Hugging Face's logo

BriereAssia
/
w2v3

Automatic Speech Recognition
Transformers
Safetensors
wav2vec2-bert
Generated from Trainer
Eval Results
Model card Files
xet
Community
BriereAssia commited on
Commit
4c6be2d
·
verified ·
1 Parent(s): febeb1f

BriereAssia/wav2vec2_finetuned_model_86_v2

Browse files
Files changed (4) hide show
  1. README.md +27 -27
  2. config.json +1 -2
  3. model.safetensors +1 -1
  4. training_args.bin +1 -1
README.md CHANGED
@@ -24,10 +24,10 @@ model-index:
24
  metrics:
25
  - name: Wer
26
  type: wer
27
- value: 0.14389960749626846
28
  - name: Bleu
29
  type: bleu
30
- value: 0.6267894360640803
31
  ---
32
 
33
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -37,11 +37,11 @@ should probably proofread and complete it, then remove this comment. -->
37
 
38
  This model is a fine-tuned version of [facebook/w2v-bert-2.0](https://huggingface.co/facebook/w2v-bert-2.0) on the common_voice_11_0 dataset.
39
  It achieves the following results on the evaluation set:
40
- - Loss: 0.1713
41
- - Wer: 0.1439
42
  - Cer: 0.0349
43
- - Bleu: 0.6268
44
- - Bert Score F1: 0.9719
45
 
46
  ## Model description
47
 
@@ -72,31 +72,31 @@ The following hyperparameters were used during training:
72
 
73
  | Training Loss | Epoch | Step | Validation Loss | Wer | Cer | Bleu | Bert Score F1 |
74
  |:-------------:|:------:|:----:|:---------------:|:------:|:------:|:------:|:-------------:|
75
- | 3.2221 | 0.0357 | 250 | 0.6698 | 0.5755 | 0.1572 | 0.1621 | 0.8585 |
76
- | 0.6124 | 0.0713 | 500 | 0.3894 | 0.2558 | 0.0702 | 0.4591 | 0.9370 |
77
- | 0.3391 | 0.1070 | 750 | 0.3347 | 0.2498 | 0.0672 | 0.4715 | 0.9382 |
78
- | 0.2579 | 0.1426 | 1000 | 0.2844 | 0.2047 | 0.0547 | 0.5279 | 0.9519 |
79
- | 0.2819 | 0.1783 | 1250 | 0.2671 | 0.2083 | 0.0538 | 0.5314 | 0.9543 |
80
- | 0.2042 | 0.2139 | 1500 | 0.2500 | 0.2029 | 0.0502 | 0.5389 | 0.9579 |
81
- | 0.2262 | 0.2496 | 1750 | 0.2324 | 0.1807 | 0.0466 | 0.5678 | 0.9611 |
82
- | 0.2001 | 0.2853 | 2000 | 0.2181 | 0.1777 | 0.0434 | 0.5733 | 0.9642 |
83
- | 0.1609 | 0.3209 | 2250 | 0.2095 | 0.1705 | 0.0430 | 0.5845 | 0.9648 |
84
- | 0.2138 | 0.3566 | 2500 | 0.2064 | 0.1561 | 0.0395 | 0.6074 | 0.9678 |
85
- | 0.1736 | 0.3922 | 2750 | 0.2018 | 0.1630 | 0.0398 | 0.5954 | 0.9675 |
86
- | 0.1622 | 0.4279 | 3000 | 0.2000 | 0.1602 | 0.0389 | 0.6004 | 0.9682 |
87
- | 0.1737 | 0.4636 | 3250 | 0.2017 | 0.1557 | 0.0383 | 0.6093 | 0.9685 |
88
- | 0.1525 | 0.4992 | 3500 | 0.1844 | 0.1460 | 0.0362 | 0.6230 | 0.9696 |
89
- | 0.1615 | 0.5349 | 3750 | 0.1947 | 0.1496 | 0.0365 | 0.6171 | 0.9704 |
90
- | 0.1558 | 0.5705 | 4000 | 0.1806 | 0.1488 | 0.0364 | 0.6184 | 0.9705 |
91
- | 0.125 | 0.6062 | 4250 | 0.1792 | 0.1468 | 0.0352 | 0.6234 | 0.9716 |
92
- | 0.1281 | 0.6418 | 4500 | 0.1785 | 0.1460 | 0.0353 | 0.6230 | 0.9717 |
93
- | 0.1321 | 0.6775 | 4750 | 0.1732 | 0.1453 | 0.0350 | 0.6245 | 0.9718 |
94
- | 0.1356 | 0.7132 | 5000 | 0.1713 | 0.1439 | 0.0349 | 0.6268 | 0.9719 |
95
 
96
 
97
  ### Framework versions
98
 
99
- - Transformers 4.49.0
100
  - Pytorch 2.5.1+cu121
101
  - Datasets 3.3.1
102
  - Tokenizers 0.21.0
 
24
  metrics:
25
  - name: Wer
26
  type: wer
27
+ value: 0.14435763249060218
28
  - name: Bleu
29
  type: bleu
30
+ value: 0.625443124553845
31
  ---
32
 
33
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
37
 
38
  This model is a fine-tuned version of [facebook/w2v-bert-2.0](https://huggingface.co/facebook/w2v-bert-2.0) on the common_voice_11_0 dataset.
39
  It achieves the following results on the evaluation set:
40
+ - Loss: 0.1747
41
+ - Wer: 0.1444
42
  - Cer: 0.0349
43
+ - Bleu: 0.6254
44
+ - Bert Score F1: 0.9721
45
 
46
  ## Model description
47
 
 
72
 
73
  | Training Loss | Epoch | Step | Validation Loss | Wer | Cer | Bleu | Bert Score F1 |
74
  |:-------------:|:------:|:----:|:---------------:|:------:|:------:|:------:|:-------------:|
75
+ | 0.3995 | 0.0357 | 250 | 0.3664 | 0.2725 | 0.0732 | 0.4414 | 0.9332 |
76
+ | 0.3065 | 0.0713 | 500 | 0.3399 | 0.2119 | 0.0593 | 0.5188 | 0.9465 |
77
+ | 0.2648 | 0.1070 | 750 | 0.3095 | 0.2327 | 0.0633 | 0.4970 | 0.9430 |
78
+ | 0.2393 | 0.1426 | 1000 | 0.2885 | 0.2134 | 0.0551 | 0.5156 | 0.9545 |
79
+ | 0.2756 | 0.1783 | 1250 | 0.2486 | 0.1817 | 0.0467 | 0.5670 | 0.9614 |
80
+ | 0.2005 | 0.2139 | 1500 | 0.2448 | 0.1935 | 0.0482 | 0.5485 | 0.9588 |
81
+ | 0.2112 | 0.2496 | 1750 | 0.2377 | 0.1823 | 0.0464 | 0.5617 | 0.9622 |
82
+ | 0.1934 | 0.2853 | 2000 | 0.2226 | 0.1674 | 0.0420 | 0.5888 | 0.9658 |
83
+ | 0.1631 | 0.3209 | 2250 | 0.2205 | 0.1660 | 0.0421 | 0.5888 | 0.9647 |
84
+ | 0.1905 | 0.3566 | 2500 | 0.2249 | 0.1679 | 0.0429 | 0.5879 | 0.9651 |
85
+ | 0.1639 | 0.3922 | 2750 | 0.2026 | 0.1625 | 0.0403 | 0.5975 | 0.9673 |
86
+ | 0.1567 | 0.4279 | 3000 | 0.1895 | 0.1516 | 0.0379 | 0.6150 | 0.9685 |
87
+ | 0.1641 | 0.4636 | 3250 | 0.1984 | 0.1555 | 0.0379 | 0.6076 | 0.9693 |
88
+ | 0.1404 | 0.4992 | 3500 | 0.1876 | 0.1528 | 0.0370 | 0.6124 | 0.9696 |
89
+ | 0.1475 | 0.5349 | 3750 | 0.1913 | 0.1568 | 0.0381 | 0.6055 | 0.9691 |
90
+ | 0.1586 | 0.5705 | 4000 | 0.1846 | 0.1510 | 0.0366 | 0.6151 | 0.9705 |
91
+ | 0.1322 | 0.6062 | 4250 | 0.1801 | 0.1475 | 0.0356 | 0.6208 | 0.9715 |
92
+ | 0.1396 | 0.6418 | 4500 | 0.1788 | 0.1454 | 0.0351 | 0.6242 | 0.9720 |
93
+ | 0.1287 | 0.6775 | 4750 | 0.1755 | 0.1455 | 0.0352 | 0.6233 | 0.9718 |
94
+ | 0.1376 | 0.7132 | 5000 | 0.1747 | 0.1444 | 0.0349 | 0.6254 | 0.9721 |
95
 
96
 
97
  ### Framework versions
98
 
99
+ - Transformers 4.50.0
100
  - Pytorch 2.5.1+cu121
101
  - Datasets 3.3.1
102
  - Tokenizers 0.21.0
config.json CHANGED
@@ -1,5 +1,4 @@
1
  {
2
- "_name_or_path": "facebook/w2v-bert-2.0",
3
  "activation_dropout": 0.0,
4
  "adapter_act": "relu",
5
  "adapter_kernel_size": 3,
@@ -74,7 +73,7 @@
74
  1
75
  ],
76
  "torch_dtype": "float32",
77
- "transformers_version": "4.49.0",
78
  "use_intermediate_ffn_before_adapter": false,
79
  "use_weighted_layer_sum": false,
80
  "vocab_size": 77,
 
1
  {
 
2
  "activation_dropout": 0.0,
3
  "adapter_act": "relu",
4
  "adapter_kernel_size": 3,
 
73
  1
74
  ],
75
  "torch_dtype": "float32",
76
+ "transformers_version": "4.50.0",
77
  "use_intermediate_ffn_before_adapter": false,
78
  "use_weighted_layer_sum": false,
79
  "vocab_size": 77,
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:4990c79313d597fe147e31917b8a8e07cbf050f2832ba941c14f0a967dc4db54
3
  size 2423130260
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:30136b522125c1592658c24a30df3d3a07be5b83b96d91e72e59e2c89584b43d
3
  size 2423130260
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:718ab1836dab9b9949b83db5b99ac96ddb6fabebe306d84a3e5d2ca2da015545
3
  size 5240
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0d54a4c2ffc67f34f5d38086221d76396ae0f75db88d94292ec0b941188ee6fe
3
  size 5240