Hugging Face's logo

aconeil
/
w2v2-lmk_updated

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
wav2vec2
Generated from Trainer
Eval Results (legacy)
Model card Files
xet
Metrics Community
aconeil commited on
Commit
b6836bc
·
verified ·
1 Parent(s): 833b20f

End of training

Browse files
Files changed (3) hide show
  1. README.md +24 -29
  2. model.safetensors +1 -1
  3. runs/Nov19_19-56-32_x1001c1s7b0n1/events.out.tfevents.1763600200.x1001c1s7b0n1.229000.0 +2 -2
README.md CHANGED
@@ -23,7 +23,7 @@ model-index:
23
  metrics:
24
  - name: Wer
25
  type: wer
26
- value: 0.5888501742160279
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-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the audiofolder dataset.
35
  It achieves the following results on the evaluation set:
36
- - Loss: 1.1431
37
- - Wer: 0.5889
38
- - Cer: 0.2034
39
 
40
  ## Model description
41
 
@@ -55,11 +55,11 @@ More information needed
55
 
56
  The following hyperparameters were used during training:
57
  - learning_rate: 0.0001
58
- - train_batch_size: 6
59
  - eval_batch_size: 8
60
  - seed: 42
61
  - gradient_accumulation_steps: 2
62
- - total_train_batch_size: 12
63
  - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
64
  - lr_scheduler_type: linear
65
  - lr_scheduler_warmup_steps: 300
@@ -68,29 +68,24 @@ The following hyperparameters were used during training:
68
 
69
  ### Training results
70
 
71
- | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
72
- |:-------------:|:-------:|:----:|:---------------:|:------:|:------:|
73
- | 10.3044 | 4.7619 | 100 | 5.0138 | 1.0 | 1.0 |
74
- | 3.1772 | 9.5238 | 200 | 2.9455 | 1.0 | 1.0 |
75
- | 2.9716 | 14.2857 | 300 | 2.8562 | 1.0 | 1.0 |
76
- | 2.8851 | 19.0476 | 400 | 2.6921 | 1.0 | 1.0 |
77
- | 2.441 | 23.8095 | 500 | 1.9387 | 0.9861 | 0.6992 |
78
- | 1.7135 | 28.5714 | 600 | 1.3350 | 0.9094 | 0.3854 |
79
- | 1.2811 | 33.3333 | 700 | 1.1663 | 0.7735 | 0.2955 |
80
- | 1.0414 | 38.0952 | 800 | 1.0695 | 0.7038 | 0.2513 |
81
- | 0.861 | 42.8571 | 900 | 1.0461 | 0.6655 | 0.2437 |
82
- | 0.7126 | 47.6190 | 1000 | 1.0540 | 0.6341 | 0.2399 |
83
- | 0.6523 | 52.3810 | 1100 | 1.0586 | 0.6098 | 0.2201 |
84
- | 0.5593 | 57.1429 | 1200 | 1.0513 | 0.6132 | 0.2193 |
85
- | 0.5007 | 61.9048 | 1300 | 1.0781 | 0.5784 | 0.2117 |
86
- | 0.4457 | 66.6667 | 1400 | 1.1428 | 0.6028 | 0.2171 |
87
- | 0.427 | 71.4286 | 1500 | 1.1510 | 0.5993 | 0.2125 |
88
- | 0.3868 | 76.1905 | 1600 | 1.1391 | 0.6098 | 0.2102 |
89
- | 0.3483 | 80.9524 | 1700 | 1.1349 | 0.5854 | 0.2011 |
90
- | 0.3526 | 85.7143 | 1800 | 1.1036 | 0.5819 | 0.1995 |
91
- | 0.3294 | 90.4762 | 1900 | 1.1188 | 0.5679 | 0.1957 |
92
- | 0.3208 | 95.2381 | 2000 | 1.1397 | 0.5819 | 0.2026 |
93
- | 0.2969 | 100.0 | 2100 | 1.1431 | 0.5889 | 0.2034 |
94
 
95
 
96
  ### Framework versions
 
23
  metrics:
24
  - name: Wer
25
  type: wer
26
+ value: 0.7700348432055749
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-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the audiofolder dataset.
35
  It achieves the following results on the evaluation set:
36
+ - Loss: 2.2629
37
+ - Wer: 0.7700
38
+ - Cer: 0.3244
39
 
40
  ## Model description
41
 
 
55
 
56
  The following hyperparameters were used during training:
57
  - learning_rate: 0.0001
58
+ - train_batch_size: 8
59
  - eval_batch_size: 8
60
  - seed: 42
61
  - gradient_accumulation_steps: 2
62
+ - total_train_batch_size: 16
63
  - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
64
  - lr_scheduler_type: linear
65
  - lr_scheduler_warmup_steps: 300
 
68
 
69
  ### Training results
70
 
71
+ | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
72
+ |:-------------:|:-----:|:----:|:---------------:|:------:|:------:|
73
+ | 5.732 | 6.25 | 100 | 3.3371 | 1.0 | 1.0 |
74
+ | 3.0052 | 12.5 | 200 | 2.8812 | 1.0 | 1.0 |
75
+ | 2.6434 | 18.75 | 300 | 2.3653 | 1.0 | 0.8850 |
76
+ | 0.8393 | 25.0 | 400 | 1.5602 | 0.7770 | 0.3488 |
77
+ | 0.2892 | 31.25 | 500 | 1.6106 | 0.7770 | 0.3298 |
78
+ | 0.1167 | 37.5 | 600 | 1.7649 | 0.7909 | 0.3267 |
79
+ | 0.0595 | 43.75 | 700 | 1.8324 | 0.7666 | 0.3138 |
80
+ | 0.0337 | 50.0 | 800 | 2.0307 | 0.7875 | 0.3351 |
81
+ | 0.0222 | 56.25 | 900 | 2.0604 | 0.7840 | 0.3305 |
82
+ | 0.015 | 62.5 | 1000 | 2.1389 | 0.7735 | 0.3313 |
83
+ | 0.0127 | 68.75 | 1100 | 2.1756 | 0.7700 | 0.3260 |
84
+ | 0.0109 | 75.0 | 1200 | 2.2084 | 0.7805 | 0.3283 |
85
+ | 0.0097 | 81.25 | 1300 | 2.2374 | 0.7805 | 0.3267 |
86
+ | 0.0088 | 87.5 | 1400 | 2.2508 | 0.7700 | 0.3252 |
87
+ | 0.008 | 93.75 | 1500 | 2.2586 | 0.7735 | 0.3252 |
88
+ | 0.0077 | 100.0 | 1600 | 2.2629 | 0.7700 | 0.3244 |
 
 
 
 
 
89
 
90
 
91
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:34a3d41922062c6d694e843bd1e945410231566c14d2a7ce870508295e1cc1fb
3
  size 1261934488
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e2925ec900fd4eb90bcd51ef2daf498431df0ca4d851c12199fcd657acb2422b
3
  size 1261934488
runs/Nov19_19-56-32_x1001c1s7b0n1/events.out.tfevents.1763600200.x1001c1s7b0n1.229000.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:44638dac2b4192e67c0b5261aa8c47f10041bac89da2865ac375f2372bf94276
3
- size 19245
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:92d32beca4388c280b67edaec12c1219c916254e0e29a1966ed971be66642766
3
+ size 19599