Closed linuxandchill closed 2 months ago
This is not referring to the training process, but an error in thecheckpoint loading process. Which checkpoints are you using, ch-hallo or joyhallo?
This is not referring to the training process, but an error in thecheckpoint loading process. Which checkpoints are you using, ch-hallo or joyhallo?
using joyhallo, downloaded from huggingface:
wget https://huggingface.co/spaces/jdh-algo/JoyHallo-v1/resolve/main/pretrained_models/joyhallo/net.pth
This is not referring to the training process, but an error in thecheckpoint loading process. Which checkpoints are you using, ch-hallo or joyhallo?
using joyhallo, downloaded from huggingface:
wget https://huggingface.co/spaces/jdh-algo/JoyHallo-v1/resolve/main/pretrained_models/joyhallo/net.pth
Can you provide the complete log?
i have been able to resolve this, seems that i didn't have the wav2vec model
am having another issues with CUDA but i will try to resolve. maybe a Docker container would help to include ?
Glad to hear that. Thanks for your suggestion, we will consider adding the Docker container.
when running
sh joyhallo-infer.sh
it appears that a training process is kicked off09/16/2024 19:31:39 - ERROR - root - Failed to execute the training process: Error(s) in loading state_dict for Net: