Update README.md
Browse files
README.md
CHANGED
|
@@ -27,10 +27,10 @@ model-index:
|
|
| 27 |
metrics:
|
| 28 |
- name: Test WER
|
| 29 |
type: wer
|
| 30 |
-
value:
|
| 31 |
- name: Test CER
|
| 32 |
type: cer
|
| 33 |
-
value: 1.
|
| 34 |
- task:
|
| 35 |
name: Speech Recognition
|
| 36 |
type: automatic-speech-recognition
|
|
@@ -41,10 +41,52 @@ model-index:
|
|
| 41 |
metrics:
|
| 42 |
- name: Test WER
|
| 43 |
type: wer
|
| 44 |
-
value: 15.
|
| 45 |
- name: Test CER
|
| 46 |
type: cer
|
| 47 |
-
value: 4.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 48 |
---
|
| 49 |
|
| 50 |
# Wav2Vec2-Large-Ru-Golos
|
|
@@ -146,13 +188,15 @@ print("Character error rate on the Farfield domain:", farfield_cer)
|
|
| 146 |
|
| 147 |
| "crowd" | "farfield" |
|
| 148 |
|---------|------------|
|
| 149 |
-
|
|
| 150 |
|
| 151 |
*Result (CER, %)*:
|
| 152 |
|
| 153 |
| "crowd" | "farfield" |
|
| 154 |
|---------|------------|
|
| 155 |
-
| 1.
|
|
|
|
|
|
|
| 156 |
|
| 157 |
## Citation
|
| 158 |
If you want to cite this model you can use this:
|
|
|
|
| 27 |
metrics:
|
| 28 |
- name: Test WER
|
| 29 |
type: wer
|
| 30 |
+
value: 6.271
|
| 31 |
- name: Test CER
|
| 32 |
type: cer
|
| 33 |
+
value: 1.292
|
| 34 |
- task:
|
| 35 |
name: Speech Recognition
|
| 36 |
type: automatic-speech-recognition
|
|
|
|
| 41 |
metrics:
|
| 42 |
- name: Test WER
|
| 43 |
type: wer
|
| 44 |
+
value: 15.653
|
| 45 |
- name: Test CER
|
| 46 |
type: cer
|
| 47 |
+
value: 4.349
|
| 48 |
+
- task:
|
| 49 |
+
name: Automatic Speech Recognition
|
| 50 |
+
type: automatic-speech-recognition
|
| 51 |
+
dataset:
|
| 52 |
+
name: Common Voice ru
|
| 53 |
+
type: common_voice
|
| 54 |
+
args: ru
|
| 55 |
+
metrics:
|
| 56 |
+
- name: Test WER
|
| 57 |
+
type: wer
|
| 58 |
+
value: 23.897
|
| 59 |
+
- name: Test CER
|
| 60 |
+
type: cer
|
| 61 |
+
value: 4.919
|
| 62 |
+
- task:
|
| 63 |
+
name: Automatic Speech Recognition
|
| 64 |
+
type: automatic-speech-recognition
|
| 65 |
+
dataset:
|
| 66 |
+
name: Sova RuDevices
|
| 67 |
+
type: bond005/sova_rudevices
|
| 68 |
+
args: ru
|
| 69 |
+
metrics:
|
| 70 |
+
- name: Test WER
|
| 71 |
+
type: wer
|
| 72 |
+
value: 41.822
|
| 73 |
+
- name: Test CER
|
| 74 |
+
type: cer
|
| 75 |
+
value: 13.624
|
| 76 |
+
- task:
|
| 77 |
+
name: Automatic Speech Recognition
|
| 78 |
+
type: automatic-speech-recognition
|
| 79 |
+
dataset:
|
| 80 |
+
name: Russian Librispeech
|
| 81 |
+
type: bond005/rulibrispeech
|
| 82 |
+
args: ru
|
| 83 |
+
metrics:
|
| 84 |
+
- name: Test WER
|
| 85 |
+
type: wer
|
| 86 |
+
value: 29.664
|
| 87 |
+
- name: Test CER
|
| 88 |
+
type: cer
|
| 89 |
+
value: 5.747
|
| 90 |
---
|
| 91 |
|
| 92 |
# Wav2Vec2-Large-Ru-Golos
|
|
|
|
| 188 |
|
| 189 |
| "crowd" | "farfield" |
|
| 190 |
|---------|------------|
|
| 191 |
+
| 6.271 | 15.653 |
|
| 192 |
|
| 193 |
*Result (CER, %)*:
|
| 194 |
|
| 195 |
| "crowd" | "farfield" |
|
| 196 |
|---------|------------|
|
| 197 |
+
| 1.292 | 4.349 |
|
| 198 |
+
|
| 199 |
+
You can see the evaluation script on other datasets, including Russian Librispeech and SOVA RuDevices, on my Kaggle web-page https://www.kaggle.com/code/bond005/wav2vec2-ru-eval
|
| 200 |
|
| 201 |
## Citation
|
| 202 |
If you want to cite this model you can use this:
|