Error

#1
by Gatozu35 - opened

On every input I try, it outputs an error. running the code locally gives me an EOFError: Ran out of input

SaruLab Speech group org

Hi, thank you for trying our demo.
We tried to reproduce the error, but the demo worked on the inputs we tried, which included different sampling rates and file formats (.wav, .raw, .flac, .aiff, .au).
It worked on both hugging face and local environment.
So can you tell us details on your inputs?

I found the problem. The inputs I was using were too long (30-60s). It works fine on <10s files.

The EOFError was from a corrupted checkpoint on my side, it doesn't output the same error if I redownload it .

Also the huggingface interface errors out on the ~30s file but not predict.py locally

Gatozu35 changed discussion status to closed

Sign up or log in to comment