lsfhuihuiff / MusicTI_AAAI2024

" Music Style Transfer with Time-Varying Inversion of Diffusion Models"
MIT License
35 stars 4 forks source link

How do I load up from a previously logged checkpoint? #9

Open Pinak1392 opened 1 month ago

Pinak1392 commented 1 month ago

After a couple of epochs of training, the model has saved a number of checkpoints. I want to resume the training from the latest checkpoint. But when I input the actual_resume parameter with the saved ckpt path, I get an error saying that the key state_dict doesn't exist. When I checked the outputs of the torch.load in load model from config, I found it to be an empty dictionary. Is there a different parameter I need to use to resume training?