search

gitmylo / audio-webui

A webui for different audio related Neural Networks
MIT License
1.03k stars 99 forks source link

[BUG REPORT] Exception: "reflection_pad1d" not implemented for 'Short' #196

Closed robonxt closed 9 months ago

robonxt commented 9 months ago

Describe the bug Been getting errors when I try to use whisper on a audio file converted with Audacity to 16000hz, signed 16 PCM, WAV format: Exception: "reflection_pad1d" not implemented for 'Short'

To Reproduce Steps to reproduce the behavior:

  1. Use Audacity to convert a english memo mp3 file into wav with 16000hz, signed 16 PCM, WAV format
  2. Upload to Audio WebUI
  3. Load any whisper model (I used base)
  4. Click Transcribe button
  5. See error

Expected behavior whisper should be transcribing my audio.

Additional context I tried it on the OG whisper.cpp and it works normally.

gitmylo commented 9 months ago

Try uploading the file into the batch section, gradio does some conversions before I get access to the audio, which are kind of unpredictable.

robonxt commented 9 months ago

Hmm, I've found that by accident. Guess it's gradio issue. Will close this.