End of training
Browse files- README.md +4 -4
- all_results.json +16 -0
- eval_results.json +10 -0
- train_results.json +9 -0
- trainer_state.json +644 -0
README.md
CHANGED
|
@@ -23,7 +23,7 @@ model-index:
|
|
| 23 |
metrics:
|
| 24 |
- name: Wer
|
| 25 |
type: wer
|
| 26 |
-
value: 0.
|
| 27 |
---
|
| 28 |
|
| 29 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
@@ -33,9 +33,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 33 |
|
| 34 |
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the common_voice_17_0 dataset.
|
| 35 |
It achieves the following results on the evaluation set:
|
| 36 |
-
- Loss: 0.
|
| 37 |
-
- Wer: 0.
|
| 38 |
-
- Cer: 0.
|
| 39 |
|
| 40 |
## Model description
|
| 41 |
|
|
|
|
| 23 |
metrics:
|
| 24 |
- name: Wer
|
| 25 |
type: wer
|
| 26 |
+
value: 0.39399943390885933
|
| 27 |
---
|
| 28 |
|
| 29 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
|
| 33 |
|
| 34 |
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the common_voice_17_0 dataset.
|
| 35 |
It achieves the following results on the evaluation set:
|
| 36 |
+
- Loss: 0.2488
|
| 37 |
+
- Wer: 0.3940
|
| 38 |
+
- Cer: 0.1776
|
| 39 |
|
| 40 |
## Model description
|
| 41 |
|
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 7.0,
|
| 3 |
+
"eval_cer": 0.1776461262464016,
|
| 4 |
+
"eval_loss": 0.24881185591220856,
|
| 5 |
+
"eval_runtime": 16.5577,
|
| 6 |
+
"eval_samples": 500,
|
| 7 |
+
"eval_samples_per_second": 30.197,
|
| 8 |
+
"eval_steps_per_second": 0.966,
|
| 9 |
+
"eval_wer": 0.39399943390885933,
|
| 10 |
+
"total_flos": 5.670215301911177e+19,
|
| 11 |
+
"train_loss": 0.7367133075233617,
|
| 12 |
+
"train_runtime": 12578.5255,
|
| 13 |
+
"train_samples": 35144,
|
| 14 |
+
"train_samples_per_second": 19.558,
|
| 15 |
+
"train_steps_per_second": 0.612
|
| 16 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 7.0,
|
| 3 |
+
"eval_cer": 0.1776461262464016,
|
| 4 |
+
"eval_loss": 0.24881185591220856,
|
| 5 |
+
"eval_runtime": 16.5577,
|
| 6 |
+
"eval_samples": 500,
|
| 7 |
+
"eval_samples_per_second": 30.197,
|
| 8 |
+
"eval_steps_per_second": 0.966,
|
| 9 |
+
"eval_wer": 0.39399943390885933
|
| 10 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 7.0,
|
| 3 |
+
"total_flos": 5.670215301911177e+19,
|
| 4 |
+
"train_loss": 0.7367133075233617,
|
| 5 |
+
"train_runtime": 12578.5255,
|
| 6 |
+
"train_samples": 35144,
|
| 7 |
+
"train_samples_per_second": 19.558,
|
| 8 |
+
"train_steps_per_second": 0.612
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,644 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 7.0,
|
| 5 |
+
"eval_steps": 1000.0,
|
| 6 |
+
"global_step": 7693,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.09099181073703366,
|
| 13 |
+
"grad_norm": 7.5102996826171875,
|
| 14 |
+
"learning_rate": 4.270214798827307e-05,
|
| 15 |
+
"loss": 9.9184,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.18198362147406733,
|
| 20 |
+
"grad_norm": 1.3230904340744019,
|
| 21 |
+
"learning_rate": 4.299387943270356e-05,
|
| 22 |
+
"loss": 3.5315,
|
| 23 |
+
"step": 200
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.272975432211101,
|
| 27 |
+
"grad_norm": 0.9986037015914917,
|
| 28 |
+
"learning_rate": 4.2420092172157666e-05,
|
| 29 |
+
"loss": 3.2339,
|
| 30 |
+
"step": 300
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.36396724294813465,
|
| 34 |
+
"grad_norm": 0.46828845143318176,
|
| 35 |
+
"learning_rate": 4.184630491161178e-05,
|
| 36 |
+
"loss": 3.1838,
|
| 37 |
+
"step": 400
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.4549590536851683,
|
| 41 |
+
"grad_norm": 0.8114163875579834,
|
| 42 |
+
"learning_rate": 4.1272517651065886e-05,
|
| 43 |
+
"loss": 3.1627,
|
| 44 |
+
"step": 500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.545950864422202,
|
| 48 |
+
"grad_norm": 0.4831515848636627,
|
| 49 |
+
"learning_rate": 4.069873039051999e-05,
|
| 50 |
+
"loss": 3.1449,
|
| 51 |
+
"step": 600
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.6369426751592356,
|
| 55 |
+
"grad_norm": 0.8736672401428223,
|
| 56 |
+
"learning_rate": 4.0124943129974106e-05,
|
| 57 |
+
"loss": 2.966,
|
| 58 |
+
"step": 700
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.7279344858962693,
|
| 62 |
+
"grad_norm": 1.4908201694488525,
|
| 63 |
+
"learning_rate": 3.955115586942821e-05,
|
| 64 |
+
"loss": 1.8526,
|
| 65 |
+
"step": 800
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.818926296633303,
|
| 69 |
+
"grad_norm": 1.4355599880218506,
|
| 70 |
+
"learning_rate": 3.8977368608882326e-05,
|
| 71 |
+
"loss": 1.1569,
|
| 72 |
+
"step": 900
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.9099181073703366,
|
| 76 |
+
"grad_norm": 2.112861394882202,
|
| 77 |
+
"learning_rate": 3.840358134833644e-05,
|
| 78 |
+
"loss": 0.9556,
|
| 79 |
+
"step": 1000
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.9099181073703366,
|
| 83 |
+
"eval_cer": 0.297551003379365,
|
| 84 |
+
"eval_loss": 0.7513839602470398,
|
| 85 |
+
"eval_runtime": 16.7143,
|
| 86 |
+
"eval_samples_per_second": 29.914,
|
| 87 |
+
"eval_steps_per_second": 0.957,
|
| 88 |
+
"eval_wer": 0.7659213133314464,
|
| 89 |
+
"step": 1000
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 1.0009099181073704,
|
| 93 |
+
"grad_norm": 1.5618329048156738,
|
| 94 |
+
"learning_rate": 3.7829794087790545e-05,
|
| 95 |
+
"loss": 0.8382,
|
| 96 |
+
"step": 1100
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.091901728844404,
|
| 100 |
+
"grad_norm": 2.229194164276123,
|
| 101 |
+
"learning_rate": 3.725600682724466e-05,
|
| 102 |
+
"loss": 0.7385,
|
| 103 |
+
"step": 1200
|
| 104 |
+
},
|
| 105 |
+
{
|
| 106 |
+
"epoch": 1.1828935395814377,
|
| 107 |
+
"grad_norm": 1.4764809608459473,
|
| 108 |
+
"learning_rate": 3.6682219566698765e-05,
|
| 109 |
+
"loss": 0.6983,
|
| 110 |
+
"step": 1300
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 1.2738853503184713,
|
| 114 |
+
"grad_norm": 3.041120767593384,
|
| 115 |
+
"learning_rate": 3.610843230615287e-05,
|
| 116 |
+
"loss": 0.6448,
|
| 117 |
+
"step": 1400
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 1.364877161055505,
|
| 121 |
+
"grad_norm": 1.5701287984848022,
|
| 122 |
+
"learning_rate": 3.5534645045606985e-05,
|
| 123 |
+
"loss": 0.6075,
|
| 124 |
+
"step": 1500
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 1.4558689717925386,
|
| 128 |
+
"grad_norm": 1.530661702156067,
|
| 129 |
+
"learning_rate": 3.496085778506109e-05,
|
| 130 |
+
"loss": 0.5734,
|
| 131 |
+
"step": 1600
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 1.5468607825295724,
|
| 135 |
+
"grad_norm": 1.5015157461166382,
|
| 136 |
+
"learning_rate": 3.4387070524515204e-05,
|
| 137 |
+
"loss": 0.5511,
|
| 138 |
+
"step": 1700
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 1.6378525932666061,
|
| 142 |
+
"grad_norm": 3.4556000232696533,
|
| 143 |
+
"learning_rate": 3.381328326396932e-05,
|
| 144 |
+
"loss": 0.5314,
|
| 145 |
+
"step": 1800
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 1.7288444040036397,
|
| 149 |
+
"grad_norm": 1.5478144884109497,
|
| 150 |
+
"learning_rate": 3.3239496003423424e-05,
|
| 151 |
+
"loss": 0.521,
|
| 152 |
+
"step": 1900
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 1.8198362147406733,
|
| 156 |
+
"grad_norm": 1.625609040260315,
|
| 157 |
+
"learning_rate": 3.266570874287753e-05,
|
| 158 |
+
"loss": 0.4941,
|
| 159 |
+
"step": 2000
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 1.8198362147406733,
|
| 163 |
+
"eval_cer": 0.21740581584546706,
|
| 164 |
+
"eval_loss": 0.3986930549144745,
|
| 165 |
+
"eval_runtime": 16.3992,
|
| 166 |
+
"eval_samples_per_second": 30.489,
|
| 167 |
+
"eval_steps_per_second": 0.976,
|
| 168 |
+
"eval_wer": 0.5471270874610812,
|
| 169 |
+
"step": 2000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 1.910828025477707,
|
| 173 |
+
"grad_norm": 1.2447913885116577,
|
| 174 |
+
"learning_rate": 3.209192148233164e-05,
|
| 175 |
+
"loss": 0.4767,
|
| 176 |
+
"step": 2100
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 2.001819836214741,
|
| 180 |
+
"grad_norm": 1.7874252796173096,
|
| 181 |
+
"learning_rate": 3.151813422178575e-05,
|
| 182 |
+
"loss": 0.4683,
|
| 183 |
+
"step": 2200
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 2.092811646951774,
|
| 187 |
+
"grad_norm": 2.041367530822754,
|
| 188 |
+
"learning_rate": 3.0944346961239864e-05,
|
| 189 |
+
"loss": 0.4324,
|
| 190 |
+
"step": 2300
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 2.183803457688808,
|
| 194 |
+
"grad_norm": 1.0832343101501465,
|
| 195 |
+
"learning_rate": 3.037055970069397e-05,
|
| 196 |
+
"loss": 0.4308,
|
| 197 |
+
"step": 2400
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 2.2747952684258417,
|
| 201 |
+
"grad_norm": 1.5363576412200928,
|
| 202 |
+
"learning_rate": 2.9796772440148083e-05,
|
| 203 |
+
"loss": 0.4154,
|
| 204 |
+
"step": 2500
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 2.3657870791628755,
|
| 208 |
+
"grad_norm": 2.582390308380127,
|
| 209 |
+
"learning_rate": 2.9222985179602193e-05,
|
| 210 |
+
"loss": 0.4049,
|
| 211 |
+
"step": 2600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 2.4567788898999092,
|
| 215 |
+
"grad_norm": 1.7013890743255615,
|
| 216 |
+
"learning_rate": 2.86491979190563e-05,
|
| 217 |
+
"loss": 0.3905,
|
| 218 |
+
"step": 2700
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 2.5477707006369426,
|
| 222 |
+
"grad_norm": 1.6584831476211548,
|
| 223 |
+
"learning_rate": 2.8075410658510413e-05,
|
| 224 |
+
"loss": 0.389,
|
| 225 |
+
"step": 2800
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 2.6387625113739763,
|
| 229 |
+
"grad_norm": 2.1508281230926514,
|
| 230 |
+
"learning_rate": 2.750162339796452e-05,
|
| 231 |
+
"loss": 0.3679,
|
| 232 |
+
"step": 2900
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 2.72975432211101,
|
| 236 |
+
"grad_norm": 1.0315163135528564,
|
| 237 |
+
"learning_rate": 2.692783613741863e-05,
|
| 238 |
+
"loss": 0.3694,
|
| 239 |
+
"step": 3000
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 2.72975432211101,
|
| 243 |
+
"eval_cer": 0.20046727022403937,
|
| 244 |
+
"eval_loss": 0.3281673192977905,
|
| 245 |
+
"eval_runtime": 16.5376,
|
| 246 |
+
"eval_samples_per_second": 30.234,
|
| 247 |
+
"eval_steps_per_second": 0.967,
|
| 248 |
+
"eval_wer": 0.4874044721200113,
|
| 249 |
+
"step": 3000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 2.8207461328480434,
|
| 253 |
+
"grad_norm": 1.1783397197723389,
|
| 254 |
+
"learning_rate": 2.6354048876872742e-05,
|
| 255 |
+
"loss": 0.3776,
|
| 256 |
+
"step": 3100
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 2.911737943585077,
|
| 260 |
+
"grad_norm": 1.5299417972564697,
|
| 261 |
+
"learning_rate": 2.578026161632685e-05,
|
| 262 |
+
"loss": 0.3671,
|
| 263 |
+
"step": 3200
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 3.002729754322111,
|
| 267 |
+
"grad_norm": 1.179459810256958,
|
| 268 |
+
"learning_rate": 2.520647435578096e-05,
|
| 269 |
+
"loss": 0.3588,
|
| 270 |
+
"step": 3300
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 3.0937215650591448,
|
| 274 |
+
"grad_norm": 2.9286251068115234,
|
| 275 |
+
"learning_rate": 2.463268709523507e-05,
|
| 276 |
+
"loss": 0.3383,
|
| 277 |
+
"step": 3400
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 3.1847133757961785,
|
| 281 |
+
"grad_norm": 4.09226131439209,
|
| 282 |
+
"learning_rate": 2.405889983468918e-05,
|
| 283 |
+
"loss": 0.3374,
|
| 284 |
+
"step": 3500
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 3.275705186533212,
|
| 288 |
+
"grad_norm": 1.4895093441009521,
|
| 289 |
+
"learning_rate": 2.348511257414329e-05,
|
| 290 |
+
"loss": 0.3231,
|
| 291 |
+
"step": 3600
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 3.3666969972702456,
|
| 295 |
+
"grad_norm": 1.2932852506637573,
|
| 296 |
+
"learning_rate": 2.2911325313597398e-05,
|
| 297 |
+
"loss": 0.3286,
|
| 298 |
+
"step": 3700
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 3.4576888080072794,
|
| 302 |
+
"grad_norm": 1.4036494493484497,
|
| 303 |
+
"learning_rate": 2.2337538053051508e-05,
|
| 304 |
+
"loss": 0.3326,
|
| 305 |
+
"step": 3800
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 3.548680618744313,
|
| 309 |
+
"grad_norm": 2.6920440196990967,
|
| 310 |
+
"learning_rate": 2.1763750792505618e-05,
|
| 311 |
+
"loss": 0.3291,
|
| 312 |
+
"step": 3900
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 3.6396724294813465,
|
| 316 |
+
"grad_norm": 2.416705369949341,
|
| 317 |
+
"learning_rate": 2.1189963531959728e-05,
|
| 318 |
+
"loss": 0.3199,
|
| 319 |
+
"step": 4000
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 3.6396724294813465,
|
| 323 |
+
"eval_cer": 0.19007885185030665,
|
| 324 |
+
"eval_loss": 0.28464144468307495,
|
| 325 |
+
"eval_runtime": 16.6491,
|
| 326 |
+
"eval_samples_per_second": 30.032,
|
| 327 |
+
"eval_steps_per_second": 0.961,
|
| 328 |
+
"eval_wer": 0.4506085479762242,
|
| 329 |
+
"step": 4000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 3.7306642402183803,
|
| 333 |
+
"grad_norm": 2.419663906097412,
|
| 334 |
+
"learning_rate": 2.0616176271413838e-05,
|
| 335 |
+
"loss": 0.3038,
|
| 336 |
+
"step": 4100
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 3.821656050955414,
|
| 340 |
+
"grad_norm": 1.5200859308242798,
|
| 341 |
+
"learning_rate": 2.0042389010867947e-05,
|
| 342 |
+
"loss": 0.3002,
|
| 343 |
+
"step": 4200
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"epoch": 3.912647861692448,
|
| 347 |
+
"grad_norm": 1.2179443836212158,
|
| 348 |
+
"learning_rate": 1.9468601750322057e-05,
|
| 349 |
+
"loss": 0.302,
|
| 350 |
+
"step": 4300
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 4.003639672429482,
|
| 354 |
+
"grad_norm": 1.685207724571228,
|
| 355 |
+
"learning_rate": 1.8894814489776167e-05,
|
| 356 |
+
"loss": 0.2969,
|
| 357 |
+
"step": 4400
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 4.094631483166515,
|
| 361 |
+
"grad_norm": 1.147800326347351,
|
| 362 |
+
"learning_rate": 1.8321027229230277e-05,
|
| 363 |
+
"loss": 0.2882,
|
| 364 |
+
"step": 4500
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 4.185623293903548,
|
| 368 |
+
"grad_norm": 2.0389134883880615,
|
| 369 |
+
"learning_rate": 1.7747239968684387e-05,
|
| 370 |
+
"loss": 0.2833,
|
| 371 |
+
"step": 4600
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 4.276615104640582,
|
| 375 |
+
"grad_norm": 2.9132614135742188,
|
| 376 |
+
"learning_rate": 1.7173452708138497e-05,
|
| 377 |
+
"loss": 0.2912,
|
| 378 |
+
"step": 4700
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 4.367606915377616,
|
| 382 |
+
"grad_norm": 1.1282093524932861,
|
| 383 |
+
"learning_rate": 1.6599665447592606e-05,
|
| 384 |
+
"loss": 0.2806,
|
| 385 |
+
"step": 4800
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 4.45859872611465,
|
| 389 |
+
"grad_norm": 0.9727948904037476,
|
| 390 |
+
"learning_rate": 1.6025878187046713e-05,
|
| 391 |
+
"loss": 0.2883,
|
| 392 |
+
"step": 4900
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 4.549590536851683,
|
| 396 |
+
"grad_norm": 1.485788106918335,
|
| 397 |
+
"learning_rate": 1.5452090926500826e-05,
|
| 398 |
+
"loss": 0.2805,
|
| 399 |
+
"step": 5000
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 4.549590536851683,
|
| 403 |
+
"eval_cer": 0.1855313112770662,
|
| 404 |
+
"eval_loss": 0.2716318368911743,
|
| 405 |
+
"eval_runtime": 16.6008,
|
| 406 |
+
"eval_samples_per_second": 30.119,
|
| 407 |
+
"eval_steps_per_second": 0.964,
|
| 408 |
+
"eval_wer": 0.4254174922162468,
|
| 409 |
+
"step": 5000
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 4.640582347588717,
|
| 413 |
+
"grad_norm": 11.880212783813477,
|
| 414 |
+
"learning_rate": 1.4878303665954934e-05,
|
| 415 |
+
"loss": 0.2827,
|
| 416 |
+
"step": 5100
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 4.731574158325751,
|
| 420 |
+
"grad_norm": 3.0947282314300537,
|
| 421 |
+
"learning_rate": 1.4304516405409044e-05,
|
| 422 |
+
"loss": 0.2759,
|
| 423 |
+
"step": 5200
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 4.822565969062785,
|
| 427 |
+
"grad_norm": 1.0977256298065186,
|
| 428 |
+
"learning_rate": 1.3730729144863152e-05,
|
| 429 |
+
"loss": 0.2785,
|
| 430 |
+
"step": 5300
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 4.9135577797998184,
|
| 434 |
+
"grad_norm": 1.035343050956726,
|
| 435 |
+
"learning_rate": 1.3156941884317266e-05,
|
| 436 |
+
"loss": 0.2713,
|
| 437 |
+
"step": 5400
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 5.004549590536851,
|
| 441 |
+
"grad_norm": 1.091783881187439,
|
| 442 |
+
"learning_rate": 1.2583154623771374e-05,
|
| 443 |
+
"loss": 0.2768,
|
| 444 |
+
"step": 5500
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 5.095541401273885,
|
| 448 |
+
"grad_norm": 1.1632474660873413,
|
| 449 |
+
"learning_rate": 1.2009367363225484e-05,
|
| 450 |
+
"loss": 0.2633,
|
| 451 |
+
"step": 5600
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 5.186533212010919,
|
| 455 |
+
"grad_norm": 1.4272072315216064,
|
| 456 |
+
"learning_rate": 1.1435580102679592e-05,
|
| 457 |
+
"loss": 0.2664,
|
| 458 |
+
"step": 5700
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 5.277525022747953,
|
| 462 |
+
"grad_norm": 1.1368160247802734,
|
| 463 |
+
"learning_rate": 1.0861792842133703e-05,
|
| 464 |
+
"loss": 0.2516,
|
| 465 |
+
"step": 5800
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 5.368516833484986,
|
| 469 |
+
"grad_norm": 10.24752426147461,
|
| 470 |
+
"learning_rate": 1.0288005581587813e-05,
|
| 471 |
+
"loss": 0.2588,
|
| 472 |
+
"step": 5900
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 5.45950864422202,
|
| 476 |
+
"grad_norm": 1.692718505859375,
|
| 477 |
+
"learning_rate": 9.714218321041923e-06,
|
| 478 |
+
"loss": 0.2572,
|
| 479 |
+
"step": 6000
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 5.45950864422202,
|
| 483 |
+
"eval_cer": 0.18102549125954356,
|
| 484 |
+
"eval_loss": 0.26223766803741455,
|
| 485 |
+
"eval_runtime": 16.6819,
|
| 486 |
+
"eval_samples_per_second": 29.973,
|
| 487 |
+
"eval_steps_per_second": 0.959,
|
| 488 |
+
"eval_wer": 0.40843475799603735,
|
| 489 |
+
"step": 6000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 5.550500454959054,
|
| 493 |
+
"grad_norm": 2.930900812149048,
|
| 494 |
+
"learning_rate": 9.140431060496031e-06,
|
| 495 |
+
"loss": 0.2559,
|
| 496 |
+
"step": 6100
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 5.641492265696087,
|
| 500 |
+
"grad_norm": 1.0715835094451904,
|
| 501 |
+
"learning_rate": 8.566643799950141e-06,
|
| 502 |
+
"loss": 0.2593,
|
| 503 |
+
"step": 6200
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 5.732484076433121,
|
| 507 |
+
"grad_norm": 1.025639533996582,
|
| 508 |
+
"learning_rate": 7.992856539404251e-06,
|
| 509 |
+
"loss": 0.2531,
|
| 510 |
+
"step": 6300
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 5.823475887170154,
|
| 514 |
+
"grad_norm": 1.0805244445800781,
|
| 515 |
+
"learning_rate": 7.4190692788583616e-06,
|
| 516 |
+
"loss": 0.2541,
|
| 517 |
+
"step": 6400
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 5.914467697907188,
|
| 521 |
+
"grad_norm": 1.0423027276992798,
|
| 522 |
+
"learning_rate": 6.845282018312471e-06,
|
| 523 |
+
"loss": 0.2523,
|
| 524 |
+
"step": 6500
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 6.005459508644222,
|
| 528 |
+
"grad_norm": 0.9967121481895447,
|
| 529 |
+
"learning_rate": 6.271494757766581e-06,
|
| 530 |
+
"loss": 0.2494,
|
| 531 |
+
"step": 6600
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 6.096451319381256,
|
| 535 |
+
"grad_norm": 1.4251012802124023,
|
| 536 |
+
"learning_rate": 5.69770749722069e-06,
|
| 537 |
+
"loss": 0.2398,
|
| 538 |
+
"step": 6700
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 6.1874431301182895,
|
| 542 |
+
"grad_norm": 1.3021987676620483,
|
| 543 |
+
"learning_rate": 5.1239202366748e-06,
|
| 544 |
+
"loss": 0.2451,
|
| 545 |
+
"step": 6800
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 6.278434940855323,
|
| 549 |
+
"grad_norm": 0.9879375100135803,
|
| 550 |
+
"learning_rate": 4.55013297612891e-06,
|
| 551 |
+
"loss": 0.2467,
|
| 552 |
+
"step": 6900
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 6.369426751592357,
|
| 556 |
+
"grad_norm": 2.257263660430908,
|
| 557 |
+
"learning_rate": 3.976345715583019e-06,
|
| 558 |
+
"loss": 0.2389,
|
| 559 |
+
"step": 7000
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 6.369426751592357,
|
| 563 |
+
"eval_cer": 0.17910634569652467,
|
| 564 |
+
"eval_loss": 0.24834765493869781,
|
| 565 |
+
"eval_runtime": 16.597,
|
| 566 |
+
"eval_samples_per_second": 30.126,
|
| 567 |
+
"eval_steps_per_second": 0.964,
|
| 568 |
+
"eval_wer": 0.4002264364562695,
|
| 569 |
+
"step": 7000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 6.460418562329391,
|
| 573 |
+
"grad_norm": 1.686767578125,
|
| 574 |
+
"learning_rate": 3.4025584550371293e-06,
|
| 575 |
+
"loss": 0.2532,
|
| 576 |
+
"step": 7100
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 6.551410373066424,
|
| 580 |
+
"grad_norm": 1.9620840549468994,
|
| 581 |
+
"learning_rate": 2.8287711944912387e-06,
|
| 582 |
+
"loss": 0.2427,
|
| 583 |
+
"step": 7200
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 6.6424021838034575,
|
| 587 |
+
"grad_norm": 1.0499871969223022,
|
| 588 |
+
"learning_rate": 2.2549839339453486e-06,
|
| 589 |
+
"loss": 0.2449,
|
| 590 |
+
"step": 7300
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 6.733393994540491,
|
| 594 |
+
"grad_norm": 2.9436795711517334,
|
| 595 |
+
"learning_rate": 1.6811966733994584e-06,
|
| 596 |
+
"loss": 0.2428,
|
| 597 |
+
"step": 7400
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 6.824385805277525,
|
| 601 |
+
"grad_norm": 1.0118967294692993,
|
| 602 |
+
"learning_rate": 1.1074094128535683e-06,
|
| 603 |
+
"loss": 0.2341,
|
| 604 |
+
"step": 7500
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 6.915377616014559,
|
| 608 |
+
"grad_norm": 1.00021493434906,
|
| 609 |
+
"learning_rate": 5.336221523076779e-07,
|
| 610 |
+
"loss": 0.2468,
|
| 611 |
+
"step": 7600
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 7.0,
|
| 615 |
+
"step": 7693,
|
| 616 |
+
"total_flos": 5.670215301911177e+19,
|
| 617 |
+
"train_loss": 0.7367133075233617,
|
| 618 |
+
"train_runtime": 12578.5255,
|
| 619 |
+
"train_samples_per_second": 19.558,
|
| 620 |
+
"train_steps_per_second": 0.612
|
| 621 |
+
}
|
| 622 |
+
],
|
| 623 |
+
"logging_steps": 100,
|
| 624 |
+
"max_steps": 7693,
|
| 625 |
+
"num_input_tokens_seen": 0,
|
| 626 |
+
"num_train_epochs": 7,
|
| 627 |
+
"save_steps": 1000,
|
| 628 |
+
"stateful_callbacks": {
|
| 629 |
+
"TrainerControl": {
|
| 630 |
+
"args": {
|
| 631 |
+
"should_epoch_stop": false,
|
| 632 |
+
"should_evaluate": false,
|
| 633 |
+
"should_log": false,
|
| 634 |
+
"should_save": true,
|
| 635 |
+
"should_training_stop": true
|
| 636 |
+
},
|
| 637 |
+
"attributes": {}
|
| 638 |
+
}
|
| 639 |
+
},
|
| 640 |
+
"total_flos": 5.670215301911177e+19,
|
| 641 |
+
"train_batch_size": 16,
|
| 642 |
+
"trial_name": null,
|
| 643 |
+
"trial_params": null
|
| 644 |
+
}
|