Prabhsimran Singh

@shubhamagarwal92 thanks for pointing this out. I'll check their implementation and see what's different. Working on the test to make it more deterministic. Will test more beam sizes too.

Hey @Mehrad0711 @shubhamagarwal92 sorry haven't gotten the time to work on this yet. You're welcome to submit a PR :)

@AtmaHou we'll see if it can be integrated in the next release. You're welcome to submit a PR. We need all the help we can get.

@varunAI Yup working on it. Will push to develop as soon as it's done.

@Diego999 no, I'm facing some issues with dimensions as mentioned and there seems to be a problem with the dynamic dictionary too. I'll get to this asap. Will post any...

@Decalogue You could pull `develop` and try implementing your script again. `Seq2seqDataset` has been introduced after #163. However things might be a little unstable as `develop` is pushed to weekly...

@enferas thanks for pointing this out.

@jhyun0919 it'll be fixed in next alpha release. Please see #161.

### Task 1 Write an interface and implement GPU batch decoding for [Kaldi ASR](https://github.com/kaldi-asr/kaldi) models in the kaldi-serve [core C++ library](https://github.com/skit-ai/kaldi-serve/tree/master/src). The [current partial version](https://github.com/skit-ai/kaldi-serve/blob/gpu-decoder/src/decoder/decoder-batch.cpp) (`gpu-decoder` branch) is buggy (stale...

Hi, sorry we were busy with the move to library v1 release. I'd be happy to test and merge this if you can refactor (or reopen) the PR based on...