End of training
Browse files- README.md +4 -1
- adapter.bem.safetensors +3 -0
- all_results.json +15 -0
- eval_results.json +9 -0
- runs/Feb14_15-58-10_srvrocgpu011.uct.ac.za/events.out.tfevents.1739548650.srvrocgpu011.uct.ac.za +3 -0
- train_results.json +9 -0
- trainer_state.json +611 -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-bsbigcgen-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: 0.4207
|
| 22 |
- Wer: 0.4623
|
|
|
|
| 3 |
license: cc-by-nc-4.0
|
| 4 |
base_model: facebook/mms-1b-all
|
| 5 |
tags:
|
| 6 |
+
- automatic-speech-recognition
|
| 7 |
+
- bsbigcgen
|
| 8 |
+
- mms
|
| 9 |
- generated_from_trainer
|
| 10 |
metrics:
|
| 11 |
- wer
|
|
|
|
| 19 |
|
| 20 |
# mms-1b-bsbigcgen-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 BSBIGCGEN - BEM dataset.
|
| 23 |
It achieves the following results on the evaluation set:
|
| 24 |
- Loss: 0.4207
|
| 25 |
- Wer: 0.4623
|
adapter.bem.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:665eb630dc9bcc974d553b41d4fffc8bbe1f50ffc0d257d2b250a449bbd5de67
|
| 3 |
+
size 8798532
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0215208034433285,
|
| 3 |
+
"eval_loss": 0.42067238688468933,
|
| 4 |
+
"eval_runtime": 60.8527,
|
| 5 |
+
"eval_samples": 942,
|
| 6 |
+
"eval_samples_per_second": 15.48,
|
| 7 |
+
"eval_steps_per_second": 3.878,
|
| 8 |
+
"eval_wer": 0.4623473541383989,
|
| 9 |
+
"total_flos": 1.7179112169673994e+19,
|
| 10 |
+
"train_loss": 1.5055460793631417,
|
| 11 |
+
"train_runtime": 6634.6238,
|
| 12 |
+
"train_samples": 5575,
|
| 13 |
+
"train_samples_per_second": 25.209,
|
| 14 |
+
"train_steps_per_second": 3.152
|
| 15 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0215208034433285,
|
| 3 |
+
"eval_loss": 0.42067238688468933,
|
| 4 |
+
"eval_runtime": 60.8527,
|
| 5 |
+
"eval_samples": 942,
|
| 6 |
+
"eval_samples_per_second": 15.48,
|
| 7 |
+
"eval_steps_per_second": 3.878,
|
| 8 |
+
"eval_wer": 0.4623473541383989
|
| 9 |
+
}
|
runs/Feb14_15-58-10_srvrocgpu011.uct.ac.za/events.out.tfevents.1739548650.srvrocgpu011.uct.ac.za
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8fc6bbb5cbd0db0287d406a3dfd54805ddb1331b6d435ef7c828d2c7b26c3011
|
| 3 |
+
size 40
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0215208034433285,
|
| 3 |
+
"total_flos": 1.7179112169673994e+19,
|
| 4 |
+
"train_loss": 1.5055460793631417,
|
| 5 |
+
"train_runtime": 6634.6238,
|
| 6 |
+
"train_samples": 5575,
|
| 7 |
+
"train_samples_per_second": 25.209,
|
| 8 |
+
"train_steps_per_second": 3.152
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,611 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.41538524627685547,
|
| 3 |
+
"best_model_checkpoint": "/scratch/skscla001/speech/results/mms-1b-bsbigcgen-combined-model/checkpoint-3100",
|
| 4 |
+
"epoch": 5.0215208034433285,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 3500,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.14347202295552366,
|
| 13 |
+
"grad_norm": 5.114016056060791,
|
| 14 |
+
"learning_rate": 0.000285,
|
| 15 |
+
"loss": 13.9355,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.14347202295552366,
|
| 20 |
+
"eval_loss": 0.8512570858001709,
|
| 21 |
+
"eval_runtime": 61.5648,
|
| 22 |
+
"eval_samples_per_second": 15.301,
|
| 23 |
+
"eval_steps_per_second": 3.833,
|
| 24 |
+
"eval_wer": 0.7487562189054726,
|
| 25 |
+
"step": 100
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.28694404591104733,
|
| 29 |
+
"grad_norm": 3.307910442352295,
|
| 30 |
+
"learning_rate": 0.0002986304661220567,
|
| 31 |
+
"loss": 1.4992,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.28694404591104733,
|
| 36 |
+
"eval_loss": 0.5381729602813721,
|
| 37 |
+
"eval_runtime": 60.603,
|
| 38 |
+
"eval_samples_per_second": 15.544,
|
| 39 |
+
"eval_steps_per_second": 3.894,
|
| 40 |
+
"eval_wer": 0.5680687471732248,
|
| 41 |
+
"step": 200
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.430416068866571,
|
| 45 |
+
"grad_norm": 3.341427803039551,
|
| 46 |
+
"learning_rate": 0.0002971888515136953,
|
| 47 |
+
"loss": 1.3313,
|
| 48 |
+
"step": 300
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 0.430416068866571,
|
| 52 |
+
"eval_loss": 0.5199733376502991,
|
| 53 |
+
"eval_runtime": 60.4866,
|
| 54 |
+
"eval_samples_per_second": 15.574,
|
| 55 |
+
"eval_steps_per_second": 3.902,
|
| 56 |
+
"eval_wer": 0.5323383084577115,
|
| 57 |
+
"step": 300
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.5738880918220947,
|
| 61 |
+
"grad_norm": 3.2797415256500244,
|
| 62 |
+
"learning_rate": 0.00029574723690533396,
|
| 63 |
+
"loss": 1.2534,
|
| 64 |
+
"step": 400
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 0.5738880918220947,
|
| 68 |
+
"eval_loss": 0.514342725276947,
|
| 69 |
+
"eval_runtime": 60.3537,
|
| 70 |
+
"eval_samples_per_second": 15.608,
|
| 71 |
+
"eval_steps_per_second": 3.91,
|
| 72 |
+
"eval_wer": 0.5271370420624152,
|
| 73 |
+
"step": 400
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.7173601147776184,
|
| 77 |
+
"grad_norm": 2.9253323078155518,
|
| 78 |
+
"learning_rate": 0.0002943056222969726,
|
| 79 |
+
"loss": 1.1283,
|
| 80 |
+
"step": 500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.7173601147776184,
|
| 84 |
+
"eval_loss": 0.5078991055488586,
|
| 85 |
+
"eval_runtime": 60.8274,
|
| 86 |
+
"eval_samples_per_second": 15.486,
|
| 87 |
+
"eval_steps_per_second": 3.88,
|
| 88 |
+
"eval_wer": 0.5200135685210312,
|
| 89 |
+
"step": 500
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.860832137733142,
|
| 93 |
+
"grad_norm": 4.737514019012451,
|
| 94 |
+
"learning_rate": 0.0002928784238346948,
|
| 95 |
+
"loss": 1.3508,
|
| 96 |
+
"step": 600
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 0.860832137733142,
|
| 100 |
+
"eval_loss": 0.5003129243850708,
|
| 101 |
+
"eval_runtime": 60.4081,
|
| 102 |
+
"eval_samples_per_second": 15.594,
|
| 103 |
+
"eval_steps_per_second": 3.907,
|
| 104 |
+
"eval_wer": 0.5192220714608774,
|
| 105 |
+
"step": 600
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.0043041606886658,
|
| 109 |
+
"grad_norm": 16.8955078125,
|
| 110 |
+
"learning_rate": 0.00029143680922633346,
|
| 111 |
+
"loss": 1.2932,
|
| 112 |
+
"step": 700
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.0043041606886658,
|
| 116 |
+
"eval_loss": 0.48886460065841675,
|
| 117 |
+
"eval_runtime": 60.5956,
|
| 118 |
+
"eval_samples_per_second": 15.546,
|
| 119 |
+
"eval_steps_per_second": 3.895,
|
| 120 |
+
"eval_wer": 0.5368611488014473,
|
| 121 |
+
"step": 700
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 1.1477761836441893,
|
| 125 |
+
"grad_norm": 0.937876284122467,
|
| 126 |
+
"learning_rate": 0.0002899951946179721,
|
| 127 |
+
"loss": 1.1563,
|
| 128 |
+
"step": 800
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.1477761836441893,
|
| 132 |
+
"eval_loss": 0.48099467158317566,
|
| 133 |
+
"eval_runtime": 60.2128,
|
| 134 |
+
"eval_samples_per_second": 15.645,
|
| 135 |
+
"eval_steps_per_second": 3.919,
|
| 136 |
+
"eval_wer": 0.5135685210312076,
|
| 137 |
+
"step": 800
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 1.291248206599713,
|
| 141 |
+
"grad_norm": 5.25884485244751,
|
| 142 |
+
"learning_rate": 0.00028855358000961075,
|
| 143 |
+
"loss": 1.3196,
|
| 144 |
+
"step": 900
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 1.291248206599713,
|
| 148 |
+
"eval_loss": 0.47826915979385376,
|
| 149 |
+
"eval_runtime": 60.796,
|
| 150 |
+
"eval_samples_per_second": 15.494,
|
| 151 |
+
"eval_steps_per_second": 3.882,
|
| 152 |
+
"eval_wer": 0.5050881953867028,
|
| 153 |
+
"step": 900
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 1.4347202295552366,
|
| 157 |
+
"grad_norm": 1.4932676553726196,
|
| 158 |
+
"learning_rate": 0.0002871119654012494,
|
| 159 |
+
"loss": 1.2192,
|
| 160 |
+
"step": 1000
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 1.4347202295552366,
|
| 164 |
+
"eval_loss": 0.4746491312980652,
|
| 165 |
+
"eval_runtime": 60.3547,
|
| 166 |
+
"eval_samples_per_second": 15.608,
|
| 167 |
+
"eval_steps_per_second": 3.91,
|
| 168 |
+
"eval_wer": 0.5063319764812302,
|
| 169 |
+
"step": 1000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 1.5781922525107603,
|
| 173 |
+
"grad_norm": 4.697901248931885,
|
| 174 |
+
"learning_rate": 0.00028567035079288803,
|
| 175 |
+
"loss": 1.2438,
|
| 176 |
+
"step": 1100
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 1.5781922525107603,
|
| 180 |
+
"eval_loss": 0.47216328978538513,
|
| 181 |
+
"eval_runtime": 60.77,
|
| 182 |
+
"eval_samples_per_second": 15.501,
|
| 183 |
+
"eval_steps_per_second": 3.883,
|
| 184 |
+
"eval_wer": 0.5063319764812302,
|
| 185 |
+
"step": 1100
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 1.721664275466284,
|
| 189 |
+
"grad_norm": 1.6676461696624756,
|
| 190 |
+
"learning_rate": 0.0002842287361845266,
|
| 191 |
+
"loss": 1.1431,
|
| 192 |
+
"step": 1200
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 1.721664275466284,
|
| 196 |
+
"eval_loss": 0.471213161945343,
|
| 197 |
+
"eval_runtime": 60.9828,
|
| 198 |
+
"eval_samples_per_second": 15.447,
|
| 199 |
+
"eval_steps_per_second": 3.87,
|
| 200 |
+
"eval_wer": 0.5124378109452736,
|
| 201 |
+
"step": 1200
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 1.8651362984218078,
|
| 205 |
+
"grad_norm": 4.146437168121338,
|
| 206 |
+
"learning_rate": 0.00028278712157616526,
|
| 207 |
+
"loss": 1.1551,
|
| 208 |
+
"step": 1300
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 1.8651362984218078,
|
| 212 |
+
"eval_loss": 0.471341609954834,
|
| 213 |
+
"eval_runtime": 61.3886,
|
| 214 |
+
"eval_samples_per_second": 15.345,
|
| 215 |
+
"eval_steps_per_second": 3.844,
|
| 216 |
+
"eval_wer": 0.5021483491632746,
|
| 217 |
+
"step": 1300
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 2.0086083213773316,
|
| 221 |
+
"grad_norm": 7.013607025146484,
|
| 222 |
+
"learning_rate": 0.0002813455069678039,
|
| 223 |
+
"loss": 1.1323,
|
| 224 |
+
"step": 1400
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 2.0086083213773316,
|
| 228 |
+
"eval_loss": 0.46611255407333374,
|
| 229 |
+
"eval_runtime": 60.7353,
|
| 230 |
+
"eval_samples_per_second": 15.51,
|
| 231 |
+
"eval_steps_per_second": 3.886,
|
| 232 |
+
"eval_wer": 0.5163952962460425,
|
| 233 |
+
"step": 1400
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 2.152080344332855,
|
| 237 |
+
"grad_norm": 5.334732532501221,
|
| 238 |
+
"learning_rate": 0.00027990389235944254,
|
| 239 |
+
"loss": 1.102,
|
| 240 |
+
"step": 1500
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 2.152080344332855,
|
| 244 |
+
"eval_loss": 0.45792442560195923,
|
| 245 |
+
"eval_runtime": 60.4707,
|
| 246 |
+
"eval_samples_per_second": 15.578,
|
| 247 |
+
"eval_steps_per_second": 3.903,
|
| 248 |
+
"eval_wer": 0.49095431931252825,
|
| 249 |
+
"step": 1500
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 2.2955523672883786,
|
| 253 |
+
"grad_norm": 12.173432350158691,
|
| 254 |
+
"learning_rate": 0.0002784622777510812,
|
| 255 |
+
"loss": 1.2423,
|
| 256 |
+
"step": 1600
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 2.2955523672883786,
|
| 260 |
+
"eval_loss": 0.46844565868377686,
|
| 261 |
+
"eval_runtime": 61.2187,
|
| 262 |
+
"eval_samples_per_second": 15.387,
|
| 263 |
+
"eval_steps_per_second": 3.855,
|
| 264 |
+
"eval_wer": 0.4955902306648575,
|
| 265 |
+
"step": 1600
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 2.4390243902439024,
|
| 269 |
+
"grad_norm": 14.777493476867676,
|
| 270 |
+
"learning_rate": 0.00027702066314271983,
|
| 271 |
+
"loss": 1.1187,
|
| 272 |
+
"step": 1700
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 2.4390243902439024,
|
| 276 |
+
"eval_loss": 0.447019100189209,
|
| 277 |
+
"eval_runtime": 61.2424,
|
| 278 |
+
"eval_samples_per_second": 15.382,
|
| 279 |
+
"eval_steps_per_second": 3.854,
|
| 280 |
+
"eval_wer": 0.48383084577114427,
|
| 281 |
+
"step": 1700
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 2.582496413199426,
|
| 285 |
+
"grad_norm": 1.6275725364685059,
|
| 286 |
+
"learning_rate": 0.00027557904853435847,
|
| 287 |
+
"loss": 1.1542,
|
| 288 |
+
"step": 1800
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 2.582496413199426,
|
| 292 |
+
"eval_loss": 0.4421312212944031,
|
| 293 |
+
"eval_runtime": 61.0964,
|
| 294 |
+
"eval_samples_per_second": 15.418,
|
| 295 |
+
"eval_steps_per_second": 3.863,
|
| 296 |
+
"eval_wer": 0.47817729534147446,
|
| 297 |
+
"step": 1800
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 2.72596843615495,
|
| 301 |
+
"grad_norm": 3.4510436058044434,
|
| 302 |
+
"learning_rate": 0.0002741374339259971,
|
| 303 |
+
"loss": 1.1252,
|
| 304 |
+
"step": 1900
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 2.72596843615495,
|
| 308 |
+
"eval_loss": 0.43616074323654175,
|
| 309 |
+
"eval_runtime": 61.0506,
|
| 310 |
+
"eval_samples_per_second": 15.43,
|
| 311 |
+
"eval_steps_per_second": 3.866,
|
| 312 |
+
"eval_wer": 0.48484848484848486,
|
| 313 |
+
"step": 1900
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 2.869440459110473,
|
| 317 |
+
"grad_norm": 2.1623735427856445,
|
| 318 |
+
"learning_rate": 0.0002726958193176357,
|
| 319 |
+
"loss": 1.018,
|
| 320 |
+
"step": 2000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 2.869440459110473,
|
| 324 |
+
"eval_loss": 0.4483390152454376,
|
| 325 |
+
"eval_runtime": 60.9286,
|
| 326 |
+
"eval_samples_per_second": 15.461,
|
| 327 |
+
"eval_steps_per_second": 3.873,
|
| 328 |
+
"eval_wer": 0.48100407055630934,
|
| 329 |
+
"step": 2000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 3.012912482065997,
|
| 333 |
+
"grad_norm": 1.0472160577774048,
|
| 334 |
+
"learning_rate": 0.0002712542047092744,
|
| 335 |
+
"loss": 1.1281,
|
| 336 |
+
"step": 2100
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 3.012912482065997,
|
| 340 |
+
"eval_loss": 0.4357198178768158,
|
| 341 |
+
"eval_runtime": 61.3477,
|
| 342 |
+
"eval_samples_per_second": 15.355,
|
| 343 |
+
"eval_steps_per_second": 3.847,
|
| 344 |
+
"eval_wer": 0.47535052012663953,
|
| 345 |
+
"step": 2100
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 3.1563845050215207,
|
| 349 |
+
"grad_norm": 10.624711036682129,
|
| 350 |
+
"learning_rate": 0.000269812590100913,
|
| 351 |
+
"loss": 1.1278,
|
| 352 |
+
"step": 2200
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 3.1563845050215207,
|
| 356 |
+
"eval_loss": 0.4405066668987274,
|
| 357 |
+
"eval_runtime": 60.9486,
|
| 358 |
+
"eval_samples_per_second": 15.456,
|
| 359 |
+
"eval_steps_per_second": 3.872,
|
| 360 |
+
"eval_wer": 0.46698326549072816,
|
| 361 |
+
"step": 2200
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 3.2998565279770444,
|
| 365 |
+
"grad_norm": 2.0468907356262207,
|
| 366 |
+
"learning_rate": 0.0002683709754925516,
|
| 367 |
+
"loss": 1.0072,
|
| 368 |
+
"step": 2300
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 3.2998565279770444,
|
| 372 |
+
"eval_loss": 0.44495889544487,
|
| 373 |
+
"eval_runtime": 61.0226,
|
| 374 |
+
"eval_samples_per_second": 15.437,
|
| 375 |
+
"eval_steps_per_second": 3.867,
|
| 376 |
+
"eval_wer": 0.4703753957485301,
|
| 377 |
+
"step": 2300
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 3.443328550932568,
|
| 381 |
+
"grad_norm": 3.0651752948760986,
|
| 382 |
+
"learning_rate": 0.00026692936088419027,
|
| 383 |
+
"loss": 1.0484,
|
| 384 |
+
"step": 2400
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 3.443328550932568,
|
| 388 |
+
"eval_loss": 0.4355350732803345,
|
| 389 |
+
"eval_runtime": 60.7745,
|
| 390 |
+
"eval_samples_per_second": 15.5,
|
| 391 |
+
"eval_steps_per_second": 3.883,
|
| 392 |
+
"eval_wer": 0.47783808231569425,
|
| 393 |
+
"step": 2400
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 3.586800573888092,
|
| 397 |
+
"grad_norm": 1.657418966293335,
|
| 398 |
+
"learning_rate": 0.0002654877462758289,
|
| 399 |
+
"loss": 1.0515,
|
| 400 |
+
"step": 2500
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 3.586800573888092,
|
| 404 |
+
"eval_loss": 0.42682021856307983,
|
| 405 |
+
"eval_runtime": 61.8866,
|
| 406 |
+
"eval_samples_per_second": 15.221,
|
| 407 |
+
"eval_steps_per_second": 3.813,
|
| 408 |
+
"eval_wer": 0.4796472184531886,
|
| 409 |
+
"step": 2500
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 3.7302725968436157,
|
| 413 |
+
"grad_norm": 2.648669719696045,
|
| 414 |
+
"learning_rate": 0.00026404613166746755,
|
| 415 |
+
"loss": 0.9878,
|
| 416 |
+
"step": 2600
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 3.7302725968436157,
|
| 420 |
+
"eval_loss": 0.43590524792671204,
|
| 421 |
+
"eval_runtime": 60.929,
|
| 422 |
+
"eval_samples_per_second": 15.461,
|
| 423 |
+
"eval_steps_per_second": 3.873,
|
| 424 |
+
"eval_wer": 0.4659656264133876,
|
| 425 |
+
"step": 2600
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 3.8737446197991394,
|
| 429 |
+
"grad_norm": 4.198331356048584,
|
| 430 |
+
"learning_rate": 0.0002626045170591062,
|
| 431 |
+
"loss": 1.1363,
|
| 432 |
+
"step": 2700
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 3.8737446197991394,
|
| 436 |
+
"eval_loss": 0.4254571497440338,
|
| 437 |
+
"eval_runtime": 60.91,
|
| 438 |
+
"eval_samples_per_second": 15.465,
|
| 439 |
+
"eval_steps_per_second": 3.875,
|
| 440 |
+
"eval_wer": 0.48575305291723203,
|
| 441 |
+
"step": 2700
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 4.017216642754663,
|
| 445 |
+
"grad_norm": 1.0629017353057861,
|
| 446 |
+
"learning_rate": 0.0002611629024507448,
|
| 447 |
+
"loss": 1.0978,
|
| 448 |
+
"step": 2800
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 4.017216642754663,
|
| 452 |
+
"eval_loss": 0.4171198308467865,
|
| 453 |
+
"eval_runtime": 60.8914,
|
| 454 |
+
"eval_samples_per_second": 15.47,
|
| 455 |
+
"eval_steps_per_second": 3.876,
|
| 456 |
+
"eval_wer": 0.46483491632745366,
|
| 457 |
+
"step": 2800
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 4.160688665710187,
|
| 461 |
+
"grad_norm": 1.54743492603302,
|
| 462 |
+
"learning_rate": 0.0002597212878423835,
|
| 463 |
+
"loss": 0.9957,
|
| 464 |
+
"step": 2900
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 4.160688665710187,
|
| 468 |
+
"eval_loss": 0.4241373836994171,
|
| 469 |
+
"eval_runtime": 61.4368,
|
| 470 |
+
"eval_samples_per_second": 15.333,
|
| 471 |
+
"eval_steps_per_second": 3.841,
|
| 472 |
+
"eval_wer": 0.47444595205789236,
|
| 473 |
+
"step": 2900
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 4.30416068866571,
|
| 477 |
+
"grad_norm": 1.1327790021896362,
|
| 478 |
+
"learning_rate": 0.00025827967323402206,
|
| 479 |
+
"loss": 1.0,
|
| 480 |
+
"step": 3000
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 4.30416068866571,
|
| 484 |
+
"eval_loss": 0.41563141345977783,
|
| 485 |
+
"eval_runtime": 60.9796,
|
| 486 |
+
"eval_samples_per_second": 15.448,
|
| 487 |
+
"eval_steps_per_second": 3.87,
|
| 488 |
+
"eval_wer": 0.4560153776571687,
|
| 489 |
+
"step": 3000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 4.4476327116212335,
|
| 493 |
+
"grad_norm": 1.5213956832885742,
|
| 494 |
+
"learning_rate": 0.0002568380586256607,
|
| 495 |
+
"loss": 1.0098,
|
| 496 |
+
"step": 3100
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 4.4476327116212335,
|
| 500 |
+
"eval_loss": 0.41538524627685547,
|
| 501 |
+
"eval_runtime": 61.3909,
|
| 502 |
+
"eval_samples_per_second": 15.344,
|
| 503 |
+
"eval_steps_per_second": 3.844,
|
| 504 |
+
"eval_wer": 0.4596336499321574,
|
| 505 |
+
"step": 3100
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 4.591104734576757,
|
| 509 |
+
"grad_norm": 2.431445598602295,
|
| 510 |
+
"learning_rate": 0.00025539644401729935,
|
| 511 |
+
"loss": 1.0682,
|
| 512 |
+
"step": 3200
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 4.591104734576757,
|
| 516 |
+
"eval_loss": 0.41864681243896484,
|
| 517 |
+
"eval_runtime": 60.8692,
|
| 518 |
+
"eval_samples_per_second": 15.476,
|
| 519 |
+
"eval_steps_per_second": 3.877,
|
| 520 |
+
"eval_wer": 0.4708276797829037,
|
| 521 |
+
"step": 3200
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 4.734576757532281,
|
| 525 |
+
"grad_norm": 17.428695678710938,
|
| 526 |
+
"learning_rate": 0.000253954829408938,
|
| 527 |
+
"loss": 1.0633,
|
| 528 |
+
"step": 3300
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 4.734576757532281,
|
| 532 |
+
"eval_loss": 0.4254568815231323,
|
| 533 |
+
"eval_runtime": 61.4388,
|
| 534 |
+
"eval_samples_per_second": 15.332,
|
| 535 |
+
"eval_steps_per_second": 3.841,
|
| 536 |
+
"eval_wer": 0.4635911352329263,
|
| 537 |
+
"step": 3300
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 4.878048780487805,
|
| 541 |
+
"grad_norm": 2.4006314277648926,
|
| 542 |
+
"learning_rate": 0.00025251321480057663,
|
| 543 |
+
"loss": 1.0339,
|
| 544 |
+
"step": 3400
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"epoch": 4.878048780487805,
|
| 548 |
+
"eval_loss": 0.4206084907054901,
|
| 549 |
+
"eval_runtime": 61.2866,
|
| 550 |
+
"eval_samples_per_second": 15.37,
|
| 551 |
+
"eval_steps_per_second": 3.851,
|
| 552 |
+
"eval_wer": 0.45805065581184984,
|
| 553 |
+
"step": 3400
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 5.0215208034433285,
|
| 557 |
+
"grad_norm": 1.0795345306396484,
|
| 558 |
+
"learning_rate": 0.00025107160019221527,
|
| 559 |
+
"loss": 1.0169,
|
| 560 |
+
"step": 3500
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 5.0215208034433285,
|
| 564 |
+
"eval_loss": 0.4206852316856384,
|
| 565 |
+
"eval_runtime": 61.1489,
|
| 566 |
+
"eval_samples_per_second": 15.405,
|
| 567 |
+
"eval_steps_per_second": 3.859,
|
| 568 |
+
"eval_wer": 0.4623473541383989,
|
| 569 |
+
"step": 3500
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 5.0215208034433285,
|
| 573 |
+
"step": 3500,
|
| 574 |
+
"total_flos": 1.7179112169673994e+19,
|
| 575 |
+
"train_loss": 1.5055460793631417,
|
| 576 |
+
"train_runtime": 6634.6238,
|
| 577 |
+
"train_samples_per_second": 25.209,
|
| 578 |
+
"train_steps_per_second": 3.152
|
| 579 |
+
}
|
| 580 |
+
],
|
| 581 |
+
"logging_steps": 100,
|
| 582 |
+
"max_steps": 20910,
|
| 583 |
+
"num_input_tokens_seen": 0,
|
| 584 |
+
"num_train_epochs": 30,
|
| 585 |
+
"save_steps": 400,
|
| 586 |
+
"stateful_callbacks": {
|
| 587 |
+
"EarlyStoppingCallback": {
|
| 588 |
+
"args": {
|
| 589 |
+
"early_stopping_patience": 4,
|
| 590 |
+
"early_stopping_threshold": 0.0
|
| 591 |
+
},
|
| 592 |
+
"attributes": {
|
| 593 |
+
"early_stopping_patience_counter": 1
|
| 594 |
+
}
|
| 595 |
+
},
|
| 596 |
+
"TrainerControl": {
|
| 597 |
+
"args": {
|
| 598 |
+
"should_epoch_stop": false,
|
| 599 |
+
"should_evaluate": false,
|
| 600 |
+
"should_log": false,
|
| 601 |
+
"should_save": true,
|
| 602 |
+
"should_training_stop": false
|
| 603 |
+
},
|
| 604 |
+
"attributes": {}
|
| 605 |
+
}
|
| 606 |
+
},
|
| 607 |
+
"total_flos": 1.7179112169673994e+19,
|
| 608 |
+
"train_batch_size": 4,
|
| 609 |
+
"trial_name": null,
|
| 610 |
+
"trial_params": null
|
| 611 |
+
}
|