clip error
Sadly my SSD died and I had to reinstall forge. Now when I start the tool, it states that it can not install clip.
I downloaded this file: Forge with CUDA 12.1 + Pytorch 2.3.1 and unpacked it into a folder "forge" (like I did in the past) and ran update.bat then run.bat.
Installing clip
Traceback (most recent call last):
File "G:\aitools\forge\webui\launch.py", line 54, in <module>
main()
File "G:\aitools\forge\webui\launch.py", line 42, in main
prepare_environment()
File "G:\aitools\forge\webui\modules\launch_utils.py", line 443, in prepare_environment
run_pip(f"install {clip_package}", "clip")
File "G:\aitools\forge\webui\modules\launch_utils.py", line 153, in run_pip
return run(f'"{python}" -m pip {command} --prefer-binary{index_url_line}', desc=f"Installing {desc}", errdesc=f"Couldn't install {desc}", live=live)
File "G:\aitools\forge\webui\modules\launch_utils.py", line 125, in run
raise RuntimeError("\n".join(error_bits))
RuntimeError: Couldn't install clip.
Command: "G:\aitools\forge\system\python\python.exe" -m pip install https://github.com/openai/CLIP/archive/d50d76daa670286dd6cacf3bcd80b5e4823fc8e1.zip --prefer-binary
Error code: 2
stdout: Collecting https://github.com/openai/CLIP/archive/d50d76daa670286dd6cacf3bcd80b5e4823fc8e1.zip
Using cached https://github.com/openai/CLIP/archive/d50d76daa670286dd6cacf3bcd80b5e4823fc8e1.zip (4.3 MB)
Installing build dependencies: started
Installing build dependencies: finished with status 'done'
Getting requirements to build wheel: started
Getting requirements to build wheel: finished with status 'done'
My workaround is to uncomment the clip check in launch_utils.py
# if not is_installed("clip"):
# run_pip(f"install {clip_package}", "clip")
# startup_timer.record("install clip")
# if not is_installed("open_clip"):
# run_pip(f"install {openclip_package}", "open_clip")
# startup_timer.record("install open_clip")
But this does not look right for me..
System: intel 12900k cpu nvidia 3080ti Windows11 dunno if this is even important :D
The actual checks for packages have always been on the lackluster side even back with A1111. I reckon it's because the severity of the issue is pretty low. Some manual intervention is required from time to time in order to adapt to every possible permutation of configs