diff options
author | Leonard Kugis <leonard@kug.is> | 2024-04-08 00:59:06 +0000 |
---|---|---|
committer | Leonard Kugis <leonard@kug.is> | 2024-04-08 00:59:06 +0000 |
commit | 6bb56e2dfd9e63e3a31da951be3161560150314f (patch) | |
tree | a3743bb34b43f83c346c6d71e6f9f9198b6f75ec | |
parent | 0a649e76a7e3989e944706c11aaa44cfc753feed (diff) | |
download | stable-diffusion-webui-gfx803-master.tar.gz stable-diffusion-webui-gfx803-master.tar.bz2 stable-diffusion-webui-gfx803-master.zip |
-rw-r--r-- | README.md | 11 |
1 files changed, 6 insertions, 5 deletions
@@ -9,12 +9,13 @@ This repo provides a stable-diffusion-webui compatible with gfx803 GPU architect ## Quick install for this version
1. Clone this repository: `git clone https://kug.is/cgit/stable-diffusion-webui-gfx803.git`.
-2. Build the docker image: `docker build -t stable-diffusion-webui-gfx803 .`.
+2. Create exchange directory for docker: `mkdir $HOME/dockerx`.
+3. Build the docker image: `docker build -t stable-diffusion-webui-gfx803 .`.
This step might take a long time and the CLI becomes inresponsive in between, but do not worry.
-3. Run the container for the first time: `docker run -it --network=host --device=/dev/kfd --device=/dev/dri --group-add=video --ipc=host --cap-add=SYS_PTRACE --security-opt seccomp=unconfined -v $HOME/dockerx:/dockerx stable-diffusion-webui-gfx803:latest`.
-4. Cd into the `stable-diffusion-webui-gfx803` folder, if you are not already in it.
-5. Source the venv: `source venv/bin/activate`.
-6. Launch the main script: `python launch.py --precision full --no-half`.
+4. Run the container for the first time: `docker run -it --network=host --device=/dev/kfd --device=/dev/dri --group-add=video --ipc=host --cap-add=SYS_PTRACE --security-opt seccomp=unconfined -v $HOME/dockerx:/dockerx stable-diffusion-webui-gfx803:latest`.
+5. Cd into the `stable-diffusion-webui-gfx803` folder, if you are not already in it.
+6. Source the venv: `source venv/bin/activate`.
+7. Launch the main script: `python launch.py --precision full --no-half`.
For consecutive runs you do not need to build the image again. Instead, do the following:
|