Hugging Face's logo

halcyonzhou
/
unispeech-sat-base

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
unispeech-sat
Generated from Trainer
Eval Results (legacy)
Model card Files
xet
 Metrics Community
halcyonzhou commited on
Commit
9c7d628
·
verified ·
1 Parent(s): 7406216

End of training

Browse files
Files changed (3) hide show
  1. README.md +29 -18
  2. model.safetensors +1 -1
  3. runs/Aug30_21-31-22_zjh/events.out.tfevents.1756560699.zjh.7320.0 +2 -2
README.md CHANGED
@@ -2,11 +2,26 @@
2
  library_name: transformers
3
  tags:
4
  - generated_from_trainer
 
 
5
  metrics:
6
  - wer
7
  model-index:
8
  - name: unispeech-sat-base
9
- results: []
 
 
 
 
 
 
 
 
 
 
 
 
 
10
  ---
11
 
12
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -14,10 +29,10 @@ should probably proofread and complete it, then remove this comment. -->
14
 
15
  # unispeech-sat-base
16
 
17
- This model was trained from scratch on an unknown dataset.
18
  It achieves the following results on the evaluation set:
19
- - Loss: 0.6338
20
- - Wer: 0.1673
21
 
22
  ## Model description
23
 
@@ -37,36 +52,32 @@ More information needed
37
 
38
  The following hyperparameters were used during training:
39
  - learning_rate: 3e-05
40
- - train_batch_size: 4
41
  - eval_batch_size: 4
42
  - seed: 42
43
  - gradient_accumulation_steps: 2
44
- - total_train_batch_size: 8
45
  - optimizer: Use adafactor and the args are:
46
  No additional optimizer arguments
47
  - lr_scheduler_type: linear
48
  - lr_scheduler_warmup_steps: 100
49
- - training_steps: 5000
50
 
51
  ### Training results
52
 
53
  | Training Loss | Epoch | Step | Validation Loss | Wer |
54
  |:-------------:|:-------:|:----:|:---------------:|:------:|
55
- | 0.1101 | 8.7788 | 500 | 0.5401 | 0.1610 |
56
- | 0.0732 | 17.5487 | 1000 | 0.5491 | 0.1642 |
57
- | 0.0825 | 26.3186 | 1500 | 0.5649 | 0.1704 |
58
- | 0.0648 | 35.0885 | 2000 | 0.5176 | 0.1597 |
59
- | 0.0444 | 43.8673 | 2500 | 0.6058 | 0.1698 |
60
- | 0.0555 | 52.6372 | 3000 | 0.6108 | 0.1660 |
61
- | 0.0418 | 61.4071 | 3500 | 0.6451 | 0.1686 |
62
- | 0.0438 | 70.1770 | 4000 | 0.6274 | 0.1667 |
63
- | 0.0379 | 78.9558 | 4500 | 0.6266 | 0.1660 |
64
- | 0.0361 | 87.7257 | 5000 | 0.6338 | 0.1673 |
65
 
66
 
67
  ### Framework versions
68
 
69
- - Transformers 4.55.1
70
  - Pytorch 2.8.0+cu129
71
  - Datasets 3.6.0
72
  - Tokenizers 0.21.4
 
2
  library_name: transformers
3
  tags:
4
  - generated_from_trainer
5
+ datasets:
6
+ - minds14
7
  metrics:
8
  - wer
9
  model-index:
10
  - name: unispeech-sat-base
11
+ results:
12
+ - task:
13
+ name: Automatic Speech Recognition
14
+ type: automatic-speech-recognition
15
+ dataset:
16
+ name: minds14
17
+ type: minds14
18
+ config: en-US
19
+ split: None
20
+ args: en-US
21
+ metrics:
22
+ - name: Wer
23
+ type: wer
24
+ value: 0.23105360443622922
25
  ---
26
 
27
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
29
 
30
  # unispeech-sat-base
31
 
32
+ This model was trained from scratch on the minds14 dataset.
33
  It achieves the following results on the evaluation set:
34
+ - Loss: 0.4374
35
+ - Wer: 0.2311
36
 
37
  ## Model description
38
 
 
52
 
53
  The following hyperparameters were used during training:
54
  - learning_rate: 3e-05
55
+ - train_batch_size: 2
56
  - eval_batch_size: 4
57
  - seed: 42
58
  - gradient_accumulation_steps: 2
59
+ - total_train_batch_size: 4
60
  - optimizer: Use adafactor and the args are:
61
  No additional optimizer arguments
62
  - lr_scheduler_type: linear
63
  - lr_scheduler_warmup_steps: 100
64
+ - training_steps: 3000
65
 
66
  ### Training results
67
 
68
  | Training Loss | Epoch | Step | Validation Loss | Wer |
69
  |:-------------:|:-------:|:----:|:---------------:|:------:|
70
+ | 0.3806 | 4.4267 | 500 | 0.2892 | 0.1793 |
71
+ | 0.3067 | 8.8533 | 1000 | 0.4070 | 0.2058 |
72
+ | 0.3009 | 13.2756 | 1500 | 0.4186 | 0.2150 |
73
+ | 0.2842 | 17.7022 | 2000 | 0.6049 | 0.2434 |
74
+ | 0.2608 | 22.1244 | 2500 | 0.4818 | 0.2335 |
75
+ | 0.2639 | 26.5511 | 3000 | 0.4374 | 0.2311 |
 
 
 
 
76
 
77
 
78
  ### Framework versions
79
 
80
+ - Transformers 4.51.0
81
  - Pytorch 2.8.0+cu129
82
  - Datasets 3.6.0
83
  - Tokenizers 0.21.4
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3cab054dd036f9318fc294b304ba8b8d3086eeb4aa2fcbdb574b2dd8c529af94
3
  size 377612176
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a604047bcfc92fcae2c3fb81106ba6e323bbe3a0c9147a24c3ac07b1b47c55e6
3
  size 377612176
runs/Aug30_21-31-22_zjh/events.out.tfevents.1756560699.zjh.7320.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5e372b9501544575f7319a52bfa074cfb62ac69c83587bd21dda912e7586be34
3
- size 10362
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4ce9bbc883b30ef32633dad5c9870de49d54da7c8bb1f4c6a3bfab91d8d8231b
3
+ size 14835