Hugging Face's logo

sabroo
/
w2v-bert-2.0-slovak-colab-CV17.0

Automatic Speech Recognition
Transformers
Safetensors
wav2vec2-bert
Generated from Trainer
Eval Results
Inference Endpoints
Model card Files Community
sabroo commited on
Commit
48587fa
·
verified ·
1 Parent(s): 0e8dc1c

End of training

Browse files
Files changed (4) hide show
  1. README.md +22 -19
  2. config.json +2 -2
  3. model.safetensors +2 -2
  4. training_args.bin +1 -1
README.md CHANGED
@@ -1,19 +1,19 @@
1
  ---
 
 
2
  base_model: facebook/w2v-bert-2.0
 
 
3
  datasets:
4
  - common_voice_17_0
5
- library_name: transformers
6
- license: mit
7
  metrics:
8
  - wer
9
- tags:
10
- - generated_from_trainer
11
  model-index:
12
  - name: w2v-bert-2.0-slovak-colab-CV17.0
13
  results:
14
  - task:
15
- type: automatic-speech-recognition
16
  name: Automatic Speech Recognition
 
17
  dataset:
18
  name: common_voice_17_0
19
  type: common_voice_17_0
@@ -21,9 +21,9 @@ model-index:
21
  split: test
22
  args: sk
23
  metrics:
24
- - type: wer
25
- value: 0.13279330117411486
26
- name: Wer
27
  ---
28
 
29
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -33,8 +33,8 @@ should probably proofread and complete it, then remove this comment. -->
33
 
34
  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_17_0 dataset.
35
  It achieves the following results on the evaluation set:
36
- - Loss: 0.3002
37
- - Wer: 0.1328
38
 
39
  ## Model description
40
 
@@ -62,19 +62,22 @@ The following hyperparameters were used during training:
62
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
63
  - lr_scheduler_type: linear
64
  - lr_scheduler_warmup_steps: 500
65
- - num_epochs: 10
66
  - mixed_precision_training: Native AMP
67
 
68
  ### Training results
69
 
70
- | Training Loss | Epoch | Step | Validation Loss | Wer |
71
- |:-------------:|:------:|:----:|:---------------:|:------:|
72
- | 2.3467 | 1.6393 | 300 | 0.3488 | 0.2605 |
73
- | 0.1905 | 3.2787 | 600 | 0.3339 | 0.2059 |
74
- | 0.1121 | 4.9180 | 900 | 0.3009 | 0.1849 |
75
- | 0.0592 | 6.5574 | 1200 | 0.2817 | 0.1482 |
76
- | 0.0264 | 8.1967 | 1500 | 0.3114 | 0.1385 |
77
- | 0.0094 | 9.8361 | 1800 | 0.3002 | 0.1328 |
 
 
 
78
 
79
 
80
  ### Framework versions
 
1
  ---
2
+ library_name: transformers
3
+ license: mit
4
  base_model: facebook/w2v-bert-2.0
5
+ tags:
6
+ - generated_from_trainer
7
  datasets:
8
  - common_voice_17_0
 
 
9
  metrics:
10
  - wer
 
 
11
  model-index:
12
  - name: w2v-bert-2.0-slovak-colab-CV17.0
13
  results:
14
  - task:
 
15
  name: Automatic Speech Recognition
16
+ type: automatic-speech-recognition
17
  dataset:
18
  name: common_voice_17_0
19
  type: common_voice_17_0
 
21
  split: test
22
  args: sk
23
  metrics:
24
+ - name: Wer
25
+ type: wer
26
+ value: 0.1358878674797488
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/w2v-bert-2.0](https://huggingface.co/facebook/w2v-bert-2.0) on the common_voice_17_0 dataset.
35
  It achieves the following results on the evaluation set:
36
+ - Loss: 0.3823
37
+ - Wer: 0.1359
38
 
39
  ## Model description
40
 
 
62
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
63
  - lr_scheduler_type: linear
64
  - lr_scheduler_warmup_steps: 500
65
+ - num_epochs: 15
66
  - mixed_precision_training: Native AMP
67
 
68
  ### Training results
69
 
70
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
71
+ |:-------------:|:-------:|:----:|:---------------:|:------:|
72
+ | 3.1601 | 1.6393 | 300 | 0.4869 | 0.2985 |
73
+ | 0.2141 | 3.2787 | 600 | 0.3886 | 0.2144 |
74
+ | 0.1323 | 4.9180 | 900 | 0.3180 | 0.1840 |
75
+ | 0.0754 | 6.5574 | 1200 | 0.3019 | 0.1750 |
76
+ | 0.0401 | 8.1967 | 1500 | 0.3717 | 0.1525 |
77
+ | 0.022 | 9.8361 | 1800 | 0.3408 | 0.1503 |
78
+ | 0.0083 | 11.4754 | 2100 | 0.3489 | 0.1413 |
79
+ | 0.0027 | 13.1148 | 2400 | 0.3681 | 0.1358 |
80
+ | 0.0011 | 14.7541 | 2700 | 0.3823 | 0.1359 |
81
 
82
 
83
  ### Framework versions
config.json CHANGED
@@ -47,7 +47,7 @@
47
  "num_hidden_layers": 24,
48
  "num_negatives": 100,
49
  "output_hidden_size": 1024,
50
- "pad_token_id": 48,
51
  "position_embeddings_type": "relative_key",
52
  "proj_codevector_dim": 768,
53
  "right_max_position_embeddings": 8,
@@ -77,6 +77,6 @@
77
  "transformers_version": "4.45.2",
78
  "use_intermediate_ffn_before_adapter": false,
79
  "use_weighted_layer_sum": false,
80
- "vocab_size": 51,
81
  "xvector_output_dim": 512
82
  }
 
47
  "num_hidden_layers": 24,
48
  "num_negatives": 100,
49
  "output_hidden_size": 1024,
50
+ "pad_token_id": 46,
51
  "position_embeddings_type": "relative_key",
52
  "proj_codevector_dim": 768,
53
  "right_max_position_embeddings": 8,
 
77
  "transformers_version": "4.45.2",
78
  "use_intermediate_ffn_before_adapter": false,
79
  "use_weighted_layer_sum": false,
80
+ "vocab_size": 49,
81
  "xvector_output_dim": 512
82
  }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0c9d1233c731a7883e4476561b845923d14a363e3a8c618c322fe19d74080c27
3
- size 2423023660
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:22c399b1ffb31d288ee510d0fb07f556f93c4ecf87c08a5ca0e51adb4e1ddb58
3
+ size 2423015460
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3c3e6a73a5edb73730efbca5e8e0e5afcbe1b71c17a1ad83d7a11896f11c034c
3
  size 5176
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:67b8d9a2dbf6c2ef5db49f5f257ebcfd3b4bed49a85f5ebc1fa76e1332f3a022
3
  size 5176