How to do ASR with this repo

Hi, I download the fairseq checkpoint from readme.

The inference/beam_search.py is for asr I think, so I trying to running the bash script/inference/evaluate_multimodel.sh, but missing some file MANIFEST_PATH (some dictionary in it?)

Would like to share the MANIFEST file?

Thank you.

Hi, thank you for your interest.

"MANIFEST_PATH" is a folder that has audio paths and transcriptions of audios. As described in Readme, you have to download a dataset and make a manifest file.

Please understand that I cannot share "MANIFEST file" as copyrights related to the dataset do not belong to me.

Please contact us if you have any problems even after following the guidelines. Thanks.

I see. the audio files and transcriptions can not share because of copyrights. How about the dictionary files? When dictionary files (e.g. dict.ltr.txt, alphabet.txt ) are available, we can run the inference code alone.