Add upcast options, full precision sampling from float16 UNet and upcasting attention for inference using SD 2.1 models without --no-half

Fix extra network thumbs label color

Fix Approx NN previews changing first generation result

Add instruct-pix2pix hijack

Update requirement for transformers to version 4.25.1 to allow instruct-pix2pix demo code to work

Allows loading instruct-pix2pix models via same method as inpainting models in sd_models.py and sd_hijack_ip2p.py

Adds ddpm_edit.py necessary for instruct-pix2pix

Adds X/Y/Z Grid Script

Adds "Upcast cross attention layer to float32" option in Stable Diffusion settings. This allows for generating images using SD 2.1 models without --no-half or xFormers.

In order to make upcasting cross attention layer optimizations possible it is necessary to indent several sections of code in sd_hijack_optimizations.py so that a context manager can be used to disable autocast. Also, even though Stable Diffusion (and Diffusers) only upcast q and k, unfortunately my findings were that most of the cross attention layer optimizations could not function unless v is upcast also.

This also handles type casting so that ROCm and MPS torch devices work correctly without --no-half. One cast is required for deepbooru in deepbooru_model.py, some explicit casting is required for img2img and inpainting. depth_model can't be converted to float16 or it won't work correctly on some systems (it's known to have issues on MPS) so in sd_models.py model.depth_model is removed for model.half().

add PIP_INSTALLER_LOCATION env var to install pip if it's not installed
remove accidental call to accelerate when venv is disabled
add another env var to skip venv - SKIP_VENV

this is mostly just so git can detect it properly

also return the removed field to sdapi/v1/upscalers because someone might have relied on it existing

Add exception handling to API image decode

Make SwinIR upscaler interruptible and skippable

Set Linux xformers 0.0.16RC425

The loading of the model for approx nn live previews can change the internal state of PyTorch, resulting in a different image. This can be avoided by preloading the approx nn model in advance.

Added white color for labels.

Add selector to interrogate categories

you can all now stop worrying about it