currently there is support for AMD GPUs in generating voice covers of audio but doesn't work for training a model under rocm on AMD cards only cuda on NVIDIA cards. if better support was added for AMD cards to train models and generate audio that would be extremely help to people with AMD RX 6000/7000 GPUs. making support for these cards would also give people more options for AI work outside of NVIDIA cards being the only route for Generative AI. I just ask that you consider improving support for AMD RX 6000/7000 Series Video cards. thanks Ozzy