fix #3451 scripts ignores file format settings for grids

Fix issue with 16xx cards

Slightly improve page scroll jumping with focus

Fixed incorrect negative prompt text in infotext

Fixes #4137 caused by race condition in training when VAE is unloaded

add `--gradio-inpaint-tool` and option to specify `color-sketch`

Use devices.autocast() and fix MPS randn issues

Fix clip skip of 1 not being restored from prompts

fix typo

Fixed AttributeError where openaimodel is not found

Fix divide by 0 error

Make VAE step sequential to prevent VRAM spikes, will fix #3059, #2082, #2561, #3462

Fixed safety checker for ckpt files written with pytorch >=1.13

Fix of the edge case 0 weight that occasionally will pop up in some specific situations. This was crashing the script.

torch.manual_seed() already sets a CPU generator, so there is no reason to create a CPU generator manually. torch.randn_like also needs a MPS fix for k-diffusion, but a torch hijack with randn_like already exists so it can also be used for that.

Previously only the first negative prompt in all_negative_prompts was
being used for infotext. This fixes that by selecting the index-th
negative prompt

fix an error that happens when you type into prompt while switching model, put queue stuff into separate file

if image on disk was deleted between being generated and request being completed, do use temporary dir to store it for the browser

resolve [name] after resolving [filewords] in training

Add CORS-allow policy launch argument using regex

Move progress info to beginning of site title

Fix support for devices other than CUDA in DeepBooru