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 +499 -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-nn-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-nn-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: 0.5884
|
| 25 |
+
- Wer: 0.5333
|
| 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:ea2306c85baea18513943b267ad30910539e3f2d2b52c1cfc5d36f927c2d9a74
|
| 3 |
+
size 8819028
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 6.320541760722348,
|
| 3 |
+
"eval_loss": 0.5884432792663574,
|
| 4 |
+
"eval_runtime": 45.0521,
|
| 5 |
+
"eval_samples": 708,
|
| 6 |
+
"eval_samples_per_second": 15.715,
|
| 7 |
+
"eval_steps_per_second": 1.975,
|
| 8 |
+
"eval_wer": 0.5332880434782609,
|
| 9 |
+
"total_flos": 1.3313268682658783e+19,
|
| 10 |
+
"train_loss": 0.972996187210083,
|
| 11 |
+
"train_runtime": 3933.908,
|
| 12 |
+
"train_samples": 3542,
|
| 13 |
+
"train_samples_per_second": 27.011,
|
| 14 |
+
"train_steps_per_second": 3.378
|
| 15 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 6.320541760722348,
|
| 3 |
+
"eval_loss": 0.5884432792663574,
|
| 4 |
+
"eval_runtime": 45.0521,
|
| 5 |
+
"eval_samples": 708,
|
| 6 |
+
"eval_samples_per_second": 15.715,
|
| 7 |
+
"eval_steps_per_second": 1.975,
|
| 8 |
+
"eval_wer": 0.5332880434782609
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 6.320541760722348,
|
| 3 |
+
"total_flos": 1.3313268682658783e+19,
|
| 4 |
+
"train_loss": 0.972996187210083,
|
| 5 |
+
"train_runtime": 3933.908,
|
| 6 |
+
"train_samples": 3542,
|
| 7 |
+
"train_samples_per_second": 27.011,
|
| 8 |
+
"train_steps_per_second": 3.378
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,499 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.5884432792663574,
|
| 3 |
+
"best_model_checkpoint": "/scratch/skscla001/results/mms-1b-all-bem-natbed-nn-model/checkpoint-2200",
|
| 4 |
+
"epoch": 6.320541760722348,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 2800,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.22573363431151242,
|
| 13 |
+
"grad_norm": 1.8729794025421143,
|
| 14 |
+
"learning_rate": 0.00028799999999999995,
|
| 15 |
+
"loss": 7.9244,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.22573363431151242,
|
| 20 |
+
"eval_loss": 1.35137140750885,
|
| 21 |
+
"eval_runtime": 45.1782,
|
| 22 |
+
"eval_samples_per_second": 15.671,
|
| 23 |
+
"eval_steps_per_second": 1.97,
|
| 24 |
+
"eval_wer": 1.0237771739130435,
|
| 25 |
+
"step": 100
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.45146726862302483,
|
| 29 |
+
"grad_norm": 2.6380980014801025,
|
| 30 |
+
"learning_rate": 0.0002978165276724791,
|
| 31 |
+
"loss": 1.0236,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.45146726862302483,
|
| 36 |
+
"eval_loss": 0.8354936838150024,
|
| 37 |
+
"eval_runtime": 44.7017,
|
| 38 |
+
"eval_samples_per_second": 15.838,
|
| 39 |
+
"eval_steps_per_second": 1.991,
|
| 40 |
+
"eval_wer": 0.6594769021739131,
|
| 41 |
+
"step": 200
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.6772009029345373,
|
| 45 |
+
"grad_norm": 1.9870244264602661,
|
| 46 |
+
"learning_rate": 0.0002955420773313116,
|
| 47 |
+
"loss": 0.8005,
|
| 48 |
+
"step": 300
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 0.6772009029345373,
|
| 52 |
+
"eval_loss": 0.7836518287658691,
|
| 53 |
+
"eval_runtime": 45.0039,
|
| 54 |
+
"eval_samples_per_second": 15.732,
|
| 55 |
+
"eval_steps_per_second": 1.978,
|
| 56 |
+
"eval_wer": 0.6141304347826086,
|
| 57 |
+
"step": 300
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.9029345372460497,
|
| 61 |
+
"grad_norm": 2.012474536895752,
|
| 62 |
+
"learning_rate": 0.00029326762699014404,
|
| 63 |
+
"loss": 0.8968,
|
| 64 |
+
"step": 400
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 0.9029345372460497,
|
| 68 |
+
"eval_loss": 0.7808529138565063,
|
| 69 |
+
"eval_runtime": 44.7704,
|
| 70 |
+
"eval_samples_per_second": 15.814,
|
| 71 |
+
"eval_steps_per_second": 1.988,
|
| 72 |
+
"eval_wer": 0.6042798913043478,
|
| 73 |
+
"step": 400
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 1.1286681715575622,
|
| 77 |
+
"grad_norm": 8.587899208068848,
|
| 78 |
+
"learning_rate": 0.00029099317664897645,
|
| 79 |
+
"loss": 0.8909,
|
| 80 |
+
"step": 500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 1.1286681715575622,
|
| 84 |
+
"eval_loss": 0.7146816849708557,
|
| 85 |
+
"eval_runtime": 44.5851,
|
| 86 |
+
"eval_samples_per_second": 15.88,
|
| 87 |
+
"eval_steps_per_second": 1.996,
|
| 88 |
+
"eval_wer": 0.5952785326086957,
|
| 89 |
+
"step": 500
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 1.3544018058690745,
|
| 93 |
+
"grad_norm": 8.078141212463379,
|
| 94 |
+
"learning_rate": 0.0002887187263078089,
|
| 95 |
+
"loss": 0.7983,
|
| 96 |
+
"step": 600
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.3544018058690745,
|
| 100 |
+
"eval_loss": 0.6989510655403137,
|
| 101 |
+
"eval_runtime": 44.4701,
|
| 102 |
+
"eval_samples_per_second": 15.921,
|
| 103 |
+
"eval_steps_per_second": 2.001,
|
| 104 |
+
"eval_wer": 0.5930706521739131,
|
| 105 |
+
"step": 600
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.580135440180587,
|
| 109 |
+
"grad_norm": 2.5096583366394043,
|
| 110 |
+
"learning_rate": 0.00028644427596664137,
|
| 111 |
+
"loss": 0.8563,
|
| 112 |
+
"step": 700
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.580135440180587,
|
| 116 |
+
"eval_loss": 0.6804757118225098,
|
| 117 |
+
"eval_runtime": 44.8305,
|
| 118 |
+
"eval_samples_per_second": 15.793,
|
| 119 |
+
"eval_steps_per_second": 1.985,
|
| 120 |
+
"eval_wer": 0.5964673913043478,
|
| 121 |
+
"step": 700
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 1.8058690744920993,
|
| 125 |
+
"grad_norm": 2.3668673038482666,
|
| 126 |
+
"learning_rate": 0.00028416982562547383,
|
| 127 |
+
"loss": 0.7094,
|
| 128 |
+
"step": 800
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.8058690744920993,
|
| 132 |
+
"eval_loss": 0.6849333047866821,
|
| 133 |
+
"eval_runtime": 44.6427,
|
| 134 |
+
"eval_samples_per_second": 15.859,
|
| 135 |
+
"eval_steps_per_second": 1.994,
|
| 136 |
+
"eval_wer": 0.5808423913043478,
|
| 137 |
+
"step": 800
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.0316027088036117,
|
| 141 |
+
"grad_norm": 1.9050077199935913,
|
| 142 |
+
"learning_rate": 0.0002819181197877179,
|
| 143 |
+
"loss": 0.7499,
|
| 144 |
+
"step": 900
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.0316027088036117,
|
| 148 |
+
"eval_loss": 0.6456880569458008,
|
| 149 |
+
"eval_runtime": 44.6644,
|
| 150 |
+
"eval_samples_per_second": 15.852,
|
| 151 |
+
"eval_steps_per_second": 1.993,
|
| 152 |
+
"eval_wer": 0.5934103260869565,
|
| 153 |
+
"step": 900
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 2.2573363431151243,
|
| 157 |
+
"grad_norm": 2.829195737838745,
|
| 158 |
+
"learning_rate": 0.0002796436694465504,
|
| 159 |
+
"loss": 0.7722,
|
| 160 |
+
"step": 1000
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 2.2573363431151243,
|
| 164 |
+
"eval_loss": 0.6565266251564026,
|
| 165 |
+
"eval_runtime": 44.7397,
|
| 166 |
+
"eval_samples_per_second": 15.825,
|
| 167 |
+
"eval_steps_per_second": 1.989,
|
| 168 |
+
"eval_wer": 0.5874660326086957,
|
| 169 |
+
"step": 1000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 2.4830699774266365,
|
| 173 |
+
"grad_norm": 0.8484971523284912,
|
| 174 |
+
"learning_rate": 0.00027736921910538284,
|
| 175 |
+
"loss": 0.7099,
|
| 176 |
+
"step": 1100
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 2.4830699774266365,
|
| 180 |
+
"eval_loss": 0.6419216394424438,
|
| 181 |
+
"eval_runtime": 44.7023,
|
| 182 |
+
"eval_samples_per_second": 15.838,
|
| 183 |
+
"eval_steps_per_second": 1.991,
|
| 184 |
+
"eval_wer": 0.5596127717391305,
|
| 185 |
+
"step": 1100
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 2.708803611738149,
|
| 189 |
+
"grad_norm": 2.4867944717407227,
|
| 190 |
+
"learning_rate": 0.0002750947687642153,
|
| 191 |
+
"loss": 0.7416,
|
| 192 |
+
"step": 1200
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 2.708803611738149,
|
| 196 |
+
"eval_loss": 0.6195096373558044,
|
| 197 |
+
"eval_runtime": 44.7966,
|
| 198 |
+
"eval_samples_per_second": 15.805,
|
| 199 |
+
"eval_steps_per_second": 1.987,
|
| 200 |
+
"eval_wer": 0.561141304347826,
|
| 201 |
+
"step": 1200
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 2.9345372460496613,
|
| 205 |
+
"grad_norm": 4.84053373336792,
|
| 206 |
+
"learning_rate": 0.00027282031842304776,
|
| 207 |
+
"loss": 0.6385,
|
| 208 |
+
"step": 1300
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 2.9345372460496613,
|
| 212 |
+
"eval_loss": 0.6227733492851257,
|
| 213 |
+
"eval_runtime": 44.9228,
|
| 214 |
+
"eval_samples_per_second": 15.76,
|
| 215 |
+
"eval_steps_per_second": 1.981,
|
| 216 |
+
"eval_wer": 0.5647078804347826,
|
| 217 |
+
"step": 1300
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 3.160270880361174,
|
| 221 |
+
"grad_norm": 0.9543440341949463,
|
| 222 |
+
"learning_rate": 0.0002705458680818802,
|
| 223 |
+
"loss": 0.6436,
|
| 224 |
+
"step": 1400
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 3.160270880361174,
|
| 228 |
+
"eval_loss": 0.6184154152870178,
|
| 229 |
+
"eval_runtime": 44.7251,
|
| 230 |
+
"eval_samples_per_second": 15.83,
|
| 231 |
+
"eval_steps_per_second": 1.99,
|
| 232 |
+
"eval_wer": 0.5509510869565217,
|
| 233 |
+
"step": 1400
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 3.386004514672686,
|
| 237 |
+
"grad_norm": 0.8133373856544495,
|
| 238 |
+
"learning_rate": 0.0002682714177407126,
|
| 239 |
+
"loss": 0.6795,
|
| 240 |
+
"step": 1500
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 3.386004514672686,
|
| 244 |
+
"eval_loss": 0.6156527996063232,
|
| 245 |
+
"eval_runtime": 45.3372,
|
| 246 |
+
"eval_samples_per_second": 15.616,
|
| 247 |
+
"eval_steps_per_second": 1.963,
|
| 248 |
+
"eval_wer": 0.553328804347826,
|
| 249 |
+
"step": 1500
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 3.6117381489841986,
|
| 253 |
+
"grad_norm": 25.58840560913086,
|
| 254 |
+
"learning_rate": 0.0002659969673995451,
|
| 255 |
+
"loss": 0.7027,
|
| 256 |
+
"step": 1600
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 3.6117381489841986,
|
| 260 |
+
"eval_loss": 0.6343082785606384,
|
| 261 |
+
"eval_runtime": 45.529,
|
| 262 |
+
"eval_samples_per_second": 15.551,
|
| 263 |
+
"eval_steps_per_second": 1.955,
|
| 264 |
+
"eval_wer": 0.5426290760869565,
|
| 265 |
+
"step": 1600
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 3.837471783295711,
|
| 269 |
+
"grad_norm": 0.6009318828582764,
|
| 270 |
+
"learning_rate": 0.00026372251705837754,
|
| 271 |
+
"loss": 0.6585,
|
| 272 |
+
"step": 1700
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 3.837471783295711,
|
| 276 |
+
"eval_loss": 0.6057115793228149,
|
| 277 |
+
"eval_runtime": 44.8336,
|
| 278 |
+
"eval_samples_per_second": 15.792,
|
| 279 |
+
"eval_steps_per_second": 1.985,
|
| 280 |
+
"eval_wer": 0.5427989130434783,
|
| 281 |
+
"step": 1700
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 4.063205417607223,
|
| 285 |
+
"grad_norm": 0.929165244102478,
|
| 286 |
+
"learning_rate": 0.00026144806671720994,
|
| 287 |
+
"loss": 0.6351,
|
| 288 |
+
"step": 1800
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 4.063205417607223,
|
| 292 |
+
"eval_loss": 0.6017059683799744,
|
| 293 |
+
"eval_runtime": 44.5067,
|
| 294 |
+
"eval_samples_per_second": 15.908,
|
| 295 |
+
"eval_steps_per_second": 2.0,
|
| 296 |
+
"eval_wer": 0.54296875,
|
| 297 |
+
"step": 1800
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 4.288939051918736,
|
| 301 |
+
"grad_norm": 1.4761062860488892,
|
| 302 |
+
"learning_rate": 0.00025917361637604246,
|
| 303 |
+
"loss": 0.6528,
|
| 304 |
+
"step": 1900
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 4.288939051918736,
|
| 308 |
+
"eval_loss": 0.6098975539207458,
|
| 309 |
+
"eval_runtime": 44.8754,
|
| 310 |
+
"eval_samples_per_second": 15.777,
|
| 311 |
+
"eval_steps_per_second": 1.983,
|
| 312 |
+
"eval_wer": 0.5339673913043478,
|
| 313 |
+
"step": 1900
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.514672686230249,
|
| 317 |
+
"grad_norm": 1.2957922220230103,
|
| 318 |
+
"learning_rate": 0.00025689916603487486,
|
| 319 |
+
"loss": 0.6603,
|
| 320 |
+
"step": 2000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 4.514672686230249,
|
| 324 |
+
"eval_loss": 0.621790885925293,
|
| 325 |
+
"eval_runtime": 45.2703,
|
| 326 |
+
"eval_samples_per_second": 15.639,
|
| 327 |
+
"eval_steps_per_second": 1.966,
|
| 328 |
+
"eval_wer": 0.5334578804347826,
|
| 329 |
+
"step": 2000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 4.74040632054176,
|
| 333 |
+
"grad_norm": 3.225343942642212,
|
| 334 |
+
"learning_rate": 0.0002546247156937073,
|
| 335 |
+
"loss": 0.6676,
|
| 336 |
+
"step": 2100
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 4.74040632054176,
|
| 340 |
+
"eval_loss": 0.5977216958999634,
|
| 341 |
+
"eval_runtime": 44.9573,
|
| 342 |
+
"eval_samples_per_second": 15.748,
|
| 343 |
+
"eval_steps_per_second": 1.98,
|
| 344 |
+
"eval_wer": 0.5322690217391305,
|
| 345 |
+
"step": 2100
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 4.966139954853273,
|
| 349 |
+
"grad_norm": 1.4750922918319702,
|
| 350 |
+
"learning_rate": 0.0002523502653525398,
|
| 351 |
+
"loss": 0.6304,
|
| 352 |
+
"step": 2200
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 4.966139954853273,
|
| 356 |
+
"eval_loss": 0.5884432792663574,
|
| 357 |
+
"eval_runtime": 45.456,
|
| 358 |
+
"eval_samples_per_second": 15.575,
|
| 359 |
+
"eval_steps_per_second": 1.958,
|
| 360 |
+
"eval_wer": 0.5332880434782609,
|
| 361 |
+
"step": 2200
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 5.191873589164786,
|
| 365 |
+
"grad_norm": 0.7652086615562439,
|
| 366 |
+
"learning_rate": 0.00025007581501137224,
|
| 367 |
+
"loss": 0.5976,
|
| 368 |
+
"step": 2300
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 5.191873589164786,
|
| 372 |
+
"eval_loss": 0.5955621600151062,
|
| 373 |
+
"eval_runtime": 45.3065,
|
| 374 |
+
"eval_samples_per_second": 15.627,
|
| 375 |
+
"eval_steps_per_second": 1.964,
|
| 376 |
+
"eval_wer": 0.5227581521739131,
|
| 377 |
+
"step": 2300
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 5.417607223476298,
|
| 381 |
+
"grad_norm": 3.0203540325164795,
|
| 382 |
+
"learning_rate": 0.00024780136467020465,
|
| 383 |
+
"loss": 0.6564,
|
| 384 |
+
"step": 2400
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 5.417607223476298,
|
| 388 |
+
"eval_loss": 0.5956509709358215,
|
| 389 |
+
"eval_runtime": 45.0964,
|
| 390 |
+
"eval_samples_per_second": 15.7,
|
| 391 |
+
"eval_steps_per_second": 1.974,
|
| 392 |
+
"eval_wer": 0.5302309782608695,
|
| 393 |
+
"step": 2400
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 5.643340857787811,
|
| 397 |
+
"grad_norm": 1.8780221939086914,
|
| 398 |
+
"learning_rate": 0.00024552691432903716,
|
| 399 |
+
"loss": 0.6717,
|
| 400 |
+
"step": 2500
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 5.643340857787811,
|
| 404 |
+
"eval_loss": 0.5767239332199097,
|
| 405 |
+
"eval_runtime": 44.9865,
|
| 406 |
+
"eval_samples_per_second": 15.738,
|
| 407 |
+
"eval_steps_per_second": 1.978,
|
| 408 |
+
"eval_wer": 0.5183423913043478,
|
| 409 |
+
"step": 2500
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 5.8690744920993225,
|
| 413 |
+
"grad_norm": 0.8542383909225464,
|
| 414 |
+
"learning_rate": 0.00024325246398786956,
|
| 415 |
+
"loss": 0.6091,
|
| 416 |
+
"step": 2600
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 5.8690744920993225,
|
| 420 |
+
"eval_loss": 0.592084527015686,
|
| 421 |
+
"eval_runtime": 44.6533,
|
| 422 |
+
"eval_samples_per_second": 15.855,
|
| 423 |
+
"eval_steps_per_second": 1.993,
|
| 424 |
+
"eval_wer": 0.52734375,
|
| 425 |
+
"step": 2600
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 6.094808126410835,
|
| 429 |
+
"grad_norm": 1.4033461809158325,
|
| 430 |
+
"learning_rate": 0.00024097801364670205,
|
| 431 |
+
"loss": 0.6168,
|
| 432 |
+
"step": 2700
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 6.094808126410835,
|
| 436 |
+
"eval_loss": 0.5894186496734619,
|
| 437 |
+
"eval_runtime": 45.0539,
|
| 438 |
+
"eval_samples_per_second": 15.715,
|
| 439 |
+
"eval_steps_per_second": 1.975,
|
| 440 |
+
"eval_wer": 0.5275135869565217,
|
| 441 |
+
"step": 2700
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 6.320541760722348,
|
| 445 |
+
"grad_norm": 10.08028507232666,
|
| 446 |
+
"learning_rate": 0.00023870356330553448,
|
| 447 |
+
"loss": 0.6495,
|
| 448 |
+
"step": 2800
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 6.320541760722348,
|
| 452 |
+
"eval_loss": 0.6036040782928467,
|
| 453 |
+
"eval_runtime": 45.0407,
|
| 454 |
+
"eval_samples_per_second": 15.719,
|
| 455 |
+
"eval_steps_per_second": 1.976,
|
| 456 |
+
"eval_wer": 0.5197010869565217,
|
| 457 |
+
"step": 2800
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 6.320541760722348,
|
| 461 |
+
"step": 2800,
|
| 462 |
+
"total_flos": 1.3313268682658783e+19,
|
| 463 |
+
"train_loss": 0.972996187210083,
|
| 464 |
+
"train_runtime": 3933.908,
|
| 465 |
+
"train_samples_per_second": 27.011,
|
| 466 |
+
"train_steps_per_second": 3.378
|
| 467 |
+
}
|
| 468 |
+
],
|
| 469 |
+
"logging_steps": 100,
|
| 470 |
+
"max_steps": 13290,
|
| 471 |
+
"num_input_tokens_seen": 0,
|
| 472 |
+
"num_train_epochs": 30,
|
| 473 |
+
"save_steps": 200,
|
| 474 |
+
"stateful_callbacks": {
|
| 475 |
+
"EarlyStoppingCallback": {
|
| 476 |
+
"args": {
|
| 477 |
+
"early_stopping_patience": 3,
|
| 478 |
+
"early_stopping_threshold": 0.0
|
| 479 |
+
},
|
| 480 |
+
"attributes": {
|
| 481 |
+
"early_stopping_patience_counter": 3
|
| 482 |
+
}
|
| 483 |
+
},
|
| 484 |
+
"TrainerControl": {
|
| 485 |
+
"args": {
|
| 486 |
+
"should_epoch_stop": false,
|
| 487 |
+
"should_evaluate": false,
|
| 488 |
+
"should_log": false,
|
| 489 |
+
"should_save": true,
|
| 490 |
+
"should_training_stop": true
|
| 491 |
+
},
|
| 492 |
+
"attributes": {}
|
| 493 |
+
}
|
| 494 |
+
},
|
| 495 |
+
"total_flos": 1.3313268682658783e+19,
|
| 496 |
+
"train_batch_size": 8,
|
| 497 |
+
"trial_name": null,
|
| 498 |
+
"trial_params": null
|
| 499 |
+
}
|