GitLab

There's no reason for the whole CrossAttention object to be repeated when
only the operation in the middle changes.

Control loras are controlnets where some of the weights are stored in
"lora" format: an up and a down low rank matrice that when multiplied
together and added to the unet weight give the controlnet weight.

This allows a much smaller memory footprint depending on the rank of the
matrices.

These controlnets are used just like regular ones.

Tomesd now uses q instead of x to decide which tokens to merge because
it seems to give better results.

Large refactor of the model detection and loading code.

Change the transformer patches function format to be more future proof.

Default pytorch Linear initializes the weights which is useless and slow.