Im refering to this guide https://www.reddit.com/r/StableDiffusion/comments/ww436j/howto_stable_diffusion_on_an_amd_gpu/ I have a problem getting auto1111 RocM i got the error :0:rocdevice.cpp :2673: 3043409144 us: 51211: [tid:0x7f6b933ff640] Device::callbackQueue aborting with error : HSA_STATUS_ERROR_MEMORY_APERTURE_VIOLATION: The agent attempted to access memory beyond the largest legal address. code: 0x29 or Memory access fault by GPU node-1 (Agent handle: 0x55ea0ebd7e00) on address 0x7ff6c9374000. Reason: Page not present or supervisor privilege. Aborted (core dumped) I follow this other guide but the black out made the tips to get it functional inaccesible https://phazertech.com/tutorials/rocm.html my card is a rx5700 xt and I have the HSA_OVERRIDE_GFX_VERSION=10.3.0 excally like in the guide

  • ottensio@lemmy.dbzer0.comOP
    link
    fedilink
    English
    arrow-up
    3
    ·
    1 year ago

    I made it work using an old version of pytorch and rocm TORCH_COMMAND=“pip install torch==1.13.0+rocm5.2 torchvision==0.14.0+rocm5.2 --index-url https://download.pytorch.org/whl/rocm5.2” and using this flags COMMANDLINE_ARGS=“–medvram --precision full --opt-sdp-attention --no-half” just follow the guide until the error and then paste the code provided here in the webui-user.sh and wait to py to download