Hugging Face's logo

shaojieli
/
icefall-asr-commonvoice-fr-pruned-transducer-stateless7-streaming-2023-04-02

Automatic Speech Recognition
TensorBoard
French
Model card Files Metrics Community
1
shaojieli commited on
Commit
91ee85c
•
1 Parent(s): 6c88444

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +10 -0
README.md CHANGED
@@ -29,3 +29,13 @@ The WERs are:
29
  | modified beam search | 640ms | 10.19 | --epoch 29 --avg 9 | simulated streaming |
30
  | fast beam search | 640ms | 10.25 | --epoch 29 --avg 9 | simulated streaming |
31
 
 
 
 
 
 
 
 
 
 
 
 
29
  | modified beam search | 640ms | 10.19 | --epoch 29 --avg 9 | simulated streaming |
30
  | fast beam search | 640ms | 10.25 | --epoch 29 --avg 9 | simulated streaming |
31
 
32
+ training on full librispeech and gigaspeech then finetune on full commonvoice
33
+
34
+ The WERs are:
35
+
36
+ | decoding method | chunk size | test | comment | decoding mode |
37
+ | -------------------- | ---------- | ------ | -------------------- | -------------------- |
38
+ | greedy search | 640ms | 9.95 | --epoch 30 --avg 9 | simulated streaming |
39
+ | modified beam search | 640ms | 9.57 | --epoch 30 --avg 9 | simulated streaming |
40
+ | fast beam search | 640ms | 9.67 | --epoch 30 --avg 9 | simulated streaming |
41
+