4 ms·

> and Qwen3-ASR Is the ASR inference engine open source as well?

• nshm 19d ago
  Yes, and it is very good one. Leading position on private leaderboard on HF: https://huggingface.co/spaces/hf-audio/open_asr_leaderboard https://huggingface.co/spaces/hf-audio/open_asr_leaderboard
  • meatmanek 19d ago
    I meant the Nari inference engine for Qwen3-ASR. I'm aware that Qwen3-ASR is open source, but I don't see a repo under https://github.com/nari-labs https://github.com/nari-labs for nari-qwen3-asr or similar. The Huggingface link on https://narilabs.com/product/stt/ https://narilabs.com/product/stt/ links to https://huggingface.co/Qwen/Qwen3-ASR-1.7B https://huggingface.co/Qwen/Qwen3-ASR-1.7B , not anything under https://huggingface.co/nari-labs https://huggingface.co/nari-labs
    • toebee 19d ago
      the qwen3-asr inference repo is not OSSed as of now. we're planning to write a paper or tech report on it as it contains some general techniques for ASR inference.
      • kshmir 19d ago
        how do I follow you? I have a small 5090 doing inference all the time and I barely use tts but a lot of asr, mostly whisper, I ported your tech report for tts and implemented some improvements on my whisper inference based on your tech report as well! would love to talk sometime!
• verdverm 19d ago
  They have a number of demos and examples in their HF space https://huggingface.co/Qwen/spaces https://huggingface.co/Qwen/spaces I saw a local-ai demo (something + gemma), where the person used ASR to get text and gemma to clean it up (like turning "question mark" into a literal "?", bullet points another one). The presenter also showed a gemma only option, that did both in one go, but had a higher WER on average, and even though the formatting statements were handled without a multi-stage pipeline, they preferred the multi-stage overall