Fine-tuned Model Versions

[ddlinh]

Dear Khai,

I have accessed the ASR models published at https://huggingface.co/leduckhai/MultiMed-ST, including 'whisper-small-vietnamese' and 'whisper-small-multilingual', to generate transcripts for the audio files. However, due to the complex mix of voices in the recordings, these versions haven't produced satisfactory transcripts.

I noticed in your paper that you mentioned other fine-tuned versions based on whisper-medium and whisper-large. I would be very interested in trying out these models for my use case.

Would it be possible for me to access them? Thank you very much.