Fixes sampler name reference

Same patch will be done for TI.

Follows up on #11569

As found by Vulture and some eyes

prevent  previews from ruining hypernetwork training

rework of hypernets: rather than via settings, hypernets are added directly to prompt as <hypernet:name:weight>

removed hiding warnings from LDSR
fixed/reworked few places that produced warnings

big rework of progressbar/preview system to allow multiple users to prompts at the same time and do not get previews of each other

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