You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am using this prompt and audio files:
prompt: soundstorm-speechtokenizer/samples/Voice Conversion/2078-142845_6345-64257/prompt.wav
audio: soundstorm-speechtokenizer/samples/Voice Conversion/2078-142845_6345-64257/raw.wav
error
rearrange(generated, 'n q -> q b n', b=semantic_tokens.size(0))
EinopsError: Error while processing rearrange-reduction pattern "n q -> q b n".
Input tensor shape: torch.Size([1, 518, 8]). Additional info: {'b': 1}.
Identifiers only on one side of expression (should be on both): {'b'}
The text was updated successfully, but these errors were encountered:
Hello @ZhangXInFD
There is an error running inference, due to shape error in einops
semantic tokens shape:
[1, 518]
prompt tokens shape
[1, 142, 8]
generated shape:
[1, 518, 8]
I am using this prompt and audio files:
prompt:
soundstorm-speechtokenizer/samples/Voice Conversion/2078-142845_6345-64257/prompt.wav
audio:
soundstorm-speechtokenizer/samples/Voice Conversion/2078-142845_6345-64257/raw.wav
error
The text was updated successfully, but these errors were encountered: