Implements variable dropout rate from #4549

Fixes hypernetwork multiplier being able to modified during training, also fixes user-errors by setting multiplier value to lower values for training.

Changes function name to match torch.nn.module standard

Fixes RNG reset issue when generating previews by restoring RNG state

allow model load if previous model failed

follow-up for pr #6466 to relax response type check enforcement in get_memory

Implement API get-memory

Move PR template so github could use it

Tests fixes and additions

Allow mixed image sizes in TI/HN training

CI: Use native actions/setup-python caching

Remove fallback for Protocol import and remove Protocol import and remove instances of Protocol in code
add some whitespace between functions to be in line with other code in the repo

Add fallback for Protocol import

Add support for loading VAEs from safetensors files

Update hints.js

add a visual effect to inactive hires fix elements

Fix quicksetting name overlapping with its controls.

Fix training for newer PyTorch builds

Use an actual label for feature requests.

remove calls to locals()
add a test for img2img with script

Run scripts from API