search

reriiasu / speech-to-text

Real-time transcription using faster-whisper
MIT License
375 stars 56 forks source link

About runtime errors #8

Open tsushimahiroaki opened 10 months ago

tsushimahiroaki commented 10 months ago

nice to meet you. When I tried using the tool, the following error appeared on the UI. What should I do? Thank you.

[Error number 2] No such file or directory: 'C:\Users\username\.cache\huggingface\hub\models--guillaumekln--faster-whisper-medium\refs\ \main'

reriiasu commented 10 months ago

@tsushimahiroaki hello tsushimahiroaki

The medium model may not have been downloaded. Please check if the folder with the error exists. If it exists, try deleting it.

tsushimahiroaki commented 10 months ago

thanks.

An error occured while synchronizing the model guillaumekln/faster-whisper-medium from the Hugging Face Hub: HTTPSConnectionPool(host='huggingface.co', port=443): Max retries exceeded with url: /api/models/guillaumekln/faster-whisper-medium/revision/main (Caused by SSLError(SSLCertVerificationError(1, '[SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: unable to get local issuer certificate (_ssl.c:1007)'))) Trying to load the model directly from the local cache, if it exists. index.html was closed

I am getting the above error. What should I do now.

reriiasu commented 10 months ago

You can manually deploy the model instead of using automatic model download. Please refer to the next step.

  1. Manually download the following file from https://huggingface.co/guillaumekln/faster-whisper-medium. .gitattributes README.md config.json model.bin tokenizer.json vocabulary.txt

  2. Move the downloaded files to the speech-to-text\local_model directory.

  3. Change parameters from the UI. Set Model Settings > Model Size or Path to local_model.