EYEDOL commited on
Commit
0827c6d
·
verified ·
1 Parent(s): a0b9afb

End of training

Browse files
README.md CHANGED
@@ -2,7 +2,7 @@
2
  language:
3
  - yo
4
  license: apache-2.0
5
- base_model: openai/whisper-small
6
  tags:
7
  - whisper-event
8
  - generated_from_trainer
@@ -25,7 +25,7 @@ model-index:
25
  metrics:
26
  - name: Wer
27
  type: wer
28
- value: 51.071878940731395
29
  ---
30
 
31
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -33,10 +33,10 @@ should probably proofread and complete it, then remove this comment. -->
33
 
34
  # WhisperYoruba
35
 
36
- This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 17.0 dataset.
37
  It achieves the following results on the evaluation set:
38
- - Loss: 1.0256
39
- - Wer: 51.0719
40
 
41
  ## Model description
42
 
@@ -62,16 +62,14 @@ 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
- - training_steps: 3500
66
  - mixed_precision_training: Native AMP
67
 
68
  ### Training results
69
 
70
- | Training Loss | Epoch | Step | Validation Loss | Wer |
71
- |:-------------:|:-------:|:----:|:---------------:|:-------:|
72
- | 0.2736 | 3.8462 | 1000 | 0.7361 | 51.3661 |
73
- | 0.0192 | 7.6923 | 2000 | 0.9077 | 51.7697 |
74
- | 0.0013 | 11.5385 | 3000 | 1.0256 | 51.0719 |
75
 
76
 
77
  ### Framework versions
 
2
  language:
3
  - yo
4
  license: apache-2.0
5
+ base_model: openai/whisper-medium
6
  tags:
7
  - whisper-event
8
  - generated_from_trainer
 
25
  metrics:
26
  - name: Wer
27
  type: wer
28
+ value: 64.04371584699453
29
  ---
30
 
31
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
33
 
34
  # WhisperYoruba
35
 
36
+ This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the Common Voice 17.0 dataset.
37
  It achieves the following results on the evaluation set:
38
+ - Loss: 0.6691
39
+ - Wer: 64.0437
40
 
41
  ## Model description
42
 
 
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
+ - training_steps: 1000
66
  - mixed_precision_training: Native AMP
67
 
68
  ### Training results
69
 
70
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
71
+ |:-------------:|:------:|:----:|:---------------:|:-------:|
72
+ | 0.1301 | 3.8462 | 1000 | 0.6691 | 64.0437 |
 
 
73
 
74
 
75
  ### Framework versions
config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "openai/whisper-small",
3
  "activation_dropout": 0.0,
4
  "activation_function": "gelu",
5
  "apply_spec_augment": false,
@@ -13,17 +13,17 @@
13
  ],
14
  "bos_token_id": 50257,
15
  "classifier_proj_size": 256,
16
- "d_model": 768,
17
- "decoder_attention_heads": 12,
18
- "decoder_ffn_dim": 3072,
19
  "decoder_layerdrop": 0.0,
20
- "decoder_layers": 12,
21
  "decoder_start_token_id": 50258,
22
  "dropout": 0.0,
23
- "encoder_attention_heads": 12,
24
- "encoder_ffn_dim": 3072,
25
  "encoder_layerdrop": 0.0,
26
- "encoder_layers": 12,
27
  "eos_token_id": 50257,
28
  "forced_decoder_ids": null,
29
  "init_std": 0.02,
@@ -39,7 +39,7 @@
39
  "max_target_positions": 448,
40
  "median_filter_width": 7,
41
  "model_type": "whisper",
42
- "num_hidden_layers": 12,
43
  "num_mel_bins": 80,
44
  "pad_token_id": 50257,
45
  "scale_embedding": false,
 
1
  {
2
+ "_name_or_path": "openai/whisper-medium",
3
  "activation_dropout": 0.0,
4
  "activation_function": "gelu",
5
  "apply_spec_augment": false,
 
13
  ],
14
  "bos_token_id": 50257,
15
  "classifier_proj_size": 256,
16
+ "d_model": 1024,
17
+ "decoder_attention_heads": 16,
18
+ "decoder_ffn_dim": 4096,
19
  "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 24,
21
  "decoder_start_token_id": 50258,
22
  "dropout": 0.0,
23
+ "encoder_attention_heads": 16,
24
+ "encoder_ffn_dim": 4096,
25
  "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 24,
27
  "eos_token_id": 50257,
28
  "forced_decoder_ids": null,
29
  "init_std": 0.02,
 
39
  "max_target_positions": 448,
40
  "median_filter_width": 7,
41
  "model_type": "whisper",
42
+ "num_hidden_layers": 24,
43
  "num_mel_bins": 80,
44
  "pad_token_id": 50257,
45
  "scale_embedding": false,
generation_config.json CHANGED
@@ -1,44 +1,28 @@
1
  {
2
  "alignment_heads": [
3
  [
4
- 5,
5
- 3
6
  ],
7
  [
8
- 5,
9
- 9
10
- ],
11
- [
12
- 8,
13
- 0
14
- ],
15
- [
16
- 8,
17
  4
18
  ],
19
  [
20
- 8,
21
- 7
22
  ],
23
  [
24
- 8,
25
- 8
26
  ],
27
  [
28
- 9,
29
  0
30
  ],
31
  [
32
- 9,
33
- 7
34
- ],
35
- [
36
- 9,
37
- 9
38
- ],
39
- [
40
- 10,
41
- 5
42
  ]
43
  ],
44
  "begin_suppress_tokens": [
 
1
  {
2
  "alignment_heads": [
3
  [
4
+ 13,
5
+ 15
6
  ],
7
  [
8
+ 15,
 
 
 
 
 
 
 
 
9
  4
10
  ],
11
  [
12
+ 15,
13
+ 15
14
  ],
15
  [
16
+ 16,
17
+ 1
18
  ],
19
  [
20
+ 20,
21
  0
22
  ],
23
  [
24
+ 23,
25
+ 4
 
 
 
 
 
 
 
 
26
  ]
27
  ],
28
  "begin_suppress_tokens": [
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2aaeb6a9f5453ef8c9f10b121fbe891589b8db71fe7ad22a0280817e9efe11fb
3
- size 966995080
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:46a7df1c1a237abf95ba1a0540d55a37894957964cac259c15fbae1e38ed241f
3
+ size 3055544304
runs/Jan10_20-51-02_c3ff67a08b40/events.out.tfevents.1736542265.c3ff67a08b40.24.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:805baccd030d1cfc41ee4d2fabd3bb108c293bd8380a2d8cd2a8ddfe936e00fe
3
+ size 14722
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:507d9af32281a031478aa4873225c61c2461c1d40d112876f755518e69632be0
3
  size 5240
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:892cb92fd398740902173be6a42eb0425a68d750c9d5ba13deebb182ebe69d0d
3
  size 5240