AudioLDM2 icon indicating copy to clipboard operation
AudioLDM2 copied to clipboard

Unable to run app.py

Open MLTQ opened this issue 1 year ago • 3 comments

After installing dependencies and run app.py I get it to download the default model and then: RuntimeError: Error(s) in loading state_dict for LatentDiffusion: Unexpected key(s) in state_dict: "clap.model.text_branch.embeddings.position_ids", "cond_stage_models.0.cond_stage_models.0.model.text_branch.embeddings.position_ids".

Any idea why this happens and how to resolve?

MLTQ avatar Aug 07 '23 21:08 MLTQ

Same error cmd_W5NDjFtNfF

ohmygoobness avatar Aug 08 '23 00:08 ohmygoobness

i did pip install git+https://github.com/haoheliu/AudioLDM2.git and then run app.py, it solves this but I saw 15.7GB VRAM used before it crashes. Crash was fixed #4

ohmygoobness avatar Aug 08 '23 00:08 ohmygoobness

RuntimeError: Error(s) in loading state_dict for LatentDiffusion: Unexpected key(s) in state_dict: "clap.model.text_branch.embeddings.position_ids", "cond_stage_models.0.model.text_branch.embeddings.position_ids".

ucas010 avatar Aug 12 '24 10:08 ucas010