I believe that there are many users who would like to try FRESCO along with their locally stored models more easily.
Therefore, I made a simple implementation so that the paths of local models can be saved and used in personal_settings.yaml.
In this implementation, I worked on the following
personal_settings.yaml is generated by initial load.
From the web UI, the path of the directory where the local model / VAE is stored can be saved in personal_settings.yaml.
The output directory of the generated results can also be specified.
In the original implementation of FRESCO's webUI.py, blend.mp4 and the generated results are overwritten each time they are generated (because the file / directory name is constant). For users who want to avoid this, I added a setting that allows incremental naming directory and blend.mp4 file name.
I have made several more changes to the FRESCO code I'm locally using.
So there may be some bugs as I made this pull request with only these features cut out.
In that case, I am sorry, but please point it out to me.
Please consider incorporating it into the main branch!
P.S. This modified version seems to work fine in my local environment.
However, when I loaded the local model for the first time, I got an error message in the CLI, so I followed the recommendation and installed omegaconf additionally.
I have not been able to verify in detail why that was necessary, but others may need that as well.
I believe that there are many users who would like to try FRESCO along with their locally stored models more easily. Therefore, I made a simple implementation so that the paths of local models can be saved and used in personal_settings.yaml.
In this implementation, I worked on the following
I have made several more changes to the FRESCO code I'm locally using. So there may be some bugs as I made this pull request with only these features cut out. In that case, I am sorry, but please point it out to me.
Please consider incorporating it into the main branch!
P.S. This modified version seems to work fine in my local environment. However, when I loaded the local model for the first time, I got an error message in the CLI, so I followed the recommendation and installed omegaconf additionally. I have not been able to verify in detail why that was necessary, but others may need that as well.