Hugging Face's logo

patrickvonplaten
/
unispeech-large-1500h-cv-timit

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
unispeech
timit_asr
generated_from_trainer
Inference Endpoints
Model card Files Metrics Community
4
patrickvonplaten commited on
Commit
7aac9c7
•
1 Parent(s): 2fb2942

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +90 -0
README.md ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - automatic-speech-recognition
4
+ - timit_asr
5
+ - generated_from_trainer
6
+ datasets:
7
+ - timit_asr
8
+ model-index:
9
+ - name: unispeech-large-1500h-cv-timit
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ # unispeech-large-1500h-cv-timit
17
+
18
+ This model is a fine-tuned version of [microsoft/unispeech-large-1500h-cv](https://huggingface.co/microsoft/unispeech-large-1500h-cv) on the TIMIT_ASR - NA dataset.
19
+ It achieves the following results on the evaluation set:
20
+ - Loss: 0.3099
21
+ - Wer: 0.2196
22
+
23
+ ## Model description
24
+
25
+ More information needed
26
+
27
+ ## Intended uses & limitations
28
+
29
+ More information needed
30
+
31
+ ## Training and evaluation data
32
+
33
+ More information needed
34
+
35
+ ## Training procedure
36
+
37
+ ### Training hyperparameters
38
+
39
+ The following hyperparameters were used during training:
40
+ - learning_rate: 0.0001
41
+ - train_batch_size: 32
42
+ - eval_batch_size: 1
43
+ - seed: 42
44
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
+ - lr_scheduler_type: linear
46
+ - lr_scheduler_warmup_steps: 1000
47
+ - num_epochs: 20.0
48
+ - mixed_precision_training: Native AMP
49
+
50
+ ### Training results
51
+
52
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
53
+ |:-------------:|:-----:|:----:|:---------------:|:------:|
54
+ | 4.64 | 0.69 | 100 | 3.9717 | 0.9981 |
55
+ | 2.6793 | 1.38 | 200 | 2.6264 | 1.0 |
56
+ | 1.2221 | 2.07 | 300 | 0.9999 | 0.7167 |
57
+ | 0.9009 | 2.76 | 400 | 0.6509 | 0.5570 |
58
+ | 0.4352 | 3.45 | 500 | 0.4682 | 0.4332 |
59
+ | 0.227 | 4.14 | 600 | 0.3661 | 0.3565 |
60
+ | 0.2169 | 4.83 | 700 | 0.3244 | 0.3203 |
61
+ | 0.2687 | 5.52 | 800 | 0.3137 | 0.2981 |
62
+ | 0.127 | 6.21 | 900 | 0.3220 | 0.2828 |
63
+ | 0.0922 | 6.9 | 1000 | 0.3075 | 0.2708 |
64
+ | 0.0965 | 7.59 | 1100 | 0.2779 | 0.2576 |
65
+ | 0.1298 | 8.28 | 1200 | 0.3111 | 0.2480 |
66
+ | 0.0855 | 8.97 | 1300 | 0.3021 | 0.2421 |
67
+ | 0.0629 | 9.66 | 1400 | 0.3122 | 0.2511 |
68
+ | 0.0471 | 10.34 | 1500 | 0.2965 | 0.2368 |
69
+ | 0.0871 | 11.03 | 1600 | 0.3247 | 0.2387 |
70
+ | 0.0503 | 11.72 | 1700 | 0.3359 | 0.2363 |
71
+ | 0.0402 | 12.41 | 1800 | 0.2976 | 0.2332 |
72
+ | 0.0336 | 13.1 | 1900 | 0.3139 | 0.2321 |
73
+ | 0.0634 | 13.79 | 2000 | 0.3188 | 0.2309 |
74
+ | 0.0429 | 14.48 | 2100 | 0.3145 | 0.2335 |
75
+ | 0.028 | 15.17 | 2200 | 0.3244 | 0.2242 |
76
+ | 0.0255 | 15.86 | 2300 | 0.2914 | 0.2196 |
77
+ | 0.0406 | 16.55 | 2400 | 0.3249 | 0.2202 |
78
+ | 0.0512 | 17.24 | 2500 | 0.3037 | 0.2198 |
79
+ | 0.0269 | 17.93 | 2600 | 0.3218 | 0.2242 |
80
+ | 0.0287 | 18.62 | 2700 | 0.3106 | 0.2185 |
81
+ | 0.0319 | 19.31 | 2800 | 0.3124 | 0.2217 |
82
+ | 0.0494 | 20.0 | 2900 | 0.3099 | 0.2196 |
83
+
84
+
85
+ ### Framework versions
86
+
87
+ - Transformers 4.12.0.dev0
88
+ - Pytorch 1.8.1
89
+ - Datasets 1.14.1.dev0
90
+ - Tokenizers 0.10.3