End of training
Browse files- README.md +4 -1
- adapter.loz.safetensors +3 -0
- all_results.json +15 -0
- eval_results.json +9 -0
- train_results.json +9 -0
- trainer_state.json +563 -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
|
|
@@ -16,7 +19,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
|
| 17 |
# mms-1b-lozgen-combined-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: 0.4288
|
| 22 |
- Wer: 0.3297
|
|
|
|
| 3 |
license: cc-by-nc-4.0
|
| 4 |
base_model: facebook/mms-1b-all
|
| 5 |
tags:
|
| 6 |
+
- automatic-speech-recognition
|
| 7 |
+
- lozgen
|
| 8 |
+
- mms
|
| 9 |
- generated_from_trainer
|
| 10 |
metrics:
|
| 11 |
- wer
|
|
|
|
| 19 |
|
| 20 |
# mms-1b-lozgen-combined-model
|
| 21 |
|
| 22 |
+
This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on the LOZGEN - LOZ dataset.
|
| 23 |
It achieves the following results on the evaluation set:
|
| 24 |
- Loss: 0.4288
|
| 25 |
- Wer: 0.3297
|
adapter.loz.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f1c9a94de9ea5c7b30d3d30bd827e3abbae356460e5bea4c6cca87d46529f0fd
|
| 3 |
+
size 8793408
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 13.008130081300813,
|
| 3 |
+
"eval_loss": 0.4287649691104889,
|
| 4 |
+
"eval_runtime": 40.7163,
|
| 5 |
+
"eval_samples": 553,
|
| 6 |
+
"eval_samples_per_second": 13.582,
|
| 7 |
+
"eval_steps_per_second": 3.414,
|
| 8 |
+
"eval_wer": 0.32969406118776245,
|
| 9 |
+
"total_flos": 9.286133731448259e+18,
|
| 10 |
+
"train_loss": 0.8163932430744171,
|
| 11 |
+
"train_runtime": 3897.3821,
|
| 12 |
+
"train_samples": 982,
|
| 13 |
+
"train_samples_per_second": 7.559,
|
| 14 |
+
"train_steps_per_second": 1.894
|
| 15 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 13.008130081300813,
|
| 3 |
+
"eval_loss": 0.4287649691104889,
|
| 4 |
+
"eval_runtime": 40.7163,
|
| 5 |
+
"eval_samples": 553,
|
| 6 |
+
"eval_samples_per_second": 13.582,
|
| 7 |
+
"eval_steps_per_second": 3.414,
|
| 8 |
+
"eval_wer": 0.32969406118776245
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 13.008130081300813,
|
| 3 |
+
"total_flos": 9.286133731448259e+18,
|
| 4 |
+
"train_loss": 0.8163932430744171,
|
| 5 |
+
"train_runtime": 3897.3821,
|
| 6 |
+
"train_samples": 982,
|
| 7 |
+
"train_samples_per_second": 7.559,
|
| 8 |
+
"train_steps_per_second": 1.894
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,563 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.42670491337776184,
|
| 3 |
+
"best_model_checkpoint": "/scratch/skscla001/speech/results/mms-1b-lozgen-combined-model/checkpoint-2900",
|
| 4 |
+
"epoch": 13.008130081300813,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 3200,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.4065040650406504,
|
| 13 |
+
"grad_norm": 5.229663372039795,
|
| 14 |
+
"learning_rate": 0.00029099999999999997,
|
| 15 |
+
"loss": 6.5686,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.4065040650406504,
|
| 20 |
+
"eval_loss": 3.0826728343963623,
|
| 21 |
+
"eval_runtime": 40.6972,
|
| 22 |
+
"eval_samples_per_second": 13.588,
|
| 23 |
+
"eval_steps_per_second": 3.415,
|
| 24 |
+
"eval_wer": 0.970125974805039,
|
| 25 |
+
"step": 100
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.8130081300813008,
|
| 29 |
+
"grad_norm": 2.938697576522827,
|
| 30 |
+
"learning_rate": 0.0002960027472527472,
|
| 31 |
+
"loss": 2.6223,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.8130081300813008,
|
| 36 |
+
"eval_loss": 2.23785662651062,
|
| 37 |
+
"eval_runtime": 40.5049,
|
| 38 |
+
"eval_samples_per_second": 13.653,
|
| 39 |
+
"eval_steps_per_second": 3.432,
|
| 40 |
+
"eval_wer": 0.9112177564487103,
|
| 41 |
+
"step": 200
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 1.2195121951219512,
|
| 45 |
+
"grad_norm": 2.3466386795043945,
|
| 46 |
+
"learning_rate": 0.0002918818681318681,
|
| 47 |
+
"loss": 1.4386,
|
| 48 |
+
"step": 300
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 1.2195121951219512,
|
| 52 |
+
"eval_loss": 0.6909840106964111,
|
| 53 |
+
"eval_runtime": 40.4202,
|
| 54 |
+
"eval_samples_per_second": 13.681,
|
| 55 |
+
"eval_steps_per_second": 3.439,
|
| 56 |
+
"eval_wer": 0.7809238152369526,
|
| 57 |
+
"step": 300
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 1.6260162601626016,
|
| 61 |
+
"grad_norm": 9.536104202270508,
|
| 62 |
+
"learning_rate": 0.000287760989010989,
|
| 63 |
+
"loss": 0.8073,
|
| 64 |
+
"step": 400
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 1.6260162601626016,
|
| 68 |
+
"eval_loss": 0.5903483629226685,
|
| 69 |
+
"eval_runtime": 40.294,
|
| 70 |
+
"eval_samples_per_second": 13.724,
|
| 71 |
+
"eval_steps_per_second": 3.45,
|
| 72 |
+
"eval_wer": 0.5698860227954409,
|
| 73 |
+
"step": 400
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 2.032520325203252,
|
| 77 |
+
"grad_norm": 1.1359059810638428,
|
| 78 |
+
"learning_rate": 0.00028364010989010986,
|
| 79 |
+
"loss": 0.651,
|
| 80 |
+
"step": 500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 2.032520325203252,
|
| 84 |
+
"eval_loss": 0.5555368065834045,
|
| 85 |
+
"eval_runtime": 40.6146,
|
| 86 |
+
"eval_samples_per_second": 13.616,
|
| 87 |
+
"eval_steps_per_second": 3.422,
|
| 88 |
+
"eval_wer": 0.5036592681463707,
|
| 89 |
+
"step": 500
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 2.4390243902439024,
|
| 93 |
+
"grad_norm": 1.2806068658828735,
|
| 94 |
+
"learning_rate": 0.00027951923076923076,
|
| 95 |
+
"loss": 0.655,
|
| 96 |
+
"step": 600
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 2.4390243902439024,
|
| 100 |
+
"eval_loss": 0.529751718044281,
|
| 101 |
+
"eval_runtime": 40.5683,
|
| 102 |
+
"eval_samples_per_second": 13.631,
|
| 103 |
+
"eval_steps_per_second": 3.426,
|
| 104 |
+
"eval_wer": 0.4818236352729454,
|
| 105 |
+
"step": 600
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 2.845528455284553,
|
| 109 |
+
"grad_norm": 0.6596723198890686,
|
| 110 |
+
"learning_rate": 0.0002753983516483516,
|
| 111 |
+
"loss": 0.6579,
|
| 112 |
+
"step": 700
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.845528455284553,
|
| 116 |
+
"eval_loss": 0.5297980308532715,
|
| 117 |
+
"eval_runtime": 40.2452,
|
| 118 |
+
"eval_samples_per_second": 13.741,
|
| 119 |
+
"eval_steps_per_second": 3.454,
|
| 120 |
+
"eval_wer": 0.4603479304139172,
|
| 121 |
+
"step": 700
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 3.252032520325203,
|
| 125 |
+
"grad_norm": 1.53997802734375,
|
| 126 |
+
"learning_rate": 0.0002712774725274725,
|
| 127 |
+
"loss": 0.5699,
|
| 128 |
+
"step": 800
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 3.252032520325203,
|
| 132 |
+
"eval_loss": 0.5159797072410583,
|
| 133 |
+
"eval_runtime": 40.3095,
|
| 134 |
+
"eval_samples_per_second": 13.719,
|
| 135 |
+
"eval_steps_per_second": 3.448,
|
| 136 |
+
"eval_wer": 0.4284343131373725,
|
| 137 |
+
"step": 800
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 3.658536585365854,
|
| 141 |
+
"grad_norm": 5.62721061706543,
|
| 142 |
+
"learning_rate": 0.00026715659340659334,
|
| 143 |
+
"loss": 0.6104,
|
| 144 |
+
"step": 900
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 3.658536585365854,
|
| 148 |
+
"eval_loss": 0.5069619417190552,
|
| 149 |
+
"eval_runtime": 40.7288,
|
| 150 |
+
"eval_samples_per_second": 13.578,
|
| 151 |
+
"eval_steps_per_second": 3.413,
|
| 152 |
+
"eval_wer": 0.4320335932813437,
|
| 153 |
+
"step": 900
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 4.065040650406504,
|
| 157 |
+
"grad_norm": 0.9546043276786804,
|
| 158 |
+
"learning_rate": 0.00026303571428571424,
|
| 159 |
+
"loss": 0.604,
|
| 160 |
+
"step": 1000
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 4.065040650406504,
|
| 164 |
+
"eval_loss": 0.49775609374046326,
|
| 165 |
+
"eval_runtime": 40.5226,
|
| 166 |
+
"eval_samples_per_second": 13.647,
|
| 167 |
+
"eval_steps_per_second": 3.43,
|
| 168 |
+
"eval_wer": 0.4098380323935213,
|
| 169 |
+
"step": 1000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 4.471544715447155,
|
| 173 |
+
"grad_norm": 1.088158130645752,
|
| 174 |
+
"learning_rate": 0.00025895604395604396,
|
| 175 |
+
"loss": 0.5681,
|
| 176 |
+
"step": 1100
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 4.471544715447155,
|
| 180 |
+
"eval_loss": 0.49754026532173157,
|
| 181 |
+
"eval_runtime": 40.4172,
|
| 182 |
+
"eval_samples_per_second": 13.682,
|
| 183 |
+
"eval_steps_per_second": 3.439,
|
| 184 |
+
"eval_wer": 0.40719856028794243,
|
| 185 |
+
"step": 1100
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 4.878048780487805,
|
| 189 |
+
"grad_norm": 1.1959774494171143,
|
| 190 |
+
"learning_rate": 0.0002548351648351648,
|
| 191 |
+
"loss": 0.5493,
|
| 192 |
+
"step": 1200
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 4.878048780487805,
|
| 196 |
+
"eval_loss": 0.4878113865852356,
|
| 197 |
+
"eval_runtime": 40.4064,
|
| 198 |
+
"eval_samples_per_second": 13.686,
|
| 199 |
+
"eval_steps_per_second": 3.44,
|
| 200 |
+
"eval_wer": 0.40383923215356926,
|
| 201 |
+
"step": 1200
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 5.284552845528455,
|
| 205 |
+
"grad_norm": 0.7603219747543335,
|
| 206 |
+
"learning_rate": 0.0002507142857142857,
|
| 207 |
+
"loss": 0.581,
|
| 208 |
+
"step": 1300
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 5.284552845528455,
|
| 212 |
+
"eval_loss": 0.48260873556137085,
|
| 213 |
+
"eval_runtime": 40.7585,
|
| 214 |
+
"eval_samples_per_second": 13.568,
|
| 215 |
+
"eval_steps_per_second": 3.41,
|
| 216 |
+
"eval_wer": 0.39652069586082783,
|
| 217 |
+
"step": 1300
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 5.691056910569106,
|
| 221 |
+
"grad_norm": 1.554537296295166,
|
| 222 |
+
"learning_rate": 0.00024659340659340654,
|
| 223 |
+
"loss": 0.5746,
|
| 224 |
+
"step": 1400
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 5.691056910569106,
|
| 228 |
+
"eval_loss": 0.47932884097099304,
|
| 229 |
+
"eval_runtime": 41.051,
|
| 230 |
+
"eval_samples_per_second": 13.471,
|
| 231 |
+
"eval_steps_per_second": 3.386,
|
| 232 |
+
"eval_wer": 0.4242351529694061,
|
| 233 |
+
"step": 1400
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 6.097560975609756,
|
| 237 |
+
"grad_norm": 1.0625553131103516,
|
| 238 |
+
"learning_rate": 0.00024247252747252747,
|
| 239 |
+
"loss": 0.5238,
|
| 240 |
+
"step": 1500
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 6.097560975609756,
|
| 244 |
+
"eval_loss": 0.472389817237854,
|
| 245 |
+
"eval_runtime": 40.6042,
|
| 246 |
+
"eval_samples_per_second": 13.619,
|
| 247 |
+
"eval_steps_per_second": 3.423,
|
| 248 |
+
"eval_wer": 0.3833233353329334,
|
| 249 |
+
"step": 1500
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 6.504065040650406,
|
| 253 |
+
"grad_norm": 1.2422752380371094,
|
| 254 |
+
"learning_rate": 0.0002383516483516483,
|
| 255 |
+
"loss": 0.5204,
|
| 256 |
+
"step": 1600
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 6.504065040650406,
|
| 260 |
+
"eval_loss": 0.48664653301239014,
|
| 261 |
+
"eval_runtime": 40.5474,
|
| 262 |
+
"eval_samples_per_second": 13.638,
|
| 263 |
+
"eval_steps_per_second": 3.428,
|
| 264 |
+
"eval_wer": 0.3864427114577085,
|
| 265 |
+
"step": 1600
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 6.9105691056910565,
|
| 269 |
+
"grad_norm": 1.324337363243103,
|
| 270 |
+
"learning_rate": 0.000234271978021978,
|
| 271 |
+
"loss": 0.5563,
|
| 272 |
+
"step": 1700
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 6.9105691056910565,
|
| 276 |
+
"eval_loss": 0.46718040108680725,
|
| 277 |
+
"eval_runtime": 40.8357,
|
| 278 |
+
"eval_samples_per_second": 13.542,
|
| 279 |
+
"eval_steps_per_second": 3.404,
|
| 280 |
+
"eval_wer": 0.38392321535692864,
|
| 281 |
+
"step": 1700
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 7.317073170731708,
|
| 285 |
+
"grad_norm": 1.2208648920059204,
|
| 286 |
+
"learning_rate": 0.0002301510989010989,
|
| 287 |
+
"loss": 0.5121,
|
| 288 |
+
"step": 1800
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 7.317073170731708,
|
| 292 |
+
"eval_loss": 0.466389536857605,
|
| 293 |
+
"eval_runtime": 40.8903,
|
| 294 |
+
"eval_samples_per_second": 13.524,
|
| 295 |
+
"eval_steps_per_second": 3.399,
|
| 296 |
+
"eval_wer": 0.3719256148770246,
|
| 297 |
+
"step": 1800
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 7.723577235772358,
|
| 301 |
+
"grad_norm": 1.342207908630371,
|
| 302 |
+
"learning_rate": 0.00022603021978021977,
|
| 303 |
+
"loss": 0.4774,
|
| 304 |
+
"step": 1900
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 7.723577235772358,
|
| 308 |
+
"eval_loss": 0.4625222682952881,
|
| 309 |
+
"eval_runtime": 40.6308,
|
| 310 |
+
"eval_samples_per_second": 13.61,
|
| 311 |
+
"eval_steps_per_second": 3.421,
|
| 312 |
+
"eval_wer": 0.36520695860827834,
|
| 313 |
+
"step": 1900
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 8.130081300813009,
|
| 317 |
+
"grad_norm": 1.1112520694732666,
|
| 318 |
+
"learning_rate": 0.00022190934065934064,
|
| 319 |
+
"loss": 0.5356,
|
| 320 |
+
"step": 2000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 8.130081300813009,
|
| 324 |
+
"eval_loss": 0.4720795750617981,
|
| 325 |
+
"eval_runtime": 40.7009,
|
| 326 |
+
"eval_samples_per_second": 13.587,
|
| 327 |
+
"eval_steps_per_second": 3.415,
|
| 328 |
+
"eval_wer": 0.3692861427714457,
|
| 329 |
+
"step": 2000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 8.536585365853659,
|
| 333 |
+
"grad_norm": 1.3869178295135498,
|
| 334 |
+
"learning_rate": 0.0002177884615384615,
|
| 335 |
+
"loss": 0.4385,
|
| 336 |
+
"step": 2100
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 8.536585365853659,
|
| 340 |
+
"eval_loss": 0.4559965133666992,
|
| 341 |
+
"eval_runtime": 40.8418,
|
| 342 |
+
"eval_samples_per_second": 13.54,
|
| 343 |
+
"eval_steps_per_second": 3.403,
|
| 344 |
+
"eval_wer": 0.3695260947810438,
|
| 345 |
+
"step": 2100
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 8.94308943089431,
|
| 349 |
+
"grad_norm": 1.1107573509216309,
|
| 350 |
+
"learning_rate": 0.0002136675824175824,
|
| 351 |
+
"loss": 0.5561,
|
| 352 |
+
"step": 2200
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 8.94308943089431,
|
| 356 |
+
"eval_loss": 0.4452870488166809,
|
| 357 |
+
"eval_runtime": 40.9093,
|
| 358 |
+
"eval_samples_per_second": 13.518,
|
| 359 |
+
"eval_steps_per_second": 3.398,
|
| 360 |
+
"eval_wer": 0.3594481103779244,
|
| 361 |
+
"step": 2200
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 9.34959349593496,
|
| 365 |
+
"grad_norm": 1.0222277641296387,
|
| 366 |
+
"learning_rate": 0.00020954670329670325,
|
| 367 |
+
"loss": 0.414,
|
| 368 |
+
"step": 2300
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 9.34959349593496,
|
| 372 |
+
"eval_loss": 0.44891127943992615,
|
| 373 |
+
"eval_runtime": 40.5136,
|
| 374 |
+
"eval_samples_per_second": 13.65,
|
| 375 |
+
"eval_steps_per_second": 3.431,
|
| 376 |
+
"eval_wer": 0.3546490701859628,
|
| 377 |
+
"step": 2300
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 9.75609756097561,
|
| 381 |
+
"grad_norm": 1.8858979940414429,
|
| 382 |
+
"learning_rate": 0.00020542582417582415,
|
| 383 |
+
"loss": 0.4763,
|
| 384 |
+
"step": 2400
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 9.75609756097561,
|
| 388 |
+
"eval_loss": 0.4524897038936615,
|
| 389 |
+
"eval_runtime": 40.6457,
|
| 390 |
+
"eval_samples_per_second": 13.605,
|
| 391 |
+
"eval_steps_per_second": 3.42,
|
| 392 |
+
"eval_wer": 0.352129574085183,
|
| 393 |
+
"step": 2400
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 10.16260162601626,
|
| 397 |
+
"grad_norm": 1.759278655052185,
|
| 398 |
+
"learning_rate": 0.00020130494505494502,
|
| 399 |
+
"loss": 0.5317,
|
| 400 |
+
"step": 2500
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 10.16260162601626,
|
| 404 |
+
"eval_loss": 0.44243165850639343,
|
| 405 |
+
"eval_runtime": 40.8637,
|
| 406 |
+
"eval_samples_per_second": 13.533,
|
| 407 |
+
"eval_steps_per_second": 3.402,
|
| 408 |
+
"eval_wer": 0.3557288542291542,
|
| 409 |
+
"step": 2500
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 10.56910569105691,
|
| 413 |
+
"grad_norm": 1.8020973205566406,
|
| 414 |
+
"learning_rate": 0.00019718406593406592,
|
| 415 |
+
"loss": 0.4939,
|
| 416 |
+
"step": 2600
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 10.56910569105691,
|
| 420 |
+
"eval_loss": 0.43977972865104675,
|
| 421 |
+
"eval_runtime": 40.8341,
|
| 422 |
+
"eval_samples_per_second": 13.543,
|
| 423 |
+
"eval_steps_per_second": 3.404,
|
| 424 |
+
"eval_wer": 0.3502099580083983,
|
| 425 |
+
"step": 2600
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 10.975609756097562,
|
| 429 |
+
"grad_norm": 1.453413724899292,
|
| 430 |
+
"learning_rate": 0.00019306318681318682,
|
| 431 |
+
"loss": 0.4456,
|
| 432 |
+
"step": 2700
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 10.975609756097562,
|
| 436 |
+
"eval_loss": 0.4414619505405426,
|
| 437 |
+
"eval_runtime": 40.6722,
|
| 438 |
+
"eval_samples_per_second": 13.597,
|
| 439 |
+
"eval_steps_per_second": 3.418,
|
| 440 |
+
"eval_wer": 0.34673065386922614,
|
| 441 |
+
"step": 2700
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 11.382113821138212,
|
| 445 |
+
"grad_norm": 1.151430606842041,
|
| 446 |
+
"learning_rate": 0.00018898351648351645,
|
| 447 |
+
"loss": 0.4583,
|
| 448 |
+
"step": 2800
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 11.382113821138212,
|
| 452 |
+
"eval_loss": 0.4501504600048065,
|
| 453 |
+
"eval_runtime": 40.6207,
|
| 454 |
+
"eval_samples_per_second": 13.614,
|
| 455 |
+
"eval_steps_per_second": 3.422,
|
| 456 |
+
"eval_wer": 0.34457108578284346,
|
| 457 |
+
"step": 2800
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 11.788617886178862,
|
| 461 |
+
"grad_norm": 4.184691429138184,
|
| 462 |
+
"learning_rate": 0.00018486263736263735,
|
| 463 |
+
"loss": 0.4573,
|
| 464 |
+
"step": 2900
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 11.788617886178862,
|
| 468 |
+
"eval_loss": 0.42670491337776184,
|
| 469 |
+
"eval_runtime": 40.8415,
|
| 470 |
+
"eval_samples_per_second": 13.54,
|
| 471 |
+
"eval_steps_per_second": 3.403,
|
| 472 |
+
"eval_wer": 0.340251949610078,
|
| 473 |
+
"step": 2900
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 12.195121951219512,
|
| 477 |
+
"grad_norm": 16.549129486083984,
|
| 478 |
+
"learning_rate": 0.00018074175824175822,
|
| 479 |
+
"loss": 0.398,
|
| 480 |
+
"step": 3000
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 12.195121951219512,
|
| 484 |
+
"eval_loss": 0.4304564595222473,
|
| 485 |
+
"eval_runtime": 40.8815,
|
| 486 |
+
"eval_samples_per_second": 13.527,
|
| 487 |
+
"eval_steps_per_second": 3.4,
|
| 488 |
+
"eval_wer": 0.3406118776244751,
|
| 489 |
+
"step": 3000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 12.601626016260163,
|
| 493 |
+
"grad_norm": 3.6663782596588135,
|
| 494 |
+
"learning_rate": 0.00017662087912087912,
|
| 495 |
+
"loss": 0.472,
|
| 496 |
+
"step": 3100
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 12.601626016260163,
|
| 500 |
+
"eval_loss": 0.42684391140937805,
|
| 501 |
+
"eval_runtime": 40.8829,
|
| 502 |
+
"eval_samples_per_second": 13.526,
|
| 503 |
+
"eval_steps_per_second": 3.4,
|
| 504 |
+
"eval_wer": 0.3319736052789442,
|
| 505 |
+
"step": 3100
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 13.008130081300813,
|
| 509 |
+
"grad_norm": 0.5571214556694031,
|
| 510 |
+
"learning_rate": 0.00017249999999999996,
|
| 511 |
+
"loss": 0.3993,
|
| 512 |
+
"step": 3200
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 13.008130081300813,
|
| 516 |
+
"eval_loss": 0.42878594994544983,
|
| 517 |
+
"eval_runtime": 40.7789,
|
| 518 |
+
"eval_samples_per_second": 13.561,
|
| 519 |
+
"eval_steps_per_second": 3.409,
|
| 520 |
+
"eval_wer": 0.32969406118776245,
|
| 521 |
+
"step": 3200
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 13.008130081300813,
|
| 525 |
+
"step": 3200,
|
| 526 |
+
"total_flos": 9.286133731448259e+18,
|
| 527 |
+
"train_loss": 0.8163932430744171,
|
| 528 |
+
"train_runtime": 3897.3821,
|
| 529 |
+
"train_samples_per_second": 7.559,
|
| 530 |
+
"train_steps_per_second": 1.894
|
| 531 |
+
}
|
| 532 |
+
],
|
| 533 |
+
"logging_steps": 100,
|
| 534 |
+
"max_steps": 7380,
|
| 535 |
+
"num_input_tokens_seen": 0,
|
| 536 |
+
"num_train_epochs": 30,
|
| 537 |
+
"save_steps": 400,
|
| 538 |
+
"stateful_callbacks": {
|
| 539 |
+
"EarlyStoppingCallback": {
|
| 540 |
+
"args": {
|
| 541 |
+
"early_stopping_patience": 3,
|
| 542 |
+
"early_stopping_threshold": 0.0
|
| 543 |
+
},
|
| 544 |
+
"attributes": {
|
| 545 |
+
"early_stopping_patience_counter": 3
|
| 546 |
+
}
|
| 547 |
+
},
|
| 548 |
+
"TrainerControl": {
|
| 549 |
+
"args": {
|
| 550 |
+
"should_epoch_stop": false,
|
| 551 |
+
"should_evaluate": false,
|
| 552 |
+
"should_log": false,
|
| 553 |
+
"should_save": true,
|
| 554 |
+
"should_training_stop": true
|
| 555 |
+
},
|
| 556 |
+
"attributes": {}
|
| 557 |
+
}
|
| 558 |
+
},
|
| 559 |
+
"total_flos": 9.286133731448259e+18,
|
| 560 |
+
"train_batch_size": 4,
|
| 561 |
+
"trial_name": null,
|
| 562 |
+
"trial_params": null
|
| 563 |
+
}
|