psychedelicious c7562dd6c0 fix(backend): mps should not use non_blocking ... We can get black outputs when moving tensors from CPU to MPS. It appears MPS to CPU is fine. See: - https://github.com/pytorch/pytorch/issues/107455 - https://discuss.pytorch.org/t/should-we-set-non-blocking-to-true/38234/28 Changes: - Add properties for each device on `TorchDevice` as a convenience. - Add `get_non_blocking` static method on `TorchDevice`. This utility takes a torch device and returns the flag to be used for non_blocking when moving a tensor to the device provided. - Update model patching and caching APIs to use this new utility. Fixes: #6545		2024-06-27 19:15:23 +10:00
..
load	fix(backend): mps should not use non_blocking	2024-06-27 19:15:23 +10:00
metadata	refactor model_install to work with refactored download queue	2024-05-13 22:49:15 -04:00
util	install model if diffusers or single file, cleaned up backend logic to not mess with existing model install	2024-03-13 21:02:29 +11:00
__init__.py	[mm] Do not write diffuser model to disk when convert_cache set to zero (#6072)	2024-03-29 16:11:08 -04:00
config.py	Initial functionality	2024-06-18 10:38:29 -04:00
convert_ckpt_to_diffusers.py	feat(mm): use same pattern for vae converter as others	2024-04-01 12:34:49 +11:00
libc_util.py	Tidy names and locations of modules	2024-03-01 10:42:33 +11:00
merge.py	[util] Add generic torch device class (#6174)	2024-04-15 13:12:49 +00:00
probe.py	[MM] Add support for probing and loading SDXL VAE checkpoint files (#6524)	2024-06-20 02:57:27 +00:00
search.py	docs(mm): update ModelSearch	2024-03-28 12:35:41 +11:00
starter_models.py	fix: update SDXL IP Adpater starter model to be ViT-H	2024-04-24 00:08:21 -04:00