End of training
Browse files- README.md +5 -2
- 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-bemgen-female-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: 0.1921
|
| 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 |
+
- bemgen
|
| 8 |
+
- mms
|
| 9 |
- generated_from_trainer
|
| 10 |
metrics:
|
| 11 |
- wer
|
|
|
|
| 19 |
|
| 20 |
# mms-1b-bemgen-female-model
|
| 21 |
|
| 22 |
+
This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on the BEMGEN - BEM dataset.
|
| 23 |
It achieves the following results on the evaluation set:
|
| 24 |
- Loss: 0.1921
|
| 25 |
+
- Wer: 0.3440
|
| 26 |
|
| 27 |
## Model description
|
| 28 |
|
adapter.bem.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:24ba3bd1bac48734803e9d8774469e7d7bd277384888d42df4db7b1153dfa5b2
|
| 3 |
+
size 8798532
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.8776978417266186,
|
| 3 |
+
"eval_loss": 0.19208292663097382,
|
| 4 |
+
"eval_runtime": 30.107,
|
| 5 |
+
"eval_samples": 485,
|
| 6 |
+
"eval_samples_per_second": 16.109,
|
| 7 |
+
"eval_steps_per_second": 4.052,
|
| 8 |
+
"eval_wer": 0.3439911797133407,
|
| 9 |
+
"total_flos": 6.630103716413751e+18,
|
| 10 |
+
"train_loss": 0.5161716863087246,
|
| 11 |
+
"train_runtime": 2738.809,
|
| 12 |
+
"train_samples": 3890,
|
| 13 |
+
"train_samples_per_second": 42.61,
|
| 14 |
+
"train_steps_per_second": 10.658
|
| 15 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.8776978417266186,
|
| 3 |
+
"eval_loss": 0.19208292663097382,
|
| 4 |
+
"eval_runtime": 30.107,
|
| 5 |
+
"eval_samples": 485,
|
| 6 |
+
"eval_samples_per_second": 16.109,
|
| 7 |
+
"eval_steps_per_second": 4.052,
|
| 8 |
+
"eval_wer": 0.3439911797133407
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.8776978417266186,
|
| 3 |
+
"total_flos": 6.630103716413751e+18,
|
| 4 |
+
"train_loss": 0.5161716863087246,
|
| 5 |
+
"train_runtime": 2738.809,
|
| 6 |
+
"train_samples": 3890,
|
| 7 |
+
"train_samples_per_second": 42.61,
|
| 8 |
+
"train_steps_per_second": 10.658
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,499 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.1877833753824234,
|
| 3 |
+
"best_model_checkpoint": "/scratch/skscla001/speech/results/mms-1b-bemgen-female-model/checkpoint-2500",
|
| 4 |
+
"epoch": 2.8776978417266186,
|
| 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": 0.10277492291880781,
|
| 13 |
+
"grad_norm": 2.668320894241333,
|
| 14 |
+
"learning_rate": 0.000285,
|
| 15 |
+
"loss": 6.6524,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.10277492291880781,
|
| 20 |
+
"eval_loss": 0.6405563950538635,
|
| 21 |
+
"eval_runtime": 30.4426,
|
| 22 |
+
"eval_samples_per_second": 15.932,
|
| 23 |
+
"eval_steps_per_second": 4.008,
|
| 24 |
+
"eval_wer": 0.6923925027563396,
|
| 25 |
+
"step": 100
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.20554984583761562,
|
| 29 |
+
"grad_norm": 3.9765446186065674,
|
| 30 |
+
"learning_rate": 0.0002990202818838088,
|
| 31 |
+
"loss": 0.4626,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.20554984583761562,
|
| 36 |
+
"eval_loss": 0.27498528361320496,
|
| 37 |
+
"eval_runtime": 30.5092,
|
| 38 |
+
"eval_samples_per_second": 15.897,
|
| 39 |
+
"eval_steps_per_second": 3.999,
|
| 40 |
+
"eval_wer": 0.4520396912899669,
|
| 41 |
+
"step": 200
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.30832476875642345,
|
| 45 |
+
"grad_norm": 1.2895652055740356,
|
| 46 |
+
"learning_rate": 0.00029798899965623925,
|
| 47 |
+
"loss": 0.3384,
|
| 48 |
+
"step": 300
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 0.30832476875642345,
|
| 52 |
+
"eval_loss": 0.2555796205997467,
|
| 53 |
+
"eval_runtime": 30.0486,
|
| 54 |
+
"eval_samples_per_second": 16.141,
|
| 55 |
+
"eval_steps_per_second": 4.06,
|
| 56 |
+
"eval_wer": 0.45865490628445427,
|
| 57 |
+
"step": 300
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.41109969167523125,
|
| 61 |
+
"grad_norm": 8.271388053894043,
|
| 62 |
+
"learning_rate": 0.0002969577174286696,
|
| 63 |
+
"loss": 0.3391,
|
| 64 |
+
"step": 400
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 0.41109969167523125,
|
| 68 |
+
"eval_loss": 0.2390519678592682,
|
| 69 |
+
"eval_runtime": 29.9294,
|
| 70 |
+
"eval_samples_per_second": 16.205,
|
| 71 |
+
"eval_steps_per_second": 4.076,
|
| 72 |
+
"eval_wer": 0.4119073869900772,
|
| 73 |
+
"step": 400
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.513874614594039,
|
| 77 |
+
"grad_norm": 1.8543354272842407,
|
| 78 |
+
"learning_rate": 0.0002959264352011,
|
| 79 |
+
"loss": 0.2928,
|
| 80 |
+
"step": 500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.513874614594039,
|
| 84 |
+
"eval_loss": 0.23259302973747253,
|
| 85 |
+
"eval_runtime": 30.0544,
|
| 86 |
+
"eval_samples_per_second": 16.137,
|
| 87 |
+
"eval_steps_per_second": 4.059,
|
| 88 |
+
"eval_wer": 0.40088202866593164,
|
| 89 |
+
"step": 500
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.6166495375128469,
|
| 93 |
+
"grad_norm": 1.9590275287628174,
|
| 94 |
+
"learning_rate": 0.0002948951529735304,
|
| 95 |
+
"loss": 0.3126,
|
| 96 |
+
"step": 600
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 0.6166495375128469,
|
| 100 |
+
"eval_loss": 0.22746512293815613,
|
| 101 |
+
"eval_runtime": 30.1494,
|
| 102 |
+
"eval_samples_per_second": 16.087,
|
| 103 |
+
"eval_steps_per_second": 4.047,
|
| 104 |
+
"eval_wer": 0.40308710033076073,
|
| 105 |
+
"step": 600
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.7194244604316546,
|
| 109 |
+
"grad_norm": 1.9041965007781982,
|
| 110 |
+
"learning_rate": 0.0002938638707459608,
|
| 111 |
+
"loss": 0.3305,
|
| 112 |
+
"step": 700
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 0.7194244604316546,
|
| 116 |
+
"eval_loss": 0.22082751989364624,
|
| 117 |
+
"eval_runtime": 29.8842,
|
| 118 |
+
"eval_samples_per_second": 16.229,
|
| 119 |
+
"eval_steps_per_second": 4.082,
|
| 120 |
+
"eval_wer": 0.3977949283351709,
|
| 121 |
+
"step": 700
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.8221993833504625,
|
| 125 |
+
"grad_norm": 1.5732944011688232,
|
| 126 |
+
"learning_rate": 0.00029283258851839117,
|
| 127 |
+
"loss": 0.3043,
|
| 128 |
+
"step": 800
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.8221993833504625,
|
| 132 |
+
"eval_loss": 0.21230436861515045,
|
| 133 |
+
"eval_runtime": 29.8824,
|
| 134 |
+
"eval_samples_per_second": 16.23,
|
| 135 |
+
"eval_steps_per_second": 4.083,
|
| 136 |
+
"eval_wer": 0.38037486218302097,
|
| 137 |
+
"step": 800
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.9249743062692704,
|
| 141 |
+
"grad_norm": 1.669620156288147,
|
| 142 |
+
"learning_rate": 0.00029180130629082154,
|
| 143 |
+
"loss": 0.2989,
|
| 144 |
+
"step": 900
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.9249743062692704,
|
| 148 |
+
"eval_loss": 0.21349409222602844,
|
| 149 |
+
"eval_runtime": 30.0481,
|
| 150 |
+
"eval_samples_per_second": 16.141,
|
| 151 |
+
"eval_steps_per_second": 4.06,
|
| 152 |
+
"eval_wer": 0.37927232635060637,
|
| 153 |
+
"step": 900
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 1.027749229188078,
|
| 157 |
+
"grad_norm": 1.0011438131332397,
|
| 158 |
+
"learning_rate": 0.00029077002406325197,
|
| 159 |
+
"loss": 0.2911,
|
| 160 |
+
"step": 1000
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 1.027749229188078,
|
| 164 |
+
"eval_loss": 0.21172800660133362,
|
| 165 |
+
"eval_runtime": 30.6543,
|
| 166 |
+
"eval_samples_per_second": 15.822,
|
| 167 |
+
"eval_steps_per_second": 3.98,
|
| 168 |
+
"eval_wer": 0.38897464167585444,
|
| 169 |
+
"step": 1000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 1.1305241521068858,
|
| 173 |
+
"grad_norm": 0.9472767114639282,
|
| 174 |
+
"learning_rate": 0.00028973874183568234,
|
| 175 |
+
"loss": 0.2994,
|
| 176 |
+
"step": 1100
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 1.1305241521068858,
|
| 180 |
+
"eval_loss": 0.20664845407009125,
|
| 181 |
+
"eval_runtime": 29.9146,
|
| 182 |
+
"eval_samples_per_second": 16.213,
|
| 183 |
+
"eval_steps_per_second": 4.078,
|
| 184 |
+
"eval_wer": 0.38676957001102535,
|
| 185 |
+
"step": 1100
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 1.2332990750256938,
|
| 189 |
+
"grad_norm": 0.9638025760650635,
|
| 190 |
+
"learning_rate": 0.0002887074596081127,
|
| 191 |
+
"loss": 0.2849,
|
| 192 |
+
"step": 1200
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 1.2332990750256938,
|
| 196 |
+
"eval_loss": 0.21126343309879303,
|
| 197 |
+
"eval_runtime": 29.8301,
|
| 198 |
+
"eval_samples_per_second": 16.259,
|
| 199 |
+
"eval_steps_per_second": 4.09,
|
| 200 |
+
"eval_wer": 0.39316427783902974,
|
| 201 |
+
"step": 1200
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 1.3360739979445015,
|
| 205 |
+
"grad_norm": 1.023758053779602,
|
| 206 |
+
"learning_rate": 0.00028767617738054314,
|
| 207 |
+
"loss": 0.2864,
|
| 208 |
+
"step": 1300
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 1.3360739979445015,
|
| 212 |
+
"eval_loss": 0.20175151526927948,
|
| 213 |
+
"eval_runtime": 30.0501,
|
| 214 |
+
"eval_samples_per_second": 16.14,
|
| 215 |
+
"eval_steps_per_second": 4.06,
|
| 216 |
+
"eval_wer": 0.3713340683572216,
|
| 217 |
+
"step": 1300
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 1.4388489208633093,
|
| 221 |
+
"grad_norm": 1.0510107278823853,
|
| 222 |
+
"learning_rate": 0.0002866448951529735,
|
| 223 |
+
"loss": 0.2611,
|
| 224 |
+
"step": 1400
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.4388489208633093,
|
| 228 |
+
"eval_loss": 0.2004697620868683,
|
| 229 |
+
"eval_runtime": 30.4532,
|
| 230 |
+
"eval_samples_per_second": 15.926,
|
| 231 |
+
"eval_steps_per_second": 4.006,
|
| 232 |
+
"eval_wer": 0.4004410143329658,
|
| 233 |
+
"step": 1400
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 1.541623843782117,
|
| 237 |
+
"grad_norm": 1.148527979850769,
|
| 238 |
+
"learning_rate": 0.0002856136129254039,
|
| 239 |
+
"loss": 0.2995,
|
| 240 |
+
"step": 1500
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 1.541623843782117,
|
| 244 |
+
"eval_loss": 0.1996380090713501,
|
| 245 |
+
"eval_runtime": 29.9115,
|
| 246 |
+
"eval_samples_per_second": 16.214,
|
| 247 |
+
"eval_steps_per_second": 4.079,
|
| 248 |
+
"eval_wer": 0.372877618522602,
|
| 249 |
+
"step": 1500
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 1.644398766700925,
|
| 253 |
+
"grad_norm": 0.7263904809951782,
|
| 254 |
+
"learning_rate": 0.00028458233069783426,
|
| 255 |
+
"loss": 0.2787,
|
| 256 |
+
"step": 1600
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 1.644398766700925,
|
| 260 |
+
"eval_loss": 0.2014673948287964,
|
| 261 |
+
"eval_runtime": 29.8924,
|
| 262 |
+
"eval_samples_per_second": 16.225,
|
| 263 |
+
"eval_steps_per_second": 4.081,
|
| 264 |
+
"eval_wer": 0.3647188533627343,
|
| 265 |
+
"step": 1600
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 1.7471736896197327,
|
| 269 |
+
"grad_norm": 0.9115646481513977,
|
| 270 |
+
"learning_rate": 0.0002835510484702647,
|
| 271 |
+
"loss": 0.2444,
|
| 272 |
+
"step": 1700
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.7471736896197327,
|
| 276 |
+
"eval_loss": 0.19884684681892395,
|
| 277 |
+
"eval_runtime": 30.122,
|
| 278 |
+
"eval_samples_per_second": 16.101,
|
| 279 |
+
"eval_steps_per_second": 4.05,
|
| 280 |
+
"eval_wer": 0.3603087100330761,
|
| 281 |
+
"step": 1700
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.8499486125385407,
|
| 285 |
+
"grad_norm": 0.9806845784187317,
|
| 286 |
+
"learning_rate": 0.00028251976624269506,
|
| 287 |
+
"loss": 0.2734,
|
| 288 |
+
"step": 1800
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 1.8499486125385407,
|
| 292 |
+
"eval_loss": 0.19500485062599182,
|
| 293 |
+
"eval_runtime": 30.0764,
|
| 294 |
+
"eval_samples_per_second": 16.126,
|
| 295 |
+
"eval_steps_per_second": 4.056,
|
| 296 |
+
"eval_wer": 0.3589856670341786,
|
| 297 |
+
"step": 1800
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 1.9527235354573484,
|
| 301 |
+
"grad_norm": 0.7178159356117249,
|
| 302 |
+
"learning_rate": 0.00028148848401512544,
|
| 303 |
+
"loss": 0.2794,
|
| 304 |
+
"step": 1900
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 1.9527235354573484,
|
| 308 |
+
"eval_loss": 0.19534997642040253,
|
| 309 |
+
"eval_runtime": 29.8724,
|
| 310 |
+
"eval_samples_per_second": 16.236,
|
| 311 |
+
"eval_steps_per_second": 4.084,
|
| 312 |
+
"eval_wer": 0.3545755237045204,
|
| 313 |
+
"step": 1900
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 2.055498458376156,
|
| 317 |
+
"grad_norm": 1.4567710161209106,
|
| 318 |
+
"learning_rate": 0.00028045720178755586,
|
| 319 |
+
"loss": 0.2708,
|
| 320 |
+
"step": 2000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 2.055498458376156,
|
| 324 |
+
"eval_loss": 0.19343802332878113,
|
| 325 |
+
"eval_runtime": 29.925,
|
| 326 |
+
"eval_samples_per_second": 16.207,
|
| 327 |
+
"eval_steps_per_second": 4.077,
|
| 328 |
+
"eval_wer": 0.36097023153252483,
|
| 329 |
+
"step": 2000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 2.158273381294964,
|
| 333 |
+
"grad_norm": 0.5352274179458618,
|
| 334 |
+
"learning_rate": 0.00027942591955998624,
|
| 335 |
+
"loss": 0.2545,
|
| 336 |
+
"step": 2100
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 2.158273381294964,
|
| 340 |
+
"eval_loss": 0.19527685642242432,
|
| 341 |
+
"eval_runtime": 30.2437,
|
| 342 |
+
"eval_samples_per_second": 16.036,
|
| 343 |
+
"eval_steps_per_second": 4.034,
|
| 344 |
+
"eval_wer": 0.36163175303197354,
|
| 345 |
+
"step": 2100
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 2.2610483042137717,
|
| 349 |
+
"grad_norm": 0.7689425349235535,
|
| 350 |
+
"learning_rate": 0.0002784049501546923,
|
| 351 |
+
"loss": 0.2529,
|
| 352 |
+
"step": 2200
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 2.2610483042137717,
|
| 356 |
+
"eval_loss": 0.19401085376739502,
|
| 357 |
+
"eval_runtime": 30.2778,
|
| 358 |
+
"eval_samples_per_second": 16.018,
|
| 359 |
+
"eval_steps_per_second": 4.029,
|
| 360 |
+
"eval_wer": 0.35589856670341785,
|
| 361 |
+
"step": 2200
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 2.3638232271325794,
|
| 365 |
+
"grad_norm": 0.518496572971344,
|
| 366 |
+
"learning_rate": 0.0002773736679271227,
|
| 367 |
+
"loss": 0.2628,
|
| 368 |
+
"step": 2300
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 2.3638232271325794,
|
| 372 |
+
"eval_loss": 0.1926334649324417,
|
| 373 |
+
"eval_runtime": 30.0716,
|
| 374 |
+
"eval_samples_per_second": 16.128,
|
| 375 |
+
"eval_steps_per_second": 4.057,
|
| 376 |
+
"eval_wer": 0.3585446527012128,
|
| 377 |
+
"step": 2300
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 2.4665981500513876,
|
| 381 |
+
"grad_norm": 0.47154027223587036,
|
| 382 |
+
"learning_rate": 0.0002763423856995531,
|
| 383 |
+
"loss": 0.2788,
|
| 384 |
+
"step": 2400
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 2.4665981500513876,
|
| 388 |
+
"eval_loss": 0.1925119012594223,
|
| 389 |
+
"eval_runtime": 29.9369,
|
| 390 |
+
"eval_samples_per_second": 16.201,
|
| 391 |
+
"eval_steps_per_second": 4.075,
|
| 392 |
+
"eval_wer": 0.35104740904079385,
|
| 393 |
+
"step": 2400
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 2.5693730729701953,
|
| 397 |
+
"grad_norm": 0.4298442304134369,
|
| 398 |
+
"learning_rate": 0.0002753111034719835,
|
| 399 |
+
"loss": 0.2473,
|
| 400 |
+
"step": 2500
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 2.5693730729701953,
|
| 404 |
+
"eval_loss": 0.1877833753824234,
|
| 405 |
+
"eval_runtime": 30.3202,
|
| 406 |
+
"eval_samples_per_second": 15.996,
|
| 407 |
+
"eval_steps_per_second": 4.024,
|
| 408 |
+
"eval_wer": 0.34994487320837925,
|
| 409 |
+
"step": 2500
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 2.672147995889003,
|
| 413 |
+
"grad_norm": 1.0587983131408691,
|
| 414 |
+
"learning_rate": 0.0002742798212444139,
|
| 415 |
+
"loss": 0.2595,
|
| 416 |
+
"step": 2600
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.672147995889003,
|
| 420 |
+
"eval_loss": 0.19105447828769684,
|
| 421 |
+
"eval_runtime": 30.2772,
|
| 422 |
+
"eval_samples_per_second": 16.019,
|
| 423 |
+
"eval_steps_per_second": 4.029,
|
| 424 |
+
"eval_wer": 0.37045203969129,
|
| 425 |
+
"step": 2600
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 2.774922918807811,
|
| 429 |
+
"grad_norm": 0.7508417963981628,
|
| 430 |
+
"learning_rate": 0.00027324853901684426,
|
| 431 |
+
"loss": 0.2516,
|
| 432 |
+
"step": 2700
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 2.774922918807811,
|
| 436 |
+
"eval_loss": 0.188262939453125,
|
| 437 |
+
"eval_runtime": 30.0618,
|
| 438 |
+
"eval_samples_per_second": 16.133,
|
| 439 |
+
"eval_steps_per_second": 4.058,
|
| 440 |
+
"eval_wer": 0.3479603087100331,
|
| 441 |
+
"step": 2700
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 2.8776978417266186,
|
| 445 |
+
"grad_norm": 1.4504516124725342,
|
| 446 |
+
"learning_rate": 0.00027221725678927463,
|
| 447 |
+
"loss": 0.2445,
|
| 448 |
+
"step": 2800
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 2.8776978417266186,
|
| 452 |
+
"eval_loss": 0.19207896292209625,
|
| 453 |
+
"eval_runtime": 30.0543,
|
| 454 |
+
"eval_samples_per_second": 16.137,
|
| 455 |
+
"eval_steps_per_second": 4.059,
|
| 456 |
+
"eval_wer": 0.3442116868798236,
|
| 457 |
+
"step": 2800
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 2.8776978417266186,
|
| 461 |
+
"step": 2800,
|
| 462 |
+
"total_flos": 6.630103716413751e+18,
|
| 463 |
+
"train_loss": 0.5161716863087246,
|
| 464 |
+
"train_runtime": 2738.809,
|
| 465 |
+
"train_samples_per_second": 42.61,
|
| 466 |
+
"train_steps_per_second": 10.658
|
| 467 |
+
}
|
| 468 |
+
],
|
| 469 |
+
"logging_steps": 100,
|
| 470 |
+
"max_steps": 29190,
|
| 471 |
+
"num_input_tokens_seen": 0,
|
| 472 |
+
"num_train_epochs": 30,
|
| 473 |
+
"save_steps": 400,
|
| 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": 6.630103716413751e+18,
|
| 496 |
+
"train_batch_size": 4,
|
| 497 |
+
"trial_name": null,
|
| 498 |
+
"trial_params": null
|
| 499 |
+
}
|