Hugging Face's logo

infinitejoy
/
wav2vec2-large-xls-r-300m-indonesian

Automatic Speech Recognition
Transformers
PyTorch
Indonesian
wav2vec2
mozilla-foundation/common_voice_7_0
Generated from Trainer
Model card Files
xet
Community
1
infinitejoy commited on
Commit
4e0810e
·
1 Parent(s): 07c24cc

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +25 -14
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_7_0 - ID dataset.
22
  It achieves the following results on the evaluation set:
23
- - Loss: 0.2750
24
- - Wer: 0.3575
25
 
26
  ## Model description
27
 
@@ -46,24 +46,35 @@ The following hyperparameters were used during training:
46
  - seed: 42
47
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
  - lr_scheduler_type: linear
49
- - lr_scheduler_warmup_steps: 2000
50
- - num_epochs: 50.0
51
  - mixed_precision_training: Native AMP
52
 
53
  ### Training results
54
 
55
  | Training Loss | Epoch | Step | Validation Loss | Wer |
56
  |:-------------:|:-----:|:-----:|:---------------:|:------:|
57
- | 2.8112 | 4.72 | 1000 | 2.7786 | 1.0 |
58
- | 1.3803 | 9.43 | 2000 | 0.4971 | 0.5788 |
59
- | 1.1894 | 14.15 | 3000 | 0.3605 | 0.4442 |
60
- | 1.1019 | 18.87 | 4000 | 0.3252 | 0.4236 |
61
- | 1.0366 | 23.58 | 5000 | 0.3127 | 0.4023 |
62
- | 1.0217 | 28.3 | 6000 | 0.3027 | 0.3953 |
63
- | 0.9774 | 33.02 | 7000 | 0.2924 | 0.3965 |
64
- | 0.9485 | 37.74 | 8000 | 0.2949 | 0.3755 |
65
- | 0.9332 | 42.45 | 9000 | 0.2815 | 0.3665 |
66
- | 0.9093 | 47.17 | 10000 | 0.2780 | 0.3623 |
 
 
 
 
 
 
 
 
 
 
 
67
 
68
 
69
  ### 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_7_0 - ID dataset.
22
  It achieves the following results on the evaluation set:
23
+ - Loss: 0.2759
24
+ - Wer: 0.3256
25
 
26
  ## Model description
27
 
 
46
  - seed: 42
47
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
  - lr_scheduler_type: linear
49
+ - lr_scheduler_warmup_steps: 4000
50
+ - num_epochs: 100.0
51
  - mixed_precision_training: Native AMP
52
 
53
  ### Training results
54
 
55
  | Training Loss | Epoch | Step | Validation Loss | Wer |
56
  |:-------------:|:-----:|:-----:|:---------------:|:------:|
57
+ | 3.0387 | 4.72 | 1000 | 3.0892 | 1.0 |
58
+ | 1.7911 | 9.43 | 2000 | 0.8451 | 0.6702 |
59
+ | 1.2826 | 14.15 | 3000 | 0.4211 | 0.4166 |
60
+ | 1.1802 | 18.87 | 4000 | 0.3508 | 0.4690 |
61
+ | 1.1065 | 23.58 | 5000 | 0.3319 | 0.4662 |
62
+ | 1.0921 | 28.3 | 6000 | 0.3056 | 0.3880 |
63
+ | 1.0366 | 33.02 | 7000 | 0.2997 | 0.3665 |
64
+ | 0.9988 | 37.74 | 8000 | 0.2972 | 0.3653 |
65
+ | 0.9864 | 42.45 | 9000 | 0.2697 | 0.3371 |
66
+ | 0.9558 | 47.17 | 10000 | 0.2739 | 0.3141 |
67
+ | 0.9094 | 51.89 | 11000 | 0.2657 | 0.3533 |
68
+ | 0.9034 | 56.6 | 12000 | 0.2699 | 0.3397 |
69
+ | 0.8907 | 61.32 | 13000 | 0.2765 | 0.3470 |
70
+ | 0.8631 | 66.04 | 14000 | 0.2774 | 0.3346 |
71
+ | 0.8389 | 70.75 | 15000 | 0.2743 | 0.3365 |
72
+ | 0.8214 | 75.47 | 16000 | 0.2778 | 0.3201 |
73
+ | 0.8195 | 80.19 | 17000 | 0.2725 | 0.3286 |
74
+ | 0.7994 | 84.91 | 18000 | 0.2782 | 0.3315 |
75
+ | 0.7816 | 89.62 | 19000 | 0.2775 | 0.3363 |
76
+ | 0.7816 | 94.34 | 20000 | 0.2731 | 0.3278 |
77
+ | 0.7635 | 99.06 | 21000 | 0.2767 | 0.3259 |
78
 
79
 
80
  ### Framework versions