Update A1111 to v1.10.1 and add ROCm support on A1111

[gafda]

This pull request introduces significant updates to the Docker setup for Stable Diffusion web UI services, with a focus on improving hardware compatibility, updating dependencies, and refactoring service definitions. The changes include upgrading base images and software versions, adding support for AMD ROCm devices, and restructuring the docker-compose.yml file to better organize service configurations for CUDA, ROCm, and CPU environments.

Hardware compatibility and service configuration:

• Added ROCm (AMD GPU) support with a new auto-rocm service and corresponding Dockerfile-rocm, including device, environment, and build settings for ROCm hardware. [1] [2]
• Refactored docker-compose.yml to split service definitions by hardware type (CUDA, ROCm, CPU) and grouped CLI arguments for each, improving maintainability and clarity. [1] [2]

Dependency and version updates:

• Updated base images and Python dependencies for all services, including the upgrade of PyTorch to 2.5.1 (CUDA) and 2.6.0 (ROCm), and bumping the Stable Diffusion web UI to version 1.10.1. [1] [2] [3] [4]
• Upgraded the Alpine Git and Bash images used for downloads and utility containers to newer versions for improved security and compatibility. [1] [2]

General improvements and bug fixes:

• Improved robustness of Gradio patching by searching for routes.py files dynamically and applying changes, with added error handling. [1] [2]
• Updated image tags and build contexts to reflect new versions and configurations for each service profile.

[pcting]

i had to uncomment this to run the profile

[gafda]

Good catch!

[pcting]

i'm getting this error on my machine:

$ docker compose --profile auto-rocm up --build -d 
...
[+] Running 1/2
 ✔ sd-auto-rocm:79                     Built                                                                              0.0s 
 ⠙ Container webui-docker-auto-rocm-1  Starting                                                                           0.1s 
DEBU[0000] otel error                                    error="<nil>"
DEBU[0000] otel error                                    error="<nil>"
Error response from daemon: could not select device driver "amd" with capabilities: [[gpu]]

any thoughts?

[gafda]

i'm getting this error on my machine:

$ docker compose --profile auto-rocm up --build -d 
...
[+] Running 1/2
 ✔ sd-auto-rocm:79                     Built                                                                              0.0s 
 ⠙ Container webui-docker-auto-rocm-1  Starting                                                                           0.1s 
DEBU[0000] otel error                                    error="<nil>"
DEBU[0000] otel error                                    error="<nil>"
Error response from daemon: could not select device driver "amd" with capabilities: [[gpu]]

any thoughts?

That is a weird error. I'll try to look into it.

Meanwhile, did you update the env vars to match your system:

    - ROCm_VERSION=6.4
    - HIP_VISIBLE_DEVICES=0
    - HSA_OVERRIDE_GFX_VERSION=11.0.0

[AJenbo]

Suggested change

	RUN . /clone.sh BLIP https://github.com/salesforce/BLIP.git 48211a1594f1321b00f14c9f7a5b4813144b2fb9
	RUN . /clone.sh k-diffusion https://github.com/crowsonkb/k-diffusion.git ab527a9a6d347f364e3d185ba6d714e22d80cb3c
	RUN . /clone.sh clip-interrogator https://github.com/pharmapsychotic/clip-interrogator 2cf03aaf6e704197fd0dae7c7f96aa59cf1b11c9
	RUN . /clone.sh generative-models https://github.com/Stability-AI/generative-models 45c443b316737a4ab6e40413d7794a7f5657c19f
	RUN . /clone.sh BLIP https://github.com/salesforce/BLIP.git 3a29b7410476bf5f2ba0955827390eb6ea1f4f9d
	RUN . /clone.sh k-diffusion https://github.com/crowsonkb/k-diffusion.git ab527a9a6d347f364e3d185ba6d714e22d80cb3c
	RUN . /clone.sh clip-interrogator https://github.com/pharmapsychotic/clip-interrogator 2cf03aaf6e704197fd0dae7c7f96aa59cf1b11c9
	RUN . /clone.sh generative-models https://github.com/Stability-AI/generative-models bc07ce62c179d3aab3053a623d96a071101d11cb

[AJenbo]

Suggested change

	pip install pyngrok xformers==0.0.26.post1 \
	git+https://github.com/TencentARC/GFPGAN.git@8d2447a2d918f8eba5a4a01463fd48e45126a379 \
	git+https://github.com/openai/CLIP.git@d50d76daa670286dd6cacf3bcd80b5e4823fc8e1 \
	git+https://github.com/mlfoundations/[email protected]
	pip install pyngrok xformers==0.0.28 \
	git+https://github.com/TencentARC/GFPGAN.git@8d2447a2d918f8eba5a4a01463fd48e45126a379 \
	git+https://github.com/openai/CLIP.git@d50d76daa670286dd6cacf3bcd80b5e4823fc8e1 \
	git+https://github.com/mlfoundations/[email protected]
	RUN pip install --upgrade --force-reinstall numpy scikit-image