Xingchen Song(宋星辰)

please try to repeat 10 times

This might be a warmup issue

>  > > 2022-03-22 10:40:24,635 INFO 20041028_124231_A013309_B013310-B-058899-059281 呃我觉得挺好的 2022-03-22 10:40:24,792 INFO 20041028_124231_A013309_B013310-B-059611-059928 巴塞罗纳我一开始我就不喜欢巴西 Traceback (most recent call last): File "tools/compute-cer.py", line 414, in default_cluster_name = default_cluster(word) File "tools/compute-cer.py", line...

> yes

Will hold this PR for anyone else who has encountered and interested in the same question.

Now it is set to 0.0005 and 0.1 for librispeech **large** : https://github.com/wenet-e2e/wenet/blob/main/examples/librispeech/s0/conf/train_conformer_bidecoder_large.yaml#L47-L48

> For binding：there is 3 questions： > > 1 get model by language type， so if we can supply small and big model for each language？（the small model could be...

Discussed it offline, close this issue

seems solved， close this issue

Solved, If u use char (rather then bpe) as modeling unit for english dataset, remember to add `split_with_space = True` in configuration.