NVIDIA/flowtron

TypeError: guvectorize() missing 1 required positional argument 'signature'

nafonelli opened this issue · 1 comments

When training from scratch using: python3 train.py -c config.json -p train_config.output_directory=outdir data_config.use_attn_prior=1

I get the following error: TypeError: guvectorize() missing 1 required positional argument 'signature'

I'm using the LJSpeech dataset, the traceback starts in 'train.py' on line 27 which appears to be when it imports data and datacollate

Apologies if this is trivial, I'm new to this!

Many thanks for any suggestions

Solved like this:

pip install resampy==0.3.1