Shinji Watanabe

We don't provide such functions, but such information is stored in `nbest_hyps` in https://github.com/espnet/espnet/blob/master/espnet/nets/beam_search.py#L329-L337. Therefore, if you modify the output format (e.g., https://github.com/espnet/espnet/blob/master/espnet2/bin/asr_inference.py#L535), you can obtain such results.

Thanks a lot for the report. This is an annoying issue. It would be great if someone help to fix the issue. Actually, I also faced the same issue and...

Right. I reran the CI.

@simpleoier, could you review this PR?

Can you reflect reviewer's comments?

Thanks. We will make this request higher if we have more requests.

We do not have so many requests, actually. We also have some internal discussions, but there are a lot of alternatives for the faster (lightweight) encoder and Squeezeformer does not...

But again, if we have more requests, we'll definitely consider it.

@espnetUser, can you check this? It looks correct to me.

@snavali567, can you reflect @ftshijt's comments?