Soshyant
Soshyant
yeah this would be great to have
there's an error that prevents your code from doing its job. take a tiny chunk of your data and try to use your phonemizer directly like the code below, it'll...
I ended up changing the dataset this time to more than 90M sentences, though I still want to know what was the problem. Also, is it natural to get such...
i'm very interested in this. but can't find any scripts in the 'examples'. enc-dec models have huge untapped potential.
yeah, i also get this, very curios about
I've not tried training this model, but enable cpu training so you can actually find out what's causing the problem when facing the cuda device-side asset error. good luck
you don't have to do it the way the author pre-processed their dataset. just use the regular .map() and set num_proc to whatever your cpu can handle
I think there's nothing wrong with the code itself and it's working as intended. the purpose of that line is probably not to take the biggest sample in the batch...
Hi. ```python def load_F0_models(path): F0_model = JDCNet(num_class=1, seq_len=192) params = torch.load(path, map_location='cpu') F0_model.load_state_dict(params) _ = F0_model.eval() return F0_model.to('cuda') F0_path = "path_to_your_model" pitch_extractor = load_F0_models(F0_path) to_mel = torchaudio.transforms.MelSpectrogram( n_mels=80, n_fft=2048, win_length=1200,...
load the JDCnet class from one of the StyleTTS repos, not the one provided here. ```RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 80 but got...