Hugging Face's logo

lucio
/
xls-r-uyghur-cv8

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
Safetensors
Uyghur
wav2vec2
Generated from Trainer
hf-asr-leaderboard
mozilla-foundation/common_voice_8_0
robust-speech-event
Eval Results
Inference Endpoints
Model card Files Metrics Community
3
lucio commited on
Commit
c8b6183
•
1 Parent(s): a658482

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +43 -24
README.md CHANGED
@@ -20,8 +20,8 @@ should probably proofread and complete it, then remove this comment. -->
20
 
21
  This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the MOZILLA-FOUNDATION/COMMON_VOICE_8_0 - UG dataset.
22
  It achieves the following results on the evaluation set:
23
- - Loss: 0.2430
24
- - Wer: 0.3804
25
 
26
  ## Model description
27
 
@@ -40,7 +40,7 @@ More information needed
40
  ### Training hyperparameters
41
 
42
  The following hyperparameters were used during training:
43
- - learning_rate: 7.5e-05
44
  - train_batch_size: 8
45
  - eval_batch_size: 8
46
  - seed: 42
@@ -49,31 +49,50 @@ The following hyperparameters were used during training:
49
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
50
  - lr_scheduler_type: linear
51
  - lr_scheduler_warmup_steps: 2000
52
- - num_epochs: 50.0
53
  - mixed_precision_training: Native AMP
54
 
55
  ### Training results
56
 
57
- | Training Loss | Epoch | Step | Validation Loss | Wer |
58
- |:-------------:|:-----:|:----:|:---------------:|:------:|
59
- | 3.6871 | 2.66 | 500 | 3.5374 | 1.0 |
60
- | 3.1501 | 5.32 | 1000 | 3.1278 | 1.0 |
61
- | 1.5843 | 7.97 | 1500 | 0.6358 | 0.6914 |
62
- | 1.3378 | 10.64 | 2000 | 0.4422 | 0.5925 |
63
- | 1.2595 | 13.3 | 2500 | 0.3921 | 0.5512 |
64
- | 1.1643 | 15.95 | 3000 | 0.3507 | 0.5149 |
65
- | 1.1352 | 18.61 | 3500 | 0.3351 | 0.5019 |
66
- | 1.1113 | 21.28 | 4000 | 0.3153 | 0.4845 |
67
- | 1.0914 | 23.93 | 4500 | 0.3050 | 0.4594 |
68
- | 1.0468 | 26.59 | 5000 | 0.2890 | 0.4470 |
69
- | 1.0473 | 29.25 | 5500 | 0.2755 | 0.4331 |
70
- | 1.0065 | 31.91 | 6000 | 0.2718 | 0.4264 |
71
- | 0.9794 | 34.57 | 6500 | 0.2646 | 0.4193 |
72
- | 0.9849 | 37.23 | 7000 | 0.2610 | 0.4058 |
73
- | 0.9496 | 39.89 | 7500 | 0.2522 | 0.3985 |
74
- | 0.9367 | 42.55 | 8000 | 0.2514 | 0.3947 |
75
- | 0.9295 | 45.21 | 8500 | 0.2458 | 0.3883 |
76
- | 0.9187 | 47.87 | 9000 | 0.2439 | 0.3833 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
77
 
78
 
79
  ### Framework versions
 
20
 
21
  This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the MOZILLA-FOUNDATION/COMMON_VOICE_8_0 - UG dataset.
22
  It achieves the following results on the evaluation set:
23
+ - Loss: 0.2240
24
+ - Wer: 0.3693
25
 
26
  ## Model description
27
 
 
40
  ### Training hyperparameters
41
 
42
  The following hyperparameters were used during training:
43
+ - learning_rate: 4e-05
44
  - train_batch_size: 8
45
  - eval_batch_size: 8
46
  - seed: 42
 
49
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
50
  - lr_scheduler_type: linear
51
  - lr_scheduler_warmup_steps: 2000
52
+ - num_epochs: 100.0
53
  - mixed_precision_training: Native AMP
54
 
55
  ### Training results
56
 
57
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
58
+ |:-------------:|:-----:|:-----:|:---------------:|:------:|
59
+ | 4.1169 | 2.66 | 500 | 4.0146 | 1.0 |
60
+ | 3.2512 | 5.32 | 1000 | 3.2342 | 1.0 |
61
+ | 2.5435 | 7.97 | 1500 | 1.8155 | 1.0286 |
62
+ | 1.5575 | 10.64 | 2000 | 0.6346 | 0.7058 |
63
+ | 1.3979 | 13.3 | 2500 | 0.4885 | 0.6320 |
64
+ | 1.2874 | 15.95 | 3000 | 0.4271 | 0.6088 |
65
+ | 1.2383 | 18.61 | 3500 | 0.3889 | 0.5869 |
66
+ | 1.2054 | 21.28 | 4000 | 0.3609 | 0.5793 |
67
+ | 1.1866 | 23.93 | 4500 | 0.3450 | 0.5513 |
68
+ | 1.1332 | 26.59 | 5000 | 0.3214 | 0.5379 |
69
+ | 1.135 | 29.25 | 5500 | 0.3122 | 0.5384 |
70
+ | 1.0992 | 31.91 | 6000 | 0.2948 | 0.5078 |
71
+ | 1.0707 | 34.57 | 6500 | 0.2928 | 0.5128 |
72
+ | 1.0754 | 37.23 | 7000 | 0.2857 | 0.5017 |
73
+ | 1.0461 | 39.89 | 7500 | 0.2791 | 0.5099 |
74
+ | 1.0328 | 42.55 | 8000 | 0.2729 | 0.5120 |
75
+ | 1.0201 | 45.21 | 8500 | 0.2654 | 0.4720 |
76
+ | 1.0035 | 47.87 | 9000 | 0.2623 | 0.4659 |
77
+ | 1.0069 | 50.53 | 9500 | 0.2569 | 0.4593 |
78
+ | 0.9998 | 53.19 | 10000 | 0.2519 | 0.4405 |
79
+ | 0.9762 | 55.85 | 10500 | 0.2505 | 0.4588 |
80
+ | 0.9755 | 58.51 | 11000 | 0.2479 | 0.4564 |
81
+ | 0.9624 | 61.17 | 11500 | 0.2460 | 0.4298 |
82
+ | 0.9494 | 63.83 | 12000 | 0.2402 | 0.4182 |
83
+ | 0.948 | 66.49 | 12500 | 0.2412 | 0.4212 |
84
+ | 0.9312 | 69.15 | 13000 | 0.2352 | 0.3970 |
85
+ | 0.9172 | 71.81 | 13500 | 0.2357 | 0.3926 |
86
+ | 0.9101 | 74.47 | 14000 | 0.2305 | 0.3905 |
87
+ | 0.9177 | 77.13 | 14500 | 0.2307 | 0.3838 |
88
+ | 0.9083 | 79.78 | 15000 | 0.2313 | 0.3800 |
89
+ | 0.9068 | 82.45 | 15500 | 0.2275 | 0.3742 |
90
+ | 0.9087 | 85.11 | 16000 | 0.2283 | 0.3747 |
91
+ | 0.8838 | 87.76 | 16500 | 0.2286 | 0.3777 |
92
+ | 0.8868 | 90.42 | 17000 | 0.2269 | 0.3722 |
93
+ | 0.8895 | 93.08 | 17500 | 0.2246 | 0.3714 |
94
+ | 0.8926 | 95.74 | 18000 | 0.2241 | 0.3705 |
95
+ | 0.8856 | 98.4 | 18500 | 0.2242 | 0.3693 |
96
 
97
 
98
  ### Framework versions