glow-tts icon indicating copy to clipboard operation
glow-tts copied to clipboard

Runtime Error: Multi speaker

Open ppanja opened this issue 3 years ago • 1 comments

Hi @jaywalnut310 , I've the following configuration for multi-speaker and I am getting the below error. I've done all other necessary changes as highlighted here in different posts like changing to TextMelSpeakerLoader, TextMelSpeakerCollate and correcting the unpacking etc.

RuntimeError: Given groups=1, weight of size 256 512 3, expected input[1, 256, 89] to have 512 channels, but got 256 channels instead

Multi Speaker Configuration:

"batch_size": 64 "hidden_channels": 256, "filter_channels": 1024, "filter_channels_dp": 256, "n_speakers": 24, "gin_channels": 256 "hidden_channels_enc": 256, "hidden_channels_dec": 256

Training records - 9102, Val records - 100

Can you please help in solving this? Please let me know if you suggest any other configuration.

Thanks in advance.

Regards, Prasanta

ppanja avatar Jun 13 '21 20:06 ppanja

I have made necessary changes to run in multi-speaker setting https://github.com/ankurdhuriya/glow-tts

ankurdhuriya avatar Jan 21 '22 13:01 ankurdhuriya