OthmaneJ
commited on
Commit
•
ff5b141
1
Parent(s):
350798d
update readme
Browse files
README
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language: en
|
| 3 |
+
datasets:
|
| 4 |
+
- librispeech_asr
|
| 5 |
+
tags:
|
| 6 |
+
- speech
|
| 7 |
+
- audio
|
| 8 |
+
- automatic-speech-recognition
|
| 9 |
+
license: apache-2.0
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Distil-wav2vec2
|
| 13 |
+
This model is a distilled version of the wav2vec2 model (https://arxiv.org/pdf/2006.11477.pdf). This model is 4 times smaller and 3 times faster than the original wav2vec2 large model.
|
| 14 |
+
|
| 15 |
+
# Evaluation results
|
| 16 |
+
When used with a light tri-gram language model head, this model achieves the following results :
|
| 17 |
+
| Dataset | WER |
|
| 18 |
+
| ------------- |:-------------:|
|
| 19 |
+
| Librispeech-clean| 12.7%|
|
| 20 |
+
|
| 21 |
+
#Usage
|
| 22 |
+
notebook (google colab) at https://github.com/OthmaneJ/distil-wav2vec2
|