+
"eval_loss": 0.5316519141197205,
|
| 501 |
+
"eval_runtime": 51.9102,
|
| 502 |
+
"eval_samples_per_second": 12.522,
|
| 503 |
+
"eval_steps_per_second": 1.58,
|
| 504 |
+
"eval_wer": 0.46781591016922347,
|
| 505 |
+
"step": 3100
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 8.98876404494382,
|
| 509 |
+
"grad_norm": 0.4698401093482971,
|
| 510 |
+
"learning_rate": 0.00021224007561436672,
|
| 511 |
+
"loss": 0.5851,
|
| 512 |
+
"step": 3200
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 8.98876404494382,
|
| 516 |
+
"eval_loss": 0.5388666987419128,
|
| 517 |
+
"eval_runtime": 51.8653,
|
| 518 |
+
"eval_samples_per_second": 12.532,
|
| 519 |
+
"eval_steps_per_second": 1.581,
|
| 520 |
+
"eval_wer": 0.4575359797564447,
|
| 521 |
+
"step": 3200
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 9.269662921348315,
|
| 525 |
+
"grad_norm": 2.0234947204589844,
|
| 526 |
+
"learning_rate": 0.00020940453686200375,
|
| 527 |
+
"loss": 0.5764,
|
| 528 |
+
"step": 3300
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 9.269662921348315,
|
| 532 |
+
"eval_loss": 0.5578745603561401,
|
| 533 |
+
"eval_runtime": 51.3849,
|
| 534 |
+
"eval_samples_per_second": 12.65,
|
| 535 |
+
"eval_steps_per_second": 1.596,
|
| 536 |
+
"eval_wer": 0.45500553534714533,
|
| 537 |
+
"step": 3300
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 9.269662921348315,
|
| 541 |
+
"step": 3300,
|
| 542 |
+
"total_flos": 1.8863577270010868e+19,
|
| 543 |
+
"train_loss": 0.8547844314575195,
|
| 544 |
+
"train_runtime": 5284.1307,
|
| 545 |
+
"train_samples_per_second": 16.152,
|
| 546 |
+
"train_steps_per_second": 2.021
|
| 547 |
+
}
|
| 548 |
+
],
|
| 549 |
+
"logging_steps": 100,
|
| 550 |
+
"max_steps": 10680,
|
| 551 |
+
"num_input_tokens_seen": 0,
|
| 552 |
+
"num_train_epochs": 30,
|
| 553 |
+
"save_steps": 200,
|
| 554 |
+
"stateful_callbacks": {
|
| 555 |
+
"EarlyStoppingCallback": {
|
| 556 |
+
"args": {
|
| 557 |
+
"early_stopping_patience": 3,
|
| 558 |
+
"early_stopping_threshold": 0.0
|
| 559 |
+
},
|
| 560 |
+
"attributes": {
|
| 561 |
+
"early_stopping_patience_counter": 2
|
| 562 |
+
}
|
| 563 |
+
},
|
| 564 |
+
"TrainerControl": {
|
| 565 |
+
"args": {
|
| 566 |
+
"should_epoch_stop": false,
|
| 567 |
+
"should_evaluate": false,
|
| 568 |
+
"should_log": false,
|
| 569 |
+
"should_save": true,
|
| 570 |
+
"should_training_stop": false
|
| 571 |
+
},
|
| 572 |
+
"attributes": {}
|
| 573 |
+
}
|
| 574 |
+
},
|
| 575 |
+
"total_flos": 1.8863577270010868e+19,
|
| 576 |
+
"train_batch_size": 8,
|
| 577 |
+
"trial_name": null,
|
| 578 |
+
"trial_params": null
|
| 579 |
+
}
|