hrbigelow / ae-wavenet Goto Github PK

Hi @hrbigelow

I am working on CPU+GPU .I got error while executing this command line

python train.py new -af par/arch.basic.json -tf par/train.basic.json -nb 4 -si 1000
-vqn 1000 $run_dir/model%.ckpt $run_dir/librispeech.dev-clean.dat $run_dir/data_slices.dat

"KeyError: '/root/data_mtbox_d/VQ_VAE/VQ_VAE_Speech/ae_wavenet/model%.ckpt'". Can you help me to resolve it

Hi,

Can you please point me to use this repository for audio feature extraction ?
Does this give a fixed representation for variable length audio signal?

I tried to train it, with using command you mentioned

python train.py new -af par/arch.basic.json -tf par/train.basic.json -nb 4 -si 1000 -vqn 1000 $run_dir/model%.ckpt $run_dir/librispeech.dev-clean.dat $run_dir/data_slices.dat
However, It causes train.py new: error: unrecognized arguments: data_slices.dat error.
there is no data_slices.dat file in here.

I just deleted that file and tried to execute it. but I found out another error mentioned in 2

Hi, nice work you've done! I was wondering if you had any visualisation of trained vq embedding space? I've been trying to reproduce the paper as well, but the model seems not to use the global condition. It seems the wavenet is too powerful to pick up local / global condition, and only uses the original wav audio. Is that what you are facing with the codebook collapse issue as well?

Hi,
Thanks for the great repo. I wanted to use the embeddings, and so wanted to use a pretrained model to get the embeddings. Could you please update the progress on the model? Thanks!