Plim commited on
Commit
a010dda
1 Parent(s): 95addc4

update README.md

Browse files
Files changed (1) hide show
  1. README.md +47 -21
README.md CHANGED
@@ -1,36 +1,48 @@
1
  ---
2
  language:
3
  - fr
 
4
  tags:
5
  - automatic-speech-recognition
6
  - mozilla-foundation/common_voice_8_0
7
  - generated_from_trainer
 
8
  model-index:
9
- - name: ''
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
- #
17
-
18
- This model is a fine-tuned version of [./checkpoint-13000](https://huggingface.co/./checkpoint-13000) on the MOZILLA-FOUNDATION/COMMON_VOICE_8_0 - FR dataset.
19
- It achieves the following results on the evaluation set:
20
- - Loss: inf
21
- - Wer: 0.2937
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
 
@@ -74,6 +86,10 @@ The following hyperparameters were used during training:
74
  | 0.8022 | 5.45 | 19000 | inf | 0.1895 |
75
  | 0.792 | 5.73 | 20000 | inf | 0.1854 |
76
 
 
 
 
 
77
 
78
  ### Framework versions
79
 
@@ -81,3 +97,13 @@ The following hyperparameters were used during training:
81
  - Pytorch 1.10.2+cu102
82
  - Datasets 1.18.3.dev0
83
  - Tokenizers 0.11.0
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  language:
3
  - fr
4
+ license: apache-2.0
5
  tags:
6
  - automatic-speech-recognition
7
  - mozilla-foundation/common_voice_8_0
8
  - generated_from_trainer
9
+ - robust-speech-event
10
  model-index:
11
+ - name: XLS-R-1B - French
12
+ results:
13
+ - task:
14
+ name: Automatic Speech Recognition
15
+ type: automatic-speech-recognition
16
+ dataset:
17
+ name: Common Voice 8
18
+ type: mozilla-foundation/common_voice_8_0
19
+ args: fr
20
+ metrics:
21
+ - name: Test WER
22
+ type: wer
23
+ value: 18.33
24
+ - name: Test CER
25
+ type: cer
26
+ value: 5.60
27
+ - task:
28
+ name: Automatic Speech Recognition
29
+ type: automatic-speech-recognition
30
+ dataset:
31
+ name: Robust Speech Event - Dev Data
32
+ type: speech-recognition-community-v2/dev_data
33
+ args: fr
34
+ metrics:
35
+ - name: Test WER
36
+ type: wer
37
+ value: 60.25
38
+ - name: Test CER
39
+ type: cer
40
+ value: 15.68
41
  ---
42
 
 
 
 
 
 
 
 
 
 
 
43
  ## Model description
44
 
45
+ This model is a fine-tuned version of [facebook/wav2vec2-xls-r-1b](https://huggingface.co/facebook/wav2vec2-xls-r-1b) on the MOZILLA-FOUNDATION/COMMON_VOICE_8_0 - FR dataset.
 
 
 
 
 
 
 
 
46
 
47
  ## Training procedure
48
 
 
86
  | 0.8022 | 5.45 | 19000 | inf | 0.1895 |
87
  | 0.792 | 5.73 | 20000 | inf | 0.1854 |
88
 
89
+ It achieves the best result on the validation set on STEP 13000:
90
+ - Wer: 0.1834
91
+
92
+ Some problem occurs when calculating the validation loss.
93
 
94
  ### Framework versions
95
 
 
97
  - Pytorch 1.10.2+cu102
98
  - Datasets 1.18.3.dev0
99
  - Tokenizers 0.11.0
100
+
101
+ ### Evaluation Commands
102
+ 1. To evaluate on `mozilla-foundation/common_voice_8` with split `test`
103
+ ```bash
104
+ python eval.py --model_id Plim/xls-r-1b-cv_8-fr --dataset mozilla-foundation/common_voice_8_0 --config fr --split test
105
+ ```
106
+ 2. To evaluate on `speech-recognition-community-v2/dev_data`
107
+ ```bash
108
+ python eval.py --model_id Plim/xls-r-1b-cv_8-fr --dataset speech-recognition-community-v2/dev_data --config fr --split validation --chunk_length_s 5.0 --stride_length_s 1.0
109
+ ```