Hugging Face's logo

Mofe
/
xls-r-hausa-40

Automatic Speech Recognition
Transformers
PyTorch
Hausa
wav2vec2
mozilla-foundation/common_voice_8_0
Generated from Trainer
robust-speech-event
hf-asr-leaderboard
Inference Endpoints
Model card Files Community
Mofe commited on
Commit
5242e86
•
1 Parent(s): f83b9ab

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +16 -16
README.md CHANGED
@@ -1,10 +1,6 @@
1
  ---
2
- language:
3
- - ha
4
  license: apache-2.0
5
  tags:
6
- - automatic-speech-recognition
7
- - mozilla-foundation/common_voice_8_0
8
  - generated_from_trainer
9
  datasets:
10
  - common_voice
@@ -18,10 +14,10 @@ should probably proofread and complete it, then remove this comment. -->
18
 
19
  #
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 - HA dataset.
22
  It achieves the following results on the evaluation set:
23
- - Loss: 0.7424
24
- - Wer: 0.6190
25
 
26
  ## Model description
27
 
@@ -40,27 +36,31 @@ More information needed
40
  ### Training hyperparameters
41
 
42
  The following hyperparameters were used during training:
43
- - learning_rate: 0.0003
44
- - train_batch_size: 16
45
  - eval_batch_size: 8
46
  - seed: 42
47
  - gradient_accumulation_steps: 4
48
- - total_train_batch_size: 64
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: 90.0
53
  - mixed_precision_training: Native AMP
54
 
55
  ### Training results
56
 
57
  | Training Loss | Epoch | Step | Validation Loss | Wer |
58
  |:-------------:|:-----:|:----:|:---------------:|:------:|
59
- | 2.721 | 16.66 | 500 | 2.6876 | 1.0 |
60
- | 1.2944 | 33.33 | 1000 | 0.6482 | 0.7135 |
61
- | 1.0515 | 49.98 | 1500 | 0.6645 | 0.6754 |
62
- | 0.9153 | 66.66 | 2000 | 0.7649 | 0.6936 |
63
- | 0.7056 | 83.33 | 2500 | 0.8048 | 0.6756 |
 
 
 
 
64
 
65
 
66
  ### Framework versions
 
1
  ---
 
 
2
  license: apache-2.0
3
  tags:
 
 
4
  - generated_from_trainer
5
  datasets:
6
  - common_voice
 
14
 
15
  #
16
 
17
+ This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the common_voice dataset.
18
  It achieves the following results on the evaluation set:
19
+ - Loss: 0.4939
20
+ - Wer: 0.5737
21
 
22
  ## Model description
23
 
 
36
  ### Training hyperparameters
37
 
38
  The following hyperparameters were used during training:
39
+ - learning_rate: 7.5e-05
40
+ - train_batch_size: 8
41
  - eval_batch_size: 8
42
  - seed: 42
43
  - gradient_accumulation_steps: 4
44
+ - total_train_batch_size: 32
45
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
46
  - lr_scheduler_type: linear
47
  - lr_scheduler_warmup_steps: 2000
48
+ - num_epochs: 80.0
49
  - mixed_precision_training: Native AMP
50
 
51
  ### Training results
52
 
53
  | Training Loss | Epoch | Step | Validation Loss | Wer |
54
  |:-------------:|:-----:|:----:|:---------------:|:------:|
55
+ | 3.1674 | 8.33 | 500 | 3.0295 | 1.0 |
56
+ | 2.6987 | 16.66 | 1000 | 2.6878 | 1.0 |
57
+ | 1.3454 | 24.99 | 1500 | 0.6814 | 0.6981 |
58
+ | 1.1227 | 33.33 | 2000 | 0.5791 | 0.6513 |
59
+ | 0.9972 | 41.66 | 2500 | 0.5235 | 0.5718 |
60
+ | 0.9123 | 49.99 | 3000 | 0.5104 | 0.5633 |
61
+ | 0.836 | 58.33 | 3500 | 0.4927 | 0.5580 |
62
+ | 0.7725 | 66.66 | 4000 | 0.5078 | 0.5779 |
63
+ | 0.7297 | 74.99 | 4500 | 0.4939 | 0.5737 |
64
 
65
 
66
  ### Framework versions