Is your feature request related to a problem? Please describe.
I did some testing of single image and noticed that the cogvlm model unloads after it does a caption. Is there any chance we can get a setting that keeps the model in the ram until we switch models or unload it via a button?
Describe the solution you'd like
Keep model in ram until I press unload button or switch models
Is your feature request related to a problem? Please describe. I did some testing of single image and noticed that the cogvlm model unloads after it does a caption. Is there any chance we can get a setting that keeps the model in the ram until we switch models or unload it via a button?
Describe the solution you'd like Keep model in ram until I press unload button or switch models