Hugging Face's logo

patrickvonplaten
/
unispeech-sat-base-timit-ft

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
unispeech-sat
timit_asr
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
1
patrickvonplaten commited on
Commit
9f029a7
•
1 Parent(s): 4971d46

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-sat-base-timit-ft
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-sat-base-timit-ft
17
+
18
+ This model is a fine-tuned version of [microsoft/unispeech-sat-base](https://huggingface.co/microsoft/unispeech-sat-base) on the TIMIT_ASR - NA dataset.
19
+ It achieves the following results on the evaluation set:
20
+ - Loss: 0.6712
21
+ - Wer: 0.4101
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
+ | 3.2582 | 0.69 | 100 | 3.1651 | 1.0 |
55
+ | 2.9542 | 1.38 | 200 | 2.9567 | 1.0 |
56
+ | 2.9656 | 2.07 | 300 | 2.9195 | 1.0 |
57
+ | 2.8946 | 2.76 | 400 | 2.8641 | 1.0 |
58
+ | 1.9305 | 3.45 | 500 | 1.7680 | 1.0029 |
59
+ | 1.0134 | 4.14 | 600 | 1.0184 | 0.6942 |
60
+ | 0.8355 | 4.83 | 700 | 0.7769 | 0.6080 |
61
+ | 0.8724 | 5.52 | 800 | 0.7182 | 0.6035 |
62
+ | 0.5619 | 6.21 | 900 | 0.6823 | 0.5406 |
63
+ | 0.4247 | 6.9 | 1000 | 0.6279 | 0.5237 |
64
+ | 0.4257 | 7.59 | 1100 | 0.6056 | 0.5000 |
65
+ | 0.5007 | 8.28 | 1200 | 0.5870 | 0.4918 |
66
+ | 0.3854 | 8.97 | 1300 | 0.6200 | 0.4804 |
67
+ | 0.264 | 9.66 | 1400 | 0.6030 | 0.4600 |
68
+ | 0.1989 | 10.34 | 1500 | 0.6049 | 0.4588 |
69
+ | 0.3196 | 11.03 | 1600 | 0.5946 | 0.4599 |
70
+ | 0.2622 | 11.72 | 1700 | 0.6282 | 0.4422 |
71
+ | 0.1697 | 12.41 | 1800 | 0.6559 | 0.4413 |
72
+ | 0.1464 | 13.1 | 1900 | 0.6349 | 0.4328 |
73
+ | 0.2277 | 13.79 | 2000 | 0.6133 | 0.4284 |
74
+ | 0.221 | 14.48 | 2100 | 0.6617 | 0.4219 |
75
+ | 0.1391 | 15.17 | 2200 | 0.6705 | 0.4235 |
76
+ | 0.112 | 15.86 | 2300 | 0.6207 | 0.4218 |
77
+ | 0.1717 | 16.55 | 2400 | 0.6749 | 0.4184 |
78
+ | 0.2081 | 17.24 | 2500 | 0.6756 | 0.4169 |
79
+ | 0.1244 | 17.93 | 2600 | 0.6750 | 0.4181 |
80
+ | 0.0978 | 18.62 | 2700 | 0.6500 | 0.4115 |
81
+ | 0.128 | 19.31 | 2800 | 0.6750 | 0.4106 |
82
+ | 0.1791 | 20.0 | 2900 | 0.6712 | 0.4101 |
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