Is it possible to use subset of Webui-Forge to convert SafeTensors into PyTorch models and continue in a Pytorch workflow

[M-wag]

Hey there, really enjoying how seamless Webui-Forge works with safetensors. Very impressed how it just works out-of-the-box. For my research I need an easy way to load in BaseModels and ControlNets and have direct access to the intermediate steps of the diffusion trajectory. I am aware you can have all the images downloaded, but I'd prefer to have direct access to the torch array.

So my question: Is it possible to isolate the conversions from Safe Tensors to Pytorch models, so I can continue working in a pure-Pytorch workflow and not rely on the API or WebUI. I tried to figure this out for both AUTOMATIC1111 and Forge, but had no luck.

Thank you very much.