But I get an error message: assert(torch.min(y.data) >= 1)
In audio_proceessing.py line 128
I printed the data:
First this is the output from audio tensor in data.py line 150:
Audio: tensor([ -3639553., -9467905., 5345536., ..., -11249665., -11092481.,
-11744769.])
And this is the output in audio_processing.py line 128 in Mel_spectrogramm
min y.data: tensor(-12156.6719)
How can I get rid of this error? Can somebody give me an advice?
Another Question: I have 30 min of own recording. Is this enough?
And is there any German dictionary for "cmudict_path": "data/cmudict_dictionary",
Hi
I try to fine tune a model with my own data.
python train.py -c config.json -p train_config.ignore_layers=["speaker_embedding.weight"] train_config.warmstart_checkpoint_path="models/flowtron_ljs.pt"
But I get an error message: assert(torch.min(y.data) >= 1)
In audio_proceessing.py line 128 I printed the data: First this is the output from audio tensor in data.py line 150: Audio: tensor([ -3639553., -9467905., 5345536., ..., -11249665., -11092481., -11744769.]) And this is the output in audio_processing.py line 128 in Mel_spectrogramm min y.data: tensor(-12156.6719)
How can I get rid of this error? Can somebody give me an advice?
Another Question: I have 30 min of own recording. Is this enough? And is there any German dictionary for "cmudict_path": "data/cmudict_dictionary",
BR and thanks.