ae_accomp.yaml missing mandatory lossconfig and other files absent when running from README

[Molanischen]

Hi，yu
Thanks for the great work！

Do you plan a code refresh in the near future to bring everything in sync?

Currently, when I try to launch LDM training from a fresh clone, I run into two blockers:

1. Missing source files
   joinaudiodataset_anylen.py is missing but required in file main.py. I found some files with same name in https://huggingface.co/spaces/AIGC-Audio/AudioLCM/tree/48feb046bd44a3f5ae913ce6b885be611fd1a8c4/ldm/data

2. Incomplete YAML configs
   The autoencoder configs (for example ae_accomp.yaml) omit the mandatory lossconfig section, so AutoencoderKL(...) raises
   TypeError: missing 1 required positional argument: 'lossconfig'

Could you please let us know if an updated release is on the way?

Thanks!

[AaronZ345]

Solved!

[Molanischen]

Solved!

Thanks， but file vocal2accomp_dataset is still missing！ TT

[AaronZ345]

Solved!

Thanks， but file vocal2accomp_dataset is still missing！ TT

I edit main.py so there is no need to import vocal2accomp_musical_drop.

[Dinxin]

Will you release the pre-trained checkpoints of other parts such as VocalBand and LyricBand?