File size: 3,447 Bytes
c30fa45
748177b
 
 
c30fa45
 
748177b
 
c30fa45
748177b
 
 
 
c30fa45
70510e9
748177b
 
 
1db58e2
748177b
5d98836
8268222
bc244ef
748177b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c30fa45
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
---
language: 
- ar

license: apache-2.0
tags:
- automatic-speech-recognition
- robust-speech-event
datasets:
- mozilla-foundation/common_voice_7_0
metrics:
- wer
- cer
model-index:
- name: wav2vec2-xls-r-300m-arabic
  results:
  - task: 
      type: automatic-speech-recognition  # Required. Example: automatic-speech-recognition
      name: Speech Recognition  # Optional. Example: Speech Recognition
    dataset:
      type: mozilla-foundation/common_voice_7_0  # Required. Example: common_voice. Use dataset id from https://hf.co/datasets
      name: Common Voice ar # Required. Example: Common Voice zh-CN
      args: ar        # Optional. Example: zh-CN
    metrics:
      - type: wer    # Required. Example: wer
        value: 31.05  # Required. Example: 20.90
        name: Test WER    # Optional. Example: Test WER
        args: 
        - learning_rate: 0.0003
        - train_batch_size: 64
        - eval_batch_size: 8
        - seed: 42
        - gradient_accumulation_steps: 2
        - total_train_batch_size: 128
        - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
        - lr_scheduler_type: linear
        - lr_scheduler_warmup_steps: 1000
        - num_epochs: 10
        - mixed_precision_training: Native AMP         # Optional. Example for BLEU: max_order
      - type: cer    # Required. Example: wer
        value: 8.78  # Required. Example: 20.90
        name: Test CER    # Optional. Example: Test WER
        args: 
        - learning_rate: 0.0003
        - train_batch_size: 64
        - eval_batch_size: 8
        - seed: 42
        - gradient_accumulation_steps: 2
        - total_train_batch_size: 128
        - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
        - lr_scheduler_type: linear
        - lr_scheduler_warmup_steps: 1000
        - num_epochs: 10
        - mixed_precision_training: Native AMP         # Optional. Example for BLEU: max_order
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# wav2vec2-large-xlsr-300-arabic

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.
It achieves the following results on the evaluation set:
- Loss: 0.3384
- Wer: 0.3105
- Cer: 0.0879

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 0.0003
- train_batch_size: 64
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- num_epochs: 10
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Wer    | Cer    |
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|
| 0.7383        | 1.8   | 500  | 0.4292          | 0.4065 | 0.1189 |
| 0.664         | 3.6   | 1000 | 0.4245          | 0.3978 | 0.1175 |
| 0.6064        | 5.4   | 1500 | 0.3854          | 0.3625 | 0.1048 |
| 0.5221        | 7.19  | 2000 | 0.3819          | 0.3400 | 0.0976 |
| 0.4591        | 8.99  | 2500 | 0.3384          | 0.3105 | 0.0879 |


### Framework versions

- Transformers 4.16.0.dev0
- Pytorch 1.10.1+cu102
- Datasets 1.17.1.dev0
- Tokenizers 0.11.0