2065 Commits

Author	SHA1	Message	Date
Yousef Rafat	c9f5c788a7	bunch of fixes	2026-02-22 23:47:49 +02:00
Yousef Rafat	253ee4c02c	fixes	2026-02-22 01:25:10 +02:00
Yousef Rafat	1fde60b2bc	debugging	2026-02-20 22:04:37 +02:00
Yousef Rafat	b3da8ed4c5	coderabbit 2	2026-02-20 21:13:13 +02:00
Yousef Rafat	f3d4125e49	code rabbit suggestions	2026-02-20 20:16:49 +02:00
Yousef Rafat	c5a750205d	.	2026-02-20 17:39:44 +02:00
Yousef Rafat	6191cd86bf	trellis2conditioning and a hidden bug	2026-02-19 22:05:33 +02:00
Yousef Rafat	ee2b66a2f2	small fixes	2026-02-19 00:48:26 +02:00
Yousef Rafat	b5feac202c	.	2026-02-18 22:01:09 +02:00
Yousef R. Gamaleldin	9257acaae3	Merge branch 'master' into trellis2	2026-02-18 21:52:39 +02:00
rattus	58dcc97dcf	ops: limit return of requants (#12506) ... This check was far too broad and the dtype is not a reliable indicator of wanting the requant (as QT returns the compute dtype as the dtype). So explictly plumb whether fp8mm wants the requant or not.	2026-02-17 15:32:27 -05:00
chaObserv	44f8598521	Fix anima LLM adapter forward when manual cast (#12504)	2026-02-17 07:56:44 -08:00
comfyanonymous	c39653163d	Fix anima preprocess text embeds not using right inference dtype. (#12501)	2026-02-17 00:29:20 -05:00
comfyanonymous	18927538a1	Implement NAG on all the models based on the Flux code. (#12500) ... Use the Normalized Attention Guidance node. Flux, Flux2, Klein, Chroma, Chroma radiance, Hunyuan Video, etc..	2026-02-16 23:30:34 -05:00
comfyanonymous	4454fab7f0	Remove code to support RMSNorm on old pytorch. (#12499)	2026-02-16 20:09:24 -05:00
Yousef Rafat	c14317d6e0	postprocessing node fixes + model small fixes	2026-02-17 00:10:48 +02:00
comfyanonymous	88e6370527	Remove workaround for old pytorch. (#12480)	2026-02-15 20:43:53 -05:00
rattus	c0370044cd	MPDynamic: force load flux img_in weight (Fixes flux1 canny+depth lora crash) (#12446) ... * lora: add weight shape calculations. This lets the loader know if a lora will change the shape of a weight so it can take appropriate action. * MPDynamic: force load flux img_in weight This weight is a bit special, in that the lora changes its geometry. This is rather unique, not handled by existing estimate and doesn't work for either offloading or dynamic_vram. Fix for dynamic_vram as a special case. Ideally we can fully precalculate these lora geometry changes at load time, but just get these models working first.	2026-02-15 20:30:09 -05:00
Yousef Rafat	91fa563b21	rewriting conditioning logic + model code addition	2026-02-16 01:53:53 +02:00
comfyanonymous	e1ede29d82	Remove unsafe pickle loading code that was used on pytorch older than 2.4 (#12473) ... ComfyUI hasn't started on pytorch 2.4 since last month.	2026-02-14 22:53:52 -05:00
krigeta	dc9822b7df	Add working Qwen 2512 ControlNet (Fun ControlNet) support (#12359)	2026-02-13 22:23:52 -05:00
comfyanonymous	712efb466b	Add left padding to LTXAV text encoder. (#12456)	2026-02-13 21:56:54 -05:00
comfyanonymous	726af73867	Fix some custom nodes. (#12455)	2026-02-13 20:21:10 -05:00
comfyanonymous	831351a29e	Support generating attention masks for left padded text encoders. (#12454)	2026-02-13 20:15:23 -05:00
comfyanonymous	e1add563f9	Use torch RMSNorm for flux models and refactor hunyuan video code. (#12432)	2026-02-13 15:35:13 -05:00
rattus	8902907d7a	dynamic_vram: Training fixes (#12442)	2026-02-13 15:29:37 -05:00
Yousef Rafat	92aa058a58	.	2026-02-13 21:05:59 +02:00
rattus	ae79e33345	llama: use a more efficient rope implementation (#12434) ... Get rid of the cat and unary negation and inplace add-cmul the two halves of the rope. Precompute -sin once at the start of the model rather than every transformer block. This is slightly faster on both GPU and CPU bound setups.	2026-02-12 19:56:42 -05:00
rattus	117e214354	ModelPatcherDynamic: force load non leaf weights (#12433) ... The current behaviour of the default ModelPatcher is to .to a model only if its fully loaded, which is how random non-leaf weights get loaded in non-LowVRAM conditions. The however means they never get loaded in dynamic_vram. In the dynamic_vram case, force load them to the GPU.	2026-02-12 19:51:50 -05:00
Yousef Rafat	0e51bee64f	more reliable detection	2026-02-13 00:10:25 +02:00
Yousef Rafat	0e239dc39b	fixed attn (couldn't use apply_rope for dino3)	2026-02-12 23:35:57 +02:00
askmyteapot	e5ae670a40	Update ace15.py to allow min_p sampling (#12373)	2026-02-11 20:28:48 -05:00
rattus	3fe61cedda	model_patcher: guard against none model_dtype (#12410) ... Handle the case where the _model_dtype exists but is none with the intended fallback.	2026-02-11 14:54:02 -05:00
rattus	2a4328d639	ace15: Use dynamic_vram friendly trange (#12409) ... Factor out the ksampler trange and use it in ACE LLM to prevent the silent stall at 0 and rate distortion due to first-step model load.	2026-02-11 14:53:42 -05:00
rattus	d297a749a2	dynamic_vram: Fix windows Aimdo crash + Fix LLM performance (#12408) ... * model_management: lazy-cache aimdo_tensor These tensors cosntructed from aimdo-allocations are CPU expensive to make on the pytorch side. Add a cache version that will be valid with signature match to fast path past whatever torch is doing. * dynamic_vram: Minimize fast path CPU work Move as much as possible inside the not resident if block and cache the formed weight and bias rather than the flat intermediates. In extreme layer weight rates this adds up.	2026-02-11 14:50:16 -05:00
Yousef Rafat	b7764479c2	debugging	2026-02-11 20:33:59 +02:00
comfyanonymous	76a7fa96db	Make built in lora training work on anima. (#12402)	2026-02-10 22:04:32 -05:00
Kohaku-Blueleaf	cdcf4119b3	[Trainer] training with proper offloading (#12189) ... * Fix bypass dtype/device moving * Force offloading mode for training * training context var * offloading implementation in training node * fix wrong input type * Support bypass load lora model, correct adapter/offloading handling	2026-02-10 21:45:19 -05:00
Yousef Rafat	f4059c189e	dinov3 fixes + other	2026-02-11 01:27:54 +02:00
rattus	123a7874a9	ops: Fix vanilla-fp8 loaded lora quality (#12390) ... This was missing the stochastic rounding required for fp8 downcast to be consistent with model_patcher.patch_weight_to_device. Missed in testing as I spend too much time with quantized tensors and overlooked the simpler ones.	2026-02-10 13:38:28 -05:00
rattus	f719f9c062	sd: delay VAE dtype archive until after override (#12388) ... VAEs have host specific dtype logic that should override the dynamic _model_dtype. Defer the archiving of model dtypes until after.	2026-02-10 13:37:46 -05:00
rattus	fe053ba5eb	mp: dont deep-clone objects from model_options (#12382) ... If there are non-trivial python objects nested in the model_options, this causes all sorts of issues. Traverse lists and dicts so clones can safely overide settings and BYO objects but stop there on the deepclone.	2026-02-10 13:37:17 -05:00
comfyanonymous	a4be04c5d7	Ace step prompts match now. (#12376)	2026-02-09 19:45:56 -05:00
blepping	baf8c87455	Iimprovements to ACE-Steps 1.5 text encoding (part 2) (#12350)	2026-02-09 19:41:49 -05:00
rattus	62315fbb15	Dynamic VRAM fixes - Ace 1.5 performance + a VRAM leak (#12368) ... * revert threaded model loader change This change was only needed to get around the pytorch 2.7 mempool bugs, and should have been reverted along with #12260. This fixes a different memory leak where pytorch gets confused about cache emptying. * load non comfy weights * MPDynamic: Pre-generate the tensors for vbars Apparently this is an expensive operation that slows down things. * bump to aimdo 1.8 New features: watermark limit feature logging enhancements -O2 build on linux	2026-02-09 16:16:08 -05:00
Yousef Rafat	2eef826def	multiple fixes	2026-02-09 22:47:50 +02:00
Yousef Rafat	704e1b5462	small bug fixes	2026-02-09 00:41:01 +02:00
comfyanonymous	f350a84261	Disable prompt weights for ltxv2. (#12354)	2026-02-07 19:16:28 -05:00
comfyanonymous	17e7df43d1	Pad ace step 1.5 ref audio if not long enough. (#12341)	2026-02-07 00:02:11 -05:00
comfyanonymous	039955c527	Some fixes to previous pr. (#12339)	2026-02-06 20:14:52 -05:00