Closed DogeLord081 closed 9 months ago
I got exact same issue with Colab. any idea now how the issue could be fixed?
I got exact same issue with Colab. any idea now how the issue could be fixed?
Yep, I found the solution, change line 19 in scripts/short_audio_transcribe.py to:
mel = whisper.log_mel_spectrogram(audio, n_mels=128).to(model.device)
I also had the same problem, but after modifying it according to the answer you gave, it was not optimized. It seems to prompt me about a CUDA problem, but I am connected to T4, so I still have a lot of doubts. Can you help me figure out how to solve this problem?
My folder structure is:
custom_character_voice ├───Megumin_1 ├ ├───Megumin_1.wav ├ ├───Megumin_2.wav ├ ├───... ├ └───Megumin_1324.wav
But on step 3 of the colab, I keep getting this error: