shell scripts in prepare_weights don't work without modification

[truedat101]

Probably there is a difference in the current developer "internal" build vs what got released to OSS. The issues are identified below with solution:

• git clone https://huggingface.co/runwayml/stable-diffusion-v1-5 ckpts/Base_Model will fail because Base_Model exists. It really should be: git clone https://huggingface.co/runwayml/stable-diffusion-v1-5 ckpts/Base_Model/stable-diffusion-v1-5
• Without previous change, later when running inference scripts, it will fail because stable-diffusion-v1-5 path is not found. fix is in previous step
• When re-running these scripts, they will fail because directories exist. TODO: identify a way to "update" ... possibly add update scripts if the models will change.
• in the prepare_weights/down_magictime_module.sh, you cannot move the repository without first deleting the .git subdirectory inside of MagicTime/.git because the clone repository has a write protected object. Why not use a git submodule? Or simply remove the .git dir in the freshly cloned subdirectory.
• The mv dir command in prepare_weights/down_magictime_module.sh because the directory exists. Either use cp -r, or simply do mv MagicTime/Magic_Weights/* ckpts/Magic_Weights/

[truedat101]

A CI server will catch all of these kinds of things. The easiest way to avoid a CI server is to use a docker container, and just make sure the commands all run and you can launch some test script.

[SHYuanBest]

Thanks for your interest. Is the problem solved now? We currently don’t have the manpower to package it into docker. This step will be completed in the future.

[truedat101]

I can resolve it, but I am pointing out the scripts are failing as currently designed. I can provide a gist with a fix or a pull request if your team will take contributions. Several other members of my team are planning to use the project but I need to make it possible to set up and install.

[SHYuanBest]

ok, please submit a PR

[truedat101]

Will work on this tonight.

[SHYuanBest]

thanks