Unable to install insightface, and manually install not work with this docker

[NicoNicoNico123]

Legacy Preprocessor init warning: Unable to install insightface automatically. Please try run pip install insightface manually,

Any solve of this issue?

[Yummiii]

Just pushed a build to try to fix this, you will need to delete your data folder and run with the new build (in docker compose just add the --pull always argument. You only need to do this once) and it should be fixed. Let me know if it worked :)

[NicoNicoNico123]

I solved it by add build-essential in dockerfile, but I faced another issue of install latest driver of nvidia, once I reverse back old driver, docker can run again, it's seem the torch issue

[Yummiii]

Can you specify where you added the build-essential, because in my tests I couldn't find a place where it made a difference. And for the drivers, I'll follow it in issue 1

[NicoNicoNico123]

Can you specify where you added the build-essential, because in my tests I couldn't find a place where it made a difference. And for the drivers, I'll follow it in issue 1

FROM nvidia/cuda:12.3.2-runtime-ubuntu22.04
LABEL org.opencontainers.image.source https://github.com/Yummiii/sd-webui-forge-docker
WORKDIR /app
RUN apt update && apt upgrade -y
RUN apt install -y wget git python3 python3-venv libgl1 libglib2.0-0 apt-transport-https libgoogle-perftools-dev bc build-essential python3-pip

COPY run.sh /app/run.sh
RUN chmod +x /app/run.sh

RUN useradd -m webui
RUN chown -R webui:webui /app
USER webui
RUN mkdir /app/sd-webui

ENTRYPOINT ["/app/run.sh"]

[Yummiii]

after testing this, adding the build-essential really didn't do anything. What probably happened was that you made the change after the c64f5e6 commit and it downloaded the latest image which had the fix