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 +531 -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-male-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: 0.3056
|
| 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-male-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: 0.3056
|
| 25 |
+
- Wer: 0.4399
|
| 26 |
|
| 27 |
## Model description
|
| 28 |
|
adapter.bem.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:515beea7f632c6767fb96e98f2af2791ece900ae37036f540a6a933ad61183db
|
| 3 |
+
size 8798532
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.1023784901758016,
|
| 3 |
+
"eval_loss": 0.30557817220687866,
|
| 4 |
+
"eval_runtime": 31.0257,
|
| 5 |
+
"eval_samples": 491,
|
| 6 |
+
"eval_samples_per_second": 15.826,
|
| 7 |
+
"eval_steps_per_second": 3.964,
|
| 8 |
+
"eval_wer": 0.4399312271652697,
|
| 9 |
+
"total_flos": 7.152937185600249e+18,
|
| 10 |
+
"train_loss": 0.6818474756876628,
|
| 11 |
+
"train_runtime": 2982.3235,
|
| 12 |
+
"train_samples": 3866,
|
| 13 |
+
"train_samples_per_second": 38.889,
|
| 14 |
+
"train_steps_per_second": 9.727
|
| 15 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.1023784901758016,
|
| 3 |
+
"eval_loss": 0.30557817220687866,
|
| 4 |
+
"eval_runtime": 31.0257,
|
| 5 |
+
"eval_samples": 491,
|
| 6 |
+
"eval_samples_per_second": 15.826,
|
| 7 |
+
"eval_steps_per_second": 3.964,
|
| 8 |
+
"eval_wer": 0.4399312271652697
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.1023784901758016,
|
| 3 |
+
"total_flos": 7.152937185600249e+18,
|
| 4 |
+
"train_loss": 0.6818474756876628,
|
| 5 |
+
"train_runtime": 2982.3235,
|
| 6 |
+
"train_samples": 3866,
|
| 7 |
+
"train_samples_per_second": 38.889,
|
| 8 |
+
"train_steps_per_second": 9.727
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,531 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.2995806932449341,
|
| 3 |
+
"best_model_checkpoint": "/scratch/skscla001/speech/results/mms-1b-bemgen-male-model/checkpoint-2700",
|
| 4 |
+
"epoch": 3.1023784901758016,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 3000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.10341261633919338,
|
| 13 |
+
"grad_norm": 3.187527894973755,
|
| 14 |
+
"learning_rate": 0.000285,
|
| 15 |
+
"loss": 7.0611,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.10341261633919338,
|
| 20 |
+
"eval_loss": 1.4202574491500854,
|
| 21 |
+
"eval_runtime": 30.7709,
|
| 22 |
+
"eval_samples_per_second": 15.957,
|
| 23 |
+
"eval_steps_per_second": 3.997,
|
| 24 |
+
"eval_wer": 0.9989254244573393,
|
| 25 |
+
"step": 100
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.20682523267838676,
|
| 29 |
+
"grad_norm": 2.534780263900757,
|
| 30 |
+
"learning_rate": 0.000299014181943964,
|
| 31 |
+
"loss": 0.7562,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.20682523267838676,
|
| 36 |
+
"eval_loss": 0.42598676681518555,
|
| 37 |
+
"eval_runtime": 30.7262,
|
| 38 |
+
"eval_samples_per_second": 15.98,
|
| 39 |
+
"eval_steps_per_second": 4.003,
|
| 40 |
+
"eval_wer": 0.5807006232538148,
|
| 41 |
+
"step": 200
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.31023784901758017,
|
| 45 |
+
"grad_norm": 1.9200867414474487,
|
| 46 |
+
"learning_rate": 0.000297976478727084,
|
| 47 |
+
"loss": 0.5464,
|
| 48 |
+
"step": 300
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 0.31023784901758017,
|
| 52 |
+
"eval_loss": 0.38771751523017883,
|
| 53 |
+
"eval_runtime": 30.4983,
|
| 54 |
+
"eval_samples_per_second": 16.099,
|
| 55 |
+
"eval_steps_per_second": 4.033,
|
| 56 |
+
"eval_wer": 0.5409413281753708,
|
| 57 |
+
"step": 300
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.4136504653567735,
|
| 61 |
+
"grad_norm": 3.3704230785369873,
|
| 62 |
+
"learning_rate": 0.0002969387755102041,
|
| 63 |
+
"loss": 0.5399,
|
| 64 |
+
"step": 400
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 0.4136504653567735,
|
| 68 |
+
"eval_loss": 0.3693784475326538,
|
| 69 |
+
"eval_runtime": 30.624,
|
| 70 |
+
"eval_samples_per_second": 16.033,
|
| 71 |
+
"eval_steps_per_second": 4.016,
|
| 72 |
+
"eval_wer": 0.539651837524178,
|
| 73 |
+
"step": 400
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.5170630816959669,
|
| 77 |
+
"grad_norm": 4.073702335357666,
|
| 78 |
+
"learning_rate": 0.0002959010722933241,
|
| 79 |
+
"loss": 0.5276,
|
| 80 |
+
"step": 500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.5170630816959669,
|
| 84 |
+
"eval_loss": 0.3597019612789154,
|
| 85 |
+
"eval_runtime": 30.927,
|
| 86 |
+
"eval_samples_per_second": 15.876,
|
| 87 |
+
"eval_steps_per_second": 3.977,
|
| 88 |
+
"eval_wer": 0.514936600042983,
|
| 89 |
+
"step": 500
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.6204756980351603,
|
| 93 |
+
"grad_norm": 1.7590080499649048,
|
| 94 |
+
"learning_rate": 0.0002948633690764441,
|
| 95 |
+
"loss": 0.4763,
|
| 96 |
+
"step": 600
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 0.6204756980351603,
|
| 100 |
+
"eval_loss": 0.3516247868537903,
|
| 101 |
+
"eval_runtime": 30.7985,
|
| 102 |
+
"eval_samples_per_second": 15.942,
|
| 103 |
+
"eval_steps_per_second": 3.994,
|
| 104 |
+
"eval_wer": 0.49258542875564154,
|
| 105 |
+
"step": 600
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.7238883143743536,
|
| 109 |
+
"grad_norm": 2.278167486190796,
|
| 110 |
+
"learning_rate": 0.00029382566585956415,
|
| 111 |
+
"loss": 0.4902,
|
| 112 |
+
"step": 700
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 0.7238883143743536,
|
| 116 |
+
"eval_loss": 0.3478095233440399,
|
| 117 |
+
"eval_runtime": 30.6978,
|
| 118 |
+
"eval_samples_per_second": 15.995,
|
| 119 |
+
"eval_steps_per_second": 4.007,
|
| 120 |
+
"eval_wer": 0.5121427036320654,
|
| 121 |
+
"step": 700
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.827300930713547,
|
| 125 |
+
"grad_norm": 2.4464170932769775,
|
| 126 |
+
"learning_rate": 0.0002927879626426842,
|
| 127 |
+
"loss": 0.4797,
|
| 128 |
+
"step": 800
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.827300930713547,
|
| 132 |
+
"eval_loss": 0.3448158800601959,
|
| 133 |
+
"eval_runtime": 30.6966,
|
| 134 |
+
"eval_samples_per_second": 15.995,
|
| 135 |
+
"eval_steps_per_second": 4.007,
|
| 136 |
+
"eval_wer": 0.4794756071351816,
|
| 137 |
+
"step": 800
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.9307135470527405,
|
| 141 |
+
"grad_norm": 2.2667806148529053,
|
| 142 |
+
"learning_rate": 0.0002917502594258042,
|
| 143 |
+
"loss": 0.4667,
|
| 144 |
+
"step": 900
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.9307135470527405,
|
| 148 |
+
"eval_loss": 0.3473311960697174,
|
| 149 |
+
"eval_runtime": 30.9122,
|
| 150 |
+
"eval_samples_per_second": 15.884,
|
| 151 |
+
"eval_steps_per_second": 3.979,
|
| 152 |
+
"eval_wer": 0.491725768321513,
|
| 153 |
+
"step": 900
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 1.0341261633919339,
|
| 157 |
+
"grad_norm": 4.21267032623291,
|
| 158 |
+
"learning_rate": 0.00029071255620892423,
|
| 159 |
+
"loss": 0.4426,
|
| 160 |
+
"step": 1000
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 1.0341261633919339,
|
| 164 |
+
"eval_loss": 0.33737972378730774,
|
| 165 |
+
"eval_runtime": 30.8508,
|
| 166 |
+
"eval_samples_per_second": 15.915,
|
| 167 |
+
"eval_steps_per_second": 3.987,
|
| 168 |
+
"eval_wer": 0.47539222007307114,
|
| 169 |
+
"step": 1000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 1.1375387797311272,
|
| 173 |
+
"grad_norm": 1.35986328125,
|
| 174 |
+
"learning_rate": 0.00028967485299204426,
|
| 175 |
+
"loss": 0.4491,
|
| 176 |
+
"step": 1100
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 1.1375387797311272,
|
| 180 |
+
"eval_loss": 0.33271169662475586,
|
| 181 |
+
"eval_runtime": 30.6479,
|
| 182 |
+
"eval_samples_per_second": 16.021,
|
| 183 |
+
"eval_steps_per_second": 4.013,
|
| 184 |
+
"eval_wer": 0.48269933376316354,
|
| 185 |
+
"step": 1100
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 1.2409513960703205,
|
| 189 |
+
"grad_norm": 0.8853399753570557,
|
| 190 |
+
"learning_rate": 0.0002886371497751643,
|
| 191 |
+
"loss": 0.4447,
|
| 192 |
+
"step": 1200
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 1.2409513960703205,
|
| 196 |
+
"eval_loss": 0.33263838291168213,
|
| 197 |
+
"eval_runtime": 30.6698,
|
| 198 |
+
"eval_samples_per_second": 16.009,
|
| 199 |
+
"eval_steps_per_second": 4.01,
|
| 200 |
+
"eval_wer": 0.4915108532129809,
|
| 201 |
+
"step": 1200
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 1.344364012409514,
|
| 205 |
+
"grad_norm": 1.4764142036437988,
|
| 206 |
+
"learning_rate": 0.0002875994465582843,
|
| 207 |
+
"loss": 0.4573,
|
| 208 |
+
"step": 1300
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 1.344364012409514,
|
| 212 |
+
"eval_loss": 0.32816657423973083,
|
| 213 |
+
"eval_runtime": 30.9107,
|
| 214 |
+
"eval_samples_per_second": 15.884,
|
| 215 |
+
"eval_steps_per_second": 3.979,
|
| 216 |
+
"eval_wer": 0.46378680421233615,
|
| 217 |
+
"step": 1300
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 1.4477766287487073,
|
| 221 |
+
"grad_norm": 1.5062446594238281,
|
| 222 |
+
"learning_rate": 0.00028656174334140434,
|
| 223 |
+
"loss": 0.4724,
|
| 224 |
+
"step": 1400
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.4477766287487073,
|
| 228 |
+
"eval_loss": 0.32294726371765137,
|
| 229 |
+
"eval_runtime": 30.9781,
|
| 230 |
+
"eval_samples_per_second": 15.85,
|
| 231 |
+
"eval_steps_per_second": 3.971,
|
| 232 |
+
"eval_wer": 0.4584139264990329,
|
| 233 |
+
"step": 1400
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 1.5511892450879006,
|
| 237 |
+
"grad_norm": 13.068098068237305,
|
| 238 |
+
"learning_rate": 0.00028552404012452436,
|
| 239 |
+
"loss": 0.4502,
|
| 240 |
+
"step": 1500
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 1.5511892450879006,
|
| 244 |
+
"eval_loss": 0.320773184299469,
|
| 245 |
+
"eval_runtime": 30.6511,
|
| 246 |
+
"eval_samples_per_second": 16.019,
|
| 247 |
+
"eval_steps_per_second": 4.013,
|
| 248 |
+
"eval_wer": 0.4558349451966473,
|
| 249 |
+
"step": 1500
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 1.654601861427094,
|
| 253 |
+
"grad_norm": 1.0407159328460693,
|
| 254 |
+
"learning_rate": 0.0002844863369076444,
|
| 255 |
+
"loss": 0.4448,
|
| 256 |
+
"step": 1600
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 1.654601861427094,
|
| 260 |
+
"eval_loss": 0.31659701466560364,
|
| 261 |
+
"eval_runtime": 30.8376,
|
| 262 |
+
"eval_samples_per_second": 15.922,
|
| 263 |
+
"eval_steps_per_second": 3.989,
|
| 264 |
+
"eval_wer": 0.4558349451966473,
|
| 265 |
+
"step": 1600
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 1.7580144777662876,
|
| 269 |
+
"grad_norm": 1.3918695449829102,
|
| 270 |
+
"learning_rate": 0.0002834486336907644,
|
| 271 |
+
"loss": 0.4549,
|
| 272 |
+
"step": 1700
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.7580144777662876,
|
| 276 |
+
"eval_loss": 0.32059985399246216,
|
| 277 |
+
"eval_runtime": 30.9591,
|
| 278 |
+
"eval_samples_per_second": 15.86,
|
| 279 |
+
"eval_steps_per_second": 3.973,
|
| 280 |
+
"eval_wer": 0.4513217279174726,
|
| 281 |
+
"step": 1700
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.861427094105481,
|
| 285 |
+
"grad_norm": 1.0892579555511475,
|
| 286 |
+
"learning_rate": 0.00028241093047388444,
|
| 287 |
+
"loss": 0.4544,
|
| 288 |
+
"step": 1800
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 1.861427094105481,
|
| 292 |
+
"eval_loss": 0.31517136096954346,
|
| 293 |
+
"eval_runtime": 30.9782,
|
| 294 |
+
"eval_samples_per_second": 15.85,
|
| 295 |
+
"eval_steps_per_second": 3.971,
|
| 296 |
+
"eval_wer": 0.4485278315065549,
|
| 297 |
+
"step": 1800
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 1.9648397104446742,
|
| 301 |
+
"grad_norm": 1.5489096641540527,
|
| 302 |
+
"learning_rate": 0.00028137322725700447,
|
| 303 |
+
"loss": 0.4233,
|
| 304 |
+
"step": 1900
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 1.9648397104446742,
|
| 308 |
+
"eval_loss": 0.3146161437034607,
|
| 309 |
+
"eval_runtime": 30.6737,
|
| 310 |
+
"eval_samples_per_second": 16.007,
|
| 311 |
+
"eval_steps_per_second": 4.01,
|
| 312 |
+
"eval_wer": 0.43950139694820545,
|
| 313 |
+
"step": 1900
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 2.0682523267838677,
|
| 317 |
+
"grad_norm": 0.7643145322799683,
|
| 318 |
+
"learning_rate": 0.0002803355240401245,
|
| 319 |
+
"loss": 0.4452,
|
| 320 |
+
"step": 2000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 2.0682523267838677,
|
| 324 |
+
"eval_loss": 0.3116036057472229,
|
| 325 |
+
"eval_runtime": 30.6446,
|
| 326 |
+
"eval_samples_per_second": 16.022,
|
| 327 |
+
"eval_steps_per_second": 4.014,
|
| 328 |
+
"eval_wer": 0.46335697399527187,
|
| 329 |
+
"step": 2000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 2.1716649431230612,
|
| 333 |
+
"grad_norm": 2.4722142219543457,
|
| 334 |
+
"learning_rate": 0.0002792978208232445,
|
| 335 |
+
"loss": 0.4171,
|
| 336 |
+
"step": 2100
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 2.1716649431230612,
|
| 340 |
+
"eval_loss": 0.31308668851852417,
|
| 341 |
+
"eval_runtime": 30.6916,
|
| 342 |
+
"eval_samples_per_second": 15.998,
|
| 343 |
+
"eval_steps_per_second": 4.008,
|
| 344 |
+
"eval_wer": 0.465291209972061,
|
| 345 |
+
"step": 2100
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 2.2750775594622543,
|
| 349 |
+
"grad_norm": 1.3434455394744873,
|
| 350 |
+
"learning_rate": 0.00027826011760636454,
|
| 351 |
+
"loss": 0.4161,
|
| 352 |
+
"step": 2200
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 2.2750775594622543,
|
| 356 |
+
"eval_loss": 0.3058585226535797,
|
| 357 |
+
"eval_runtime": 30.8818,
|
| 358 |
+
"eval_samples_per_second": 15.899,
|
| 359 |
+
"eval_steps_per_second": 3.983,
|
| 360 |
+
"eval_wer": 0.4455190199871051,
|
| 361 |
+
"step": 2200
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 2.378490175801448,
|
| 365 |
+
"grad_norm": 1.7992069721221924,
|
| 366 |
+
"learning_rate": 0.00027722241438948457,
|
| 367 |
+
"loss": 0.4126,
|
| 368 |
+
"step": 2300
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 2.378490175801448,
|
| 372 |
+
"eval_loss": 0.304585337638855,
|
| 373 |
+
"eval_runtime": 31.0224,
|
| 374 |
+
"eval_samples_per_second": 15.827,
|
| 375 |
+
"eval_steps_per_second": 3.965,
|
| 376 |
+
"eval_wer": 0.437137330754352,
|
| 377 |
+
"step": 2300
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 2.481902792140641,
|
| 381 |
+
"grad_norm": 0.9206439852714539,
|
| 382 |
+
"learning_rate": 0.0002761847111726046,
|
| 383 |
+
"loss": 0.4337,
|
| 384 |
+
"step": 2400
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 2.481902792140641,
|
| 388 |
+
"eval_loss": 0.3068229854106903,
|
| 389 |
+
"eval_runtime": 30.6081,
|
| 390 |
+
"eval_samples_per_second": 16.041,
|
| 391 |
+
"eval_steps_per_second": 4.019,
|
| 392 |
+
"eval_wer": 0.43434343434343436,
|
| 393 |
+
"step": 2400
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 2.5853154084798344,
|
| 397 |
+
"grad_norm": 0.7960453629493713,
|
| 398 |
+
"learning_rate": 0.0002751470079557246,
|
| 399 |
+
"loss": 0.4175,
|
| 400 |
+
"step": 2500
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 2.5853154084798344,
|
| 404 |
+
"eval_loss": 0.3018108904361725,
|
| 405 |
+
"eval_runtime": 31.0479,
|
| 406 |
+
"eval_samples_per_second": 15.814,
|
| 407 |
+
"eval_steps_per_second": 3.962,
|
| 408 |
+
"eval_wer": 0.43606275521169136,
|
| 409 |
+
"step": 2500
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 2.688728024819028,
|
| 413 |
+
"grad_norm": 0.8274104595184326,
|
| 414 |
+
"learning_rate": 0.0002741093047388447,
|
| 415 |
+
"loss": 0.4097,
|
| 416 |
+
"step": 2600
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.688728024819028,
|
| 420 |
+
"eval_loss": 0.3014011085033417,
|
| 421 |
+
"eval_runtime": 30.9733,
|
| 422 |
+
"eval_samples_per_second": 15.852,
|
| 423 |
+
"eval_steps_per_second": 3.971,
|
| 424 |
+
"eval_wer": 0.4410058027079304,
|
| 425 |
+
"step": 2600
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 2.7921406411582215,
|
| 429 |
+
"grad_norm": 0.6941657066345215,
|
| 430 |
+
"learning_rate": 0.00027307160152196467,
|
| 431 |
+
"loss": 0.4137,
|
| 432 |
+
"step": 2700
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 2.7921406411582215,
|
| 436 |
+
"eval_loss": 0.2995806932449341,
|
| 437 |
+
"eval_runtime": 30.8058,
|
| 438 |
+
"eval_samples_per_second": 15.939,
|
| 439 |
+
"eval_steps_per_second": 3.993,
|
| 440 |
+
"eval_wer": 0.43541800988609497,
|
| 441 |
+
"step": 2700
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 2.8955532574974145,
|
| 445 |
+
"grad_norm": 1.414363145828247,
|
| 446 |
+
"learning_rate": 0.0002720442753372535,
|
| 447 |
+
"loss": 0.4167,
|
| 448 |
+
"step": 2800
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 2.8955532574974145,
|
| 452 |
+
"eval_loss": 0.2997954189777374,
|
| 453 |
+
"eval_runtime": 30.7397,
|
| 454 |
+
"eval_samples_per_second": 15.973,
|
| 455 |
+
"eval_steps_per_second": 4.001,
|
| 456 |
+
"eval_wer": 0.4399312271652697,
|
| 457 |
+
"step": 2800
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 2.998965873836608,
|
| 461 |
+
"grad_norm": 8.619386672973633,
|
| 462 |
+
"learning_rate": 0.00027100657212037355,
|
| 463 |
+
"loss": 0.427,
|
| 464 |
+
"step": 2900
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.998965873836608,
|
| 468 |
+
"eval_loss": 0.30078718066215515,
|
| 469 |
+
"eval_runtime": 31.0317,
|
| 470 |
+
"eval_samples_per_second": 15.823,
|
| 471 |
+
"eval_steps_per_second": 3.964,
|
| 472 |
+
"eval_wer": 0.4435847840103159,
|
| 473 |
+
"step": 2900
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 3.1023784901758016,
|
| 477 |
+
"grad_norm": 35.736480712890625,
|
| 478 |
+
"learning_rate": 0.0002699688689034936,
|
| 479 |
+
"loss": 0.4082,
|
| 480 |
+
"step": 3000
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 3.1023784901758016,
|
| 484 |
+
"eval_loss": 0.3055952489376068,
|
| 485 |
+
"eval_runtime": 30.9565,
|
| 486 |
+
"eval_samples_per_second": 15.861,
|
| 487 |
+
"eval_steps_per_second": 3.973,
|
| 488 |
+
"eval_wer": 0.440361057382334,
|
| 489 |
+
"step": 3000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 3.1023784901758016,
|
| 493 |
+
"step": 3000,
|
| 494 |
+
"total_flos": 7.152937185600249e+18,
|
| 495 |
+
"train_loss": 0.6818474756876628,
|
| 496 |
+
"train_runtime": 2982.3235,
|
| 497 |
+
"train_samples_per_second": 38.889,
|
| 498 |
+
"train_steps_per_second": 9.727
|
| 499 |
+
}
|
| 500 |
+
],
|
| 501 |
+
"logging_steps": 100,
|
| 502 |
+
"max_steps": 29010,
|
| 503 |
+
"num_input_tokens_seen": 0,
|
| 504 |
+
"num_train_epochs": 30,
|
| 505 |
+
"save_steps": 400,
|
| 506 |
+
"stateful_callbacks": {
|
| 507 |
+
"EarlyStoppingCallback": {
|
| 508 |
+
"args": {
|
| 509 |
+
"early_stopping_patience": 3,
|
| 510 |
+
"early_stopping_threshold": 0.0
|
| 511 |
+
},
|
| 512 |
+
"attributes": {
|
| 513 |
+
"early_stopping_patience_counter": 1
|
| 514 |
+
}
|
| 515 |
+
},
|
| 516 |
+
"TrainerControl": {
|
| 517 |
+
"args": {
|
| 518 |
+
"should_epoch_stop": false,
|
| 519 |
+
"should_evaluate": false,
|
| 520 |
+
"should_log": false,
|
| 521 |
+
"should_save": true,
|
| 522 |
+
"should_training_stop": false
|
| 523 |
+
},
|
| 524 |
+
"attributes": {}
|
| 525 |
+
}
|
| 526 |
+
},
|
| 527 |
+
"total_flos": 7.152937185600249e+18,
|
| 528 |
+
"train_batch_size": 4,
|
| 529 |
+
"trial_name": null,
|
| 530 |
+
"trial_params": null
|
| 531 |
+
}
|