Open csukuangfj opened 4 months ago
I can contribute a recipe for a streaming model for one of the languages. Do you need it?
I can contribute a recipe for a streaming model for one of the languages. Do you need it?
Yes, definitely we need it. Thank you!
MSR-86K: An Evolving, Multilingual Corpus with 86,300 Hours of Transcribed Audio for Speech Recognition Research The above paper has just open-sourced a dataset for 15 languages and is available at https://huggingface.co/datasets/Alex-Song/MSR-86K
It would be great if someone could train a (streaming or/and a non-streaming) zipformer model with it.