How can I change the LLM model or add unloading fine-tuning parameters during a web run?

The LLM model is loaded once before the web is run, so how do I replace it with another LLM model for later use. In addition to that, how do I add or unload parameters for fine tuning (LoRA, P-tunging V2) to the current LLM model.

I’m new to all of this and still learning, but the information wasn’t available, so thank you very much for reading. :slight_smile: