End of training
Browse files- README.md +5 -2
- adapter.swa.safetensors +3 -0
- all_results.json +15 -0
- eval_results.json +9 -0
- runs/Jan03_13-43-49_srvrocgpu011.uct.ac.za/events.out.tfevents.1735911107.srvrocgpu011.uct.ac.za +3 -0
- train_results.json +9 -0
- trainer_state.json +579 -0
README.md
CHANGED
|
@@ -3,6 +3,9 @@ library_name: transformers
|
|
| 3 |
license: cc-by-nc-4.0
|
| 4 |
base_model: facebook/mms-1b-all
|
| 5 |
tags:
|
|
|
|
|
|
|
|
|
|
| 6 |
- generated_from_trainer
|
| 7 |
metrics:
|
| 8 |
- wer
|
|
@@ -16,9 +19,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
|
| 17 |
# mms-1b-swagen-combined-30hrs-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: 0.1922
|
| 23 |
|
| 24 |
## Model description
|
|
|
|
| 3 |
license: cc-by-nc-4.0
|
| 4 |
base_model: facebook/mms-1b-all
|
| 5 |
tags:
|
| 6 |
+
- automatic-speech-recognition
|
| 7 |
+
- swagen
|
| 8 |
+
- mms
|
| 9 |
- generated_from_trainer
|
| 10 |
metrics:
|
| 11 |
- wer
|
|
|
|
| 19 |
|
| 20 |
# mms-1b-swagen-combined-30hrs-model
|
| 21 |
|
| 22 |
+
This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on the SWAGEN - SWA dataset.
|
| 23 |
It achieves the following results on the evaluation set:
|
| 24 |
+
- Loss: 0.2278
|
| 25 |
- Wer: 0.1922
|
| 26 |
|
| 27 |
## Model description
|
adapter.swa.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a9a88290061e087b88bc2cf2f7084a7c632050f51a5ddd087704da49df572d51
|
| 3 |
+
size 8906144
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.315852442671984,
|
| 3 |
+
"eval_loss": 0.22776462137699127,
|
| 4 |
+
"eval_runtime": 63.2383,
|
| 5 |
+
"eval_samples": 1132,
|
| 6 |
+
"eval_samples_per_second": 17.901,
|
| 7 |
+
"eval_steps_per_second": 4.475,
|
| 8 |
+
"eval_wer": 0.19219308700834326,
|
| 9 |
+
"total_flos": 1.351677187667784e+19,
|
| 10 |
+
"train_loss": 2.917839215596517,
|
| 11 |
+
"train_runtime": 6068.548,
|
| 12 |
+
"train_samples": 20059,
|
| 13 |
+
"train_samples_per_second": 99.162,
|
| 14 |
+
"train_steps_per_second": 12.393
|
| 15 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.315852442671984,
|
| 3 |
+
"eval_loss": 0.22776462137699127,
|
| 4 |
+
"eval_runtime": 63.2383,
|
| 5 |
+
"eval_samples": 1132,
|
| 6 |
+
"eval_samples_per_second": 17.901,
|
| 7 |
+
"eval_steps_per_second": 4.475,
|
| 8 |
+
"eval_wer": 0.19219308700834326
|
| 9 |
+
}
|
runs/Jan03_13-43-49_srvrocgpu011.uct.ac.za/events.out.tfevents.1735911107.srvrocgpu011.uct.ac.za
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e52390a4c550e144484e2b3b2c54d34657f3f11eb899b652d6dff9691af86577
|
| 3 |
+
size 40
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.315852442671984,
|
| 3 |
+
"total_flos": 1.351677187667784e+19,
|
| 4 |
+
"train_loss": 2.917839215596517,
|
| 5 |
+
"train_runtime": 6068.548,
|
| 6 |
+
"train_samples": 20059,
|
| 7 |
+
"train_samples_per_second": 99.162,
|
| 8 |
+
"train_steps_per_second": 12.393
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,579 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.2257038652896881,
|
| 3 |
+
"best_model_checkpoint": "/scratch/skscla001/speech/results/mms-1b-swagen-combined-30hrs-model/checkpoint-2900",
|
| 4 |
+
"epoch": 1.315852442671984,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 3300,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.03988035892323031,
|
| 13 |
+
"grad_norm": 51.95917510986328,
|
| 14 |
+
"learning_rate": 0.00028799999999999995,
|
| 15 |
+
"loss": 16.9208,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.03988035892323031,
|
| 20 |
+
"eval_loss": 4.268084526062012,
|
| 21 |
+
"eval_runtime": 62.3338,
|
| 22 |
+
"eval_samples_per_second": 18.16,
|
| 23 |
+
"eval_steps_per_second": 4.54,
|
| 24 |
+
"eval_wer": 0.9999006754072308,
|
| 25 |
+
"step": 100
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.07976071784646062,
|
| 29 |
+
"grad_norm": 27.331029891967773,
|
| 30 |
+
"learning_rate": 0.00029962055651710823,
|
| 31 |
+
"loss": 7.2021,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.07976071784646062,
|
| 36 |
+
"eval_loss": 3.4067752361297607,
|
| 37 |
+
"eval_runtime": 62.1258,
|
| 38 |
+
"eval_samples_per_second": 18.221,
|
| 39 |
+
"eval_steps_per_second": 4.555,
|
| 40 |
+
"eval_wer": 1.0026817640047676,
|
| 41 |
+
"step": 200
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.11964107676969092,
|
| 45 |
+
"grad_norm": 32.389076232910156,
|
| 46 |
+
"learning_rate": 0.0002992211423245906,
|
| 47 |
+
"loss": 6.725,
|
| 48 |
+
"step": 300
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 0.11964107676969092,
|
| 52 |
+
"eval_loss": 3.1580252647399902,
|
| 53 |
+
"eval_runtime": 62.3765,
|
| 54 |
+
"eval_samples_per_second": 18.148,
|
| 55 |
+
"eval_steps_per_second": 4.537,
|
| 56 |
+
"eval_wer": 1.014600715137068,
|
| 57 |
+
"step": 300
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.15952143569292124,
|
| 61 |
+
"grad_norm": 34.745182037353516,
|
| 62 |
+
"learning_rate": 0.0002988217281320729,
|
| 63 |
+
"loss": 6.2214,
|
| 64 |
+
"step": 400
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 0.15952143569292124,
|
| 68 |
+
"eval_loss": 3.043919324874878,
|
| 69 |
+
"eval_runtime": 62.1466,
|
| 70 |
+
"eval_samples_per_second": 18.215,
|
| 71 |
+
"eval_steps_per_second": 4.554,
|
| 72 |
+
"eval_wer": 1.0026817640047676,
|
| 73 |
+
"step": 400
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.19940179461615154,
|
| 77 |
+
"grad_norm": 25.307340621948242,
|
| 78 |
+
"learning_rate": 0.0002984223139395553,
|
| 79 |
+
"loss": 5.9996,
|
| 80 |
+
"step": 500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.19940179461615154,
|
| 84 |
+
"eval_loss": 3.004955768585205,
|
| 85 |
+
"eval_runtime": 62.9799,
|
| 86 |
+
"eval_samples_per_second": 17.974,
|
| 87 |
+
"eval_steps_per_second": 4.493,
|
| 88 |
+
"eval_wer": 1.0015891934843066,
|
| 89 |
+
"step": 500
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.23928215353938184,
|
| 93 |
+
"grad_norm": 6.2275800704956055,
|
| 94 |
+
"learning_rate": 0.00029802289974703767,
|
| 95 |
+
"loss": 5.8133,
|
| 96 |
+
"step": 600
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 0.23928215353938184,
|
| 100 |
+
"eval_loss": 2.9414522647857666,
|
| 101 |
+
"eval_runtime": 62.5714,
|
| 102 |
+
"eval_samples_per_second": 18.091,
|
| 103 |
+
"eval_steps_per_second": 4.523,
|
| 104 |
+
"eval_wer": 1.0,
|
| 105 |
+
"step": 600
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.27916251246261214,
|
| 109 |
+
"grad_norm": 6.14309549331665,
|
| 110 |
+
"learning_rate": 0.00029762348555452,
|
| 111 |
+
"loss": 5.805,
|
| 112 |
+
"step": 700
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 0.27916251246261214,
|
| 116 |
+
"eval_loss": 2.8951313495635986,
|
| 117 |
+
"eval_runtime": 62.4303,
|
| 118 |
+
"eval_samples_per_second": 18.132,
|
| 119 |
+
"eval_steps_per_second": 4.533,
|
| 120 |
+
"eval_wer": 0.9997020262216925,
|
| 121 |
+
"step": 700
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.3190428713858425,
|
| 125 |
+
"grad_norm": 14.912216186523438,
|
| 126 |
+
"learning_rate": 0.00029722407136200237,
|
| 127 |
+
"loss": 5.7621,
|
| 128 |
+
"step": 800
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.3190428713858425,
|
| 132 |
+
"eval_loss": 2.905362844467163,
|
| 133 |
+
"eval_runtime": 62.3651,
|
| 134 |
+
"eval_samples_per_second": 18.151,
|
| 135 |
+
"eval_steps_per_second": 4.538,
|
| 136 |
+
"eval_wer": 0.9996027016289233,
|
| 137 |
+
"step": 800
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.3589232303090728,
|
| 141 |
+
"grad_norm": 3.54911470413208,
|
| 142 |
+
"learning_rate": 0.00029682465716948474,
|
| 143 |
+
"loss": 5.6973,
|
| 144 |
+
"step": 900
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.3589232303090728,
|
| 148 |
+
"eval_loss": 2.810933828353882,
|
| 149 |
+
"eval_runtime": 63.0608,
|
| 150 |
+
"eval_samples_per_second": 17.951,
|
| 151 |
+
"eval_steps_per_second": 4.488,
|
| 152 |
+
"eval_wer": 0.9908621374652364,
|
| 153 |
+
"step": 900
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.3988035892323031,
|
| 157 |
+
"grad_norm": 5.402283668518066,
|
| 158 |
+
"learning_rate": 0.00029642524297696706,
|
| 159 |
+
"loss": 5.6541,
|
| 160 |
+
"step": 1000
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 0.3988035892323031,
|
| 164 |
+
"eval_loss": 2.835520029067993,
|
| 165 |
+
"eval_runtime": 62.7046,
|
| 166 |
+
"eval_samples_per_second": 18.053,
|
| 167 |
+
"eval_steps_per_second": 4.513,
|
| 168 |
+
"eval_wer": 0.9910607866507747,
|
| 169 |
+
"step": 1000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 0.4386839481555334,
|
| 173 |
+
"grad_norm": 5.038330554962158,
|
| 174 |
+
"learning_rate": 0.00029602582878444943,
|
| 175 |
+
"loss": 5.4159,
|
| 176 |
+
"step": 1100
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.4386839481555334,
|
| 180 |
+
"eval_loss": 2.7631733417510986,
|
| 181 |
+
"eval_runtime": 62.482,
|
| 182 |
+
"eval_samples_per_second": 18.117,
|
| 183 |
+
"eval_steps_per_second": 4.529,
|
| 184 |
+
"eval_wer": 0.978347238776321,
|
| 185 |
+
"step": 1100
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.4785643070787637,
|
| 189 |
+
"grad_norm": 10.235861778259277,
|
| 190 |
+
"learning_rate": 0.0002956264145919318,
|
| 191 |
+
"loss": 5.4112,
|
| 192 |
+
"step": 1200
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.4785643070787637,
|
| 196 |
+
"eval_loss": 2.6498703956604004,
|
| 197 |
+
"eval_runtime": 62.6482,
|
| 198 |
+
"eval_samples_per_second": 18.069,
|
| 199 |
+
"eval_steps_per_second": 4.517,
|
| 200 |
+
"eval_wer": 0.9793404847040127,
|
| 201 |
+
"step": 1200
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.518444666001994,
|
| 205 |
+
"grad_norm": 15.807747840881348,
|
| 206 |
+
"learning_rate": 0.0002952270003994142,
|
| 207 |
+
"loss": 4.2059,
|
| 208 |
+
"step": 1300
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 0.518444666001994,
|
| 212 |
+
"eval_loss": 0.36752405762672424,
|
| 213 |
+
"eval_runtime": 63.4538,
|
| 214 |
+
"eval_samples_per_second": 17.84,
|
| 215 |
+
"eval_steps_per_second": 4.46,
|
| 216 |
+
"eval_wer": 0.261720301946762,
|
| 217 |
+
"step": 1300
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 0.5583250249252243,
|
| 221 |
+
"grad_norm": 3.0918314456939697,
|
| 222 |
+
"learning_rate": 0.0002948275862068965,
|
| 223 |
+
"loss": 0.6493,
|
| 224 |
+
"step": 1400
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.5583250249252243,
|
| 228 |
+
"eval_loss": 0.27466678619384766,
|
| 229 |
+
"eval_runtime": 62.8469,
|
| 230 |
+
"eval_samples_per_second": 18.012,
|
| 231 |
+
"eval_steps_per_second": 4.503,
|
| 232 |
+
"eval_wer": 0.20768772348033374,
|
| 233 |
+
"step": 1400
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.5982053838484547,
|
| 237 |
+
"grad_norm": 3.152963638305664,
|
| 238 |
+
"learning_rate": 0.0002944281720143789,
|
| 239 |
+
"loss": 0.5624,
|
| 240 |
+
"step": 1500
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.5982053838484547,
|
| 244 |
+
"eval_loss": 0.2648875117301941,
|
| 245 |
+
"eval_runtime": 63.2142,
|
| 246 |
+
"eval_samples_per_second": 17.907,
|
| 247 |
+
"eval_steps_per_second": 4.477,
|
| 248 |
+
"eval_wer": 0.20232419547079858,
|
| 249 |
+
"step": 1500
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.638085742771685,
|
| 253 |
+
"grad_norm": 2.553431749343872,
|
| 254 |
+
"learning_rate": 0.00029402875782186125,
|
| 255 |
+
"loss": 0.5197,
|
| 256 |
+
"step": 1600
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 0.638085742771685,
|
| 260 |
+
"eval_loss": 0.2619471848011017,
|
| 261 |
+
"eval_runtime": 63.2314,
|
| 262 |
+
"eval_samples_per_second": 17.902,
|
| 263 |
+
"eval_steps_per_second": 4.476,
|
| 264 |
+
"eval_wer": 0.19884783472387763,
|
| 265 |
+
"step": 1600
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.6779661016949152,
|
| 269 |
+
"grad_norm": 11.26433277130127,
|
| 270 |
+
"learning_rate": 0.00029363333777126876,
|
| 271 |
+
"loss": 0.4715,
|
| 272 |
+
"step": 1700
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.6779661016949152,
|
| 276 |
+
"eval_loss": 0.2589423656463623,
|
| 277 |
+
"eval_runtime": 63.8104,
|
| 278 |
+
"eval_samples_per_second": 17.74,
|
| 279 |
+
"eval_steps_per_second": 4.435,
|
| 280 |
+
"eval_wer": 0.19815256257449346,
|
| 281 |
+
"step": 1700
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 0.7178464606181456,
|
| 285 |
+
"grad_norm": 2.320037603378296,
|
| 286 |
+
"learning_rate": 0.00029323392357875113,
|
| 287 |
+
"loss": 0.5126,
|
| 288 |
+
"step": 1800
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 0.7178464606181456,
|
| 292 |
+
"eval_loss": 0.2518182694911957,
|
| 293 |
+
"eval_runtime": 63.1237,
|
| 294 |
+
"eval_samples_per_second": 17.933,
|
| 295 |
+
"eval_steps_per_second": 4.483,
|
| 296 |
+
"eval_wer": 0.19785458879618595,
|
| 297 |
+
"step": 1800
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.7577268195413759,
|
| 301 |
+
"grad_norm": 2.7758800983428955,
|
| 302 |
+
"learning_rate": 0.0002928345093862335,
|
| 303 |
+
"loss": 0.4916,
|
| 304 |
+
"step": 1900
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.7577268195413759,
|
| 308 |
+
"eval_loss": 0.254902720451355,
|
| 309 |
+
"eval_runtime": 63.0039,
|
| 310 |
+
"eval_samples_per_second": 17.967,
|
| 311 |
+
"eval_steps_per_second": 4.492,
|
| 312 |
+
"eval_wer": 0.19576877234803336,
|
| 313 |
+
"step": 1900
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 0.7976071784646062,
|
| 317 |
+
"grad_norm": 3.670294761657715,
|
| 318 |
+
"learning_rate": 0.0002924350951937158,
|
| 319 |
+
"loss": 0.4667,
|
| 320 |
+
"step": 2000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.7976071784646062,
|
| 324 |
+
"eval_loss": 0.2501005530357361,
|
| 325 |
+
"eval_runtime": 63.1402,
|
| 326 |
+
"eval_samples_per_second": 17.928,
|
| 327 |
+
"eval_steps_per_second": 4.482,
|
| 328 |
+
"eval_wer": 0.1946762018275725,
|
| 329 |
+
"step": 2000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 0.8374875373878365,
|
| 333 |
+
"grad_norm": 3.5203053951263428,
|
| 334 |
+
"learning_rate": 0.0002920356810011982,
|
| 335 |
+
"loss": 0.4713,
|
| 336 |
+
"step": 2100
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 0.8374875373878365,
|
| 340 |
+
"eval_loss": 0.24789908528327942,
|
| 341 |
+
"eval_runtime": 63.6613,
|
| 342 |
+
"eval_samples_per_second": 17.782,
|
| 343 |
+
"eval_steps_per_second": 4.445,
|
| 344 |
+
"eval_wer": 0.19427890345649582,
|
| 345 |
+
"step": 2100
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.8773678963110668,
|
| 349 |
+
"grad_norm": 3.342277765274048,
|
| 350 |
+
"learning_rate": 0.0002916362668086806,
|
| 351 |
+
"loss": 0.4875,
|
| 352 |
+
"step": 2200
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.8773678963110668,
|
| 356 |
+
"eval_loss": 0.24493974447250366,
|
| 357 |
+
"eval_runtime": 63.0142,
|
| 358 |
+
"eval_samples_per_second": 17.964,
|
| 359 |
+
"eval_steps_per_second": 4.491,
|
| 360 |
+
"eval_wer": 0.19308700834326578,
|
| 361 |
+
"step": 2200
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 0.9172482552342971,
|
| 365 |
+
"grad_norm": 4.0236711502075195,
|
| 366 |
+
"learning_rate": 0.0002912368526161629,
|
| 367 |
+
"loss": 0.4611,
|
| 368 |
+
"step": 2300
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.9172482552342971,
|
| 372 |
+
"eval_loss": 0.24361559748649597,
|
| 373 |
+
"eval_runtime": 62.9448,
|
| 374 |
+
"eval_samples_per_second": 17.984,
|
| 375 |
+
"eval_steps_per_second": 4.496,
|
| 376 |
+
"eval_wer": 0.19348430671434247,
|
| 377 |
+
"step": 2300
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 0.9571286141575274,
|
| 381 |
+
"grad_norm": 2.9615795612335205,
|
| 382 |
+
"learning_rate": 0.00029083743842364527,
|
| 383 |
+
"loss": 0.4587,
|
| 384 |
+
"step": 2400
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 0.9571286141575274,
|
| 388 |
+
"eval_loss": 0.24336393177509308,
|
| 389 |
+
"eval_runtime": 63.2457,
|
| 390 |
+
"eval_samples_per_second": 17.898,
|
| 391 |
+
"eval_steps_per_second": 4.475,
|
| 392 |
+
"eval_wer": 0.19278903456495827,
|
| 393 |
+
"step": 2400
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.9970089730807578,
|
| 397 |
+
"grad_norm": 2.3204922676086426,
|
| 398 |
+
"learning_rate": 0.00029043802423112764,
|
| 399 |
+
"loss": 0.4679,
|
| 400 |
+
"step": 2500
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 0.9970089730807578,
|
| 404 |
+
"eval_loss": 0.24085378646850586,
|
| 405 |
+
"eval_runtime": 63.8549,
|
| 406 |
+
"eval_samples_per_second": 17.728,
|
| 407 |
+
"eval_steps_per_second": 4.432,
|
| 408 |
+
"eval_wer": 0.18951132300357568,
|
| 409 |
+
"step": 2500
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 1.036689930209372,
|
| 413 |
+
"grad_norm": 1.1865150928497314,
|
| 414 |
+
"learning_rate": 0.00029003861003861,
|
| 415 |
+
"loss": 0.4141,
|
| 416 |
+
"step": 2600
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 1.036689930209372,
|
| 420 |
+
"eval_loss": 0.23312124609947205,
|
| 421 |
+
"eval_runtime": 63.2647,
|
| 422 |
+
"eval_samples_per_second": 17.893,
|
| 423 |
+
"eval_steps_per_second": 4.473,
|
| 424 |
+
"eval_wer": 0.18961064759634486,
|
| 425 |
+
"step": 2600
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 1.076570289132602,
|
| 429 |
+
"grad_norm": 1.3028316497802734,
|
| 430 |
+
"learning_rate": 0.0002896391958460924,
|
| 431 |
+
"loss": 0.4263,
|
| 432 |
+
"step": 2700
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 1.076570289132602,
|
| 436 |
+
"eval_loss": 0.23293058574199677,
|
| 437 |
+
"eval_runtime": 62.8596,
|
| 438 |
+
"eval_samples_per_second": 18.008,
|
| 439 |
+
"eval_steps_per_second": 4.502,
|
| 440 |
+
"eval_wer": 0.19199443782280493,
|
| 441 |
+
"step": 2700
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 1.1164506480558325,
|
| 445 |
+
"grad_norm": 7.91491174697876,
|
| 446 |
+
"learning_rate": 0.0002892397816535747,
|
| 447 |
+
"loss": 0.4142,
|
| 448 |
+
"step": 2800
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 1.1164506480558325,
|
| 452 |
+
"eval_loss": 0.23239396512508392,
|
| 453 |
+
"eval_runtime": 63.0701,
|
| 454 |
+
"eval_samples_per_second": 17.948,
|
| 455 |
+
"eval_steps_per_second": 4.487,
|
| 456 |
+
"eval_wer": 0.1917957886372666,
|
| 457 |
+
"step": 2800
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 1.1563310069790629,
|
| 461 |
+
"grad_norm": 1.7904026508331299,
|
| 462 |
+
"learning_rate": 0.0002888403674610571,
|
| 463 |
+
"loss": 0.4606,
|
| 464 |
+
"step": 2900
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 1.1563310069790629,
|
| 468 |
+
"eval_loss": 0.2257038652896881,
|
| 469 |
+
"eval_runtime": 63.835,
|
| 470 |
+
"eval_samples_per_second": 17.733,
|
| 471 |
+
"eval_steps_per_second": 4.433,
|
| 472 |
+
"eval_wer": 0.19427890345649582,
|
| 473 |
+
"step": 2900
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 1.196211365902293,
|
| 477 |
+
"grad_norm": 2.778794527053833,
|
| 478 |
+
"learning_rate": 0.00028844095326853946,
|
| 479 |
+
"loss": 0.4048,
|
| 480 |
+
"step": 3000
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 1.196211365902293,
|
| 484 |
+
"eval_loss": 0.2288675457239151,
|
| 485 |
+
"eval_runtime": 63.4733,
|
| 486 |
+
"eval_samples_per_second": 17.834,
|
| 487 |
+
"eval_steps_per_second": 4.459,
|
| 488 |
+
"eval_wer": 0.19278903456495827,
|
| 489 |
+
"step": 3000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.2360917248255234,
|
| 493 |
+
"grad_norm": 1.2311575412750244,
|
| 494 |
+
"learning_rate": 0.00028804153907602183,
|
| 495 |
+
"loss": 0.4172,
|
| 496 |
+
"step": 3100
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 1.2360917248255234,
|
| 500 |
+
"eval_loss": 0.2326343059539795,
|
| 501 |
+
"eval_runtime": 63.2571,
|
| 502 |
+
"eval_samples_per_second": 17.895,
|
| 503 |
+
"eval_steps_per_second": 4.474,
|
| 504 |
+
"eval_wer": 0.19378228049264998,
|
| 505 |
+
"step": 3100
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 1.2759720837487538,
|
| 509 |
+
"grad_norm": 1.667809247970581,
|
| 510 |
+
"learning_rate": 0.00028764212488350415,
|
| 511 |
+
"loss": 0.4294,
|
| 512 |
+
"step": 3200
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 1.2759720837487538,
|
| 516 |
+
"eval_loss": 0.232680082321167,
|
| 517 |
+
"eval_runtime": 63.7285,
|
| 518 |
+
"eval_samples_per_second": 17.763,
|
| 519 |
+
"eval_steps_per_second": 4.441,
|
| 520 |
+
"eval_wer": 0.1940802542709575,
|
| 521 |
+
"step": 3200
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 1.315852442671984,
|
| 525 |
+
"grad_norm": 2.067225456237793,
|
| 526 |
+
"learning_rate": 0.00028724271069098653,
|
| 527 |
+
"loss": 0.468,
|
| 528 |
+
"step": 3300
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 1.315852442671984,
|
| 532 |
+
"eval_loss": 0.22773675620555878,
|
| 533 |
+
"eval_runtime": 63.9749,
|
| 534 |
+
"eval_samples_per_second": 17.694,
|
| 535 |
+
"eval_steps_per_second": 4.424,
|
| 536 |
+
"eval_wer": 0.19219308700834326,
|
| 537 |
+
"step": 3300
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.315852442671984,
|
| 541 |
+
"step": 3300,
|
| 542 |
+
"total_flos": 1.351677187667784e+19,
|
| 543 |
+
"train_loss": 2.917839215596517,
|
| 544 |
+
"train_runtime": 6068.548,
|
| 545 |
+
"train_samples_per_second": 99.162,
|
| 546 |
+
"train_steps_per_second": 12.393
|
| 547 |
+
}
|
| 548 |
+
],
|
| 549 |
+
"logging_steps": 100,
|
| 550 |
+
"max_steps": 75210,
|
| 551 |
+
"num_input_tokens_seen": 0,
|
| 552 |
+
"num_train_epochs": 30,
|
| 553 |
+
"save_steps": 400,
|
| 554 |
+
"stateful_callbacks": {
|
| 555 |
+
"EarlyStoppingCallback": {
|
| 556 |
+
"args": {
|
| 557 |
+
"early_stopping_patience": 4,
|
| 558 |
+
"early_stopping_threshold": 0.0
|
| 559 |
+
},
|
| 560 |
+
"attributes": {
|
| 561 |
+
"early_stopping_patience_counter": 3
|
| 562 |
+
}
|
| 563 |
+
},
|
| 564 |
+
"TrainerControl": {
|
| 565 |
+
"args": {
|
| 566 |
+
"should_epoch_stop": false,
|
| 567 |
+
"should_evaluate": false,
|
| 568 |
+
"should_log": false,
|
| 569 |
+
"should_save": true,
|
| 570 |
+
"should_training_stop": false
|
| 571 |
+
},
|
| 572 |
+
"attributes": {}
|
| 573 |
+
}
|
| 574 |
+
},
|
| 575 |
+
"total_flos": 1.351677187667784e+19,
|
| 576 |
+
"train_batch_size": 4,
|
| 577 |
+
"trial_name": null,
|
| 578 |
+
"trial_params": null
|
| 579 |
+
}
|