Sv. Lockal
0e0f1ed5bc
Allow bf16 computations on CPUs with BF16 support
...
Modern CPUs have native AVX512 BF16 instructions, which significantly improves matmul and conv2d operations.
With Bfloat16 instructions UNET steps are 40-50% faster on both AMD and Intel CPUs. There are minor visible changes with bf16, but no avalanche effects, so this feature is enabled by default with new `--use-cpu-bf16=auto` option. It can be disabled with `--use-cpu-bf16=no`.
Signed-off-by: Sv. Lockal <lockalsash@gmail.com>
2025-07-02 16:07:02 +00:00
comfyanonymous
111f583e00
Fallback to regular op when fp8 op throws exception. ( #8761 )
2025-07-02 00:57:13 -04:00
chaObserv
b22e97dcfa
Migrate ER-SDE from VE to VP algorithm and add its sampler node ( #8744 )
...
Apply alpha scaling in the algorithm for reverse-time SDE and add custom ER-SDE sampler node for other solver types (SDE, ODE).
2025-07-01 02:38:52 -04:00
comfyanonymous
170c7bb90c
Fix contiguous issue with pytorch nightly. ( #8729 )
2025-06-29 06:38:40 -04:00
comfyanonymous
396454fa41
Reorder the schedulers so simple is the default one. ( #8722 )
2025-06-28 18:12:56 -04:00
xufeng
ba9548f756
“--whitelist-custom-nodes” args for comfy core to go with “--disable-all-custom-nodes” for development purposes ( #8592 )
...
* feat: “--whitelist-custom-nodes” args for comfy core to go with “--disable-all-custom-nodes” for development purposes
* feat: Simplify custom nodes whitelist logic to use consistent code paths
2025-06-28 15:24:02 -04:00
comfyanonymous
c36be0ea09
Fix memory estimation bug with kontext. ( #8709 )
2025-06-27 17:21:12 -04:00
comfyanonymous
9093301a49
Don't add tiny bit of random noise when VAE encoding. ( #8705 )
...
Shouldn't change outputs but might make things a tiny bit more
deterministic.
2025-06-27 14:14:56 -04:00
comfyanonymous
ef5266b1c1
Support Flux Kontext Dev model. ( #8679 )
2025-06-26 11:28:41 -04:00
comfyanonymous
a96e65df18
Disable omnigen2 fp16 on older pytorch versions. ( #8672 )
2025-06-26 03:39:09 -04:00
comfyanonymous
ec70ed6aea
Omnigen2 model implementation. ( #8669 )
2025-06-25 19:35:57 -04:00
comfyanonymous
7a13f74220
unet -> diffusion model ( #8659 )
2025-06-25 04:52:34 -04:00
chaObserv
8042eb20c6
Singlestep DPM++ SDE for RF ( #8627 )
...
Refactor the algorithm, and apply alpha scaling.
2025-06-24 14:59:09 -04:00
comfyanonymous
1883e70b43
Fix exception when using a noise mask with cosmos predict2. ( #8621 )
...
* Fix exception when using a noise mask with cosmos predict2.
* Fix ruff.
2025-06-21 03:30:39 -04:00
comfyanonymous
f7fb193712
Small flux optimization. ( #8611 )
2025-06-20 05:37:32 -04:00
comfyanonymous
7e9267fa77
Make flux controlnet work with sd3 text enc. ( #8599 )
2025-06-19 18:50:05 -04:00
comfyanonymous
91d40086db
Fix pytorch warning. ( #8593 )
2025-06-19 11:04:52 -04:00
chaObserv
8e81c507d2
Multistep DPM++ SDE samplers for RF ( #8541 )
...
Include alpha in sampling and minor refactoring
2025-06-16 14:47:10 -04:00
comfyanonymous
e1c6dc720e
Allow setting min_length with tokenizer_data. ( #8547 )
2025-06-16 13:43:52 -04:00
comfyanonymous
7ea79ebb9d
Add correct eps to ltxv rmsnorm. ( #8542 )
2025-06-15 12:21:25 -04:00
comfyanonymous
d6a2137fc3
Support Cosmos predict2 image to video models. ( #8535 )
...
Use the CosmosPredict2ImageToVideoLatent node.
2025-06-14 21:37:07 -04:00
chaObserv
53e8d8193c
Generalize SEEDS samplers ( #8529 )
...
Restore VP algorithm for RF and refactor noise_coeffs and half-logSNR calculations
2025-06-14 16:58:16 -04:00
comfyanonymous
29596bd53f
Small cosmos attention code refactor. ( #8530 )
2025-06-14 05:02:05 -04:00
Kohaku-Blueleaf
520eb77b72
LoRA Trainer: LoRA training node in weight adapter scheme ( #8446 )
2025-06-13 19:25:59 -04:00
comfyanonymous
c69af655aa
Uncap cosmos predict2 res and fix mem estimation. ( #8518 )
2025-06-13 07:30:18 -04:00
comfyanonymous
251f54a2ad
Basic initial support for cosmos predict2 text to image 2B and 14B models. ( #8517 )
2025-06-13 07:05:23 -04:00
pythongosssss
50c605e957
Add support for sqlite database ( #8444 )
...
* Add support for sqlite database
* fix
2025-06-11 16:43:39 -04:00
comfyanonymous
8a4ff747bd
Fix mistake in last commit. ( #8496 )
...
* Move to right place.
2025-06-11 15:13:29 -04:00
comfyanonymous
af1eb58be8
Fix black images on some flux models in fp16. ( #8495 )
2025-06-11 15:09:11 -04:00
comfyanonymous
6e28a46454
Apple most likely is never fixing the fp16 attention bug. ( #8485 )
2025-06-10 13:06:24 -04:00
comfyanonymous
7f800d04fa
Enable AMD fp8 and pytorch attention on some GPUs. ( #8474 )
...
Information is from the pytorch source code.
2025-06-09 12:50:39 -04:00
comfyanonymous
97755eed46
Enable fp8 ops by default on gfx1201 ( #8464 )
2025-06-08 14:15:34 -04:00
comfyanonymous
daf9d25ee2
Cleaner torch version comparisons. ( #8453 )
2025-06-07 10:01:15 -04:00
comfyanonymous
3b4b171e18
Alternate fix for #8435 ( #8442 )
2025-06-06 09:43:27 -04:00
comfyanonymous
4248b1618f
Let chroma TE work on regular flux. ( #8429 )
2025-06-05 10:07:17 -04:00
comfyanonymous
fb4754624d
Make the casting in lists the same as regular inputs. ( #8373 )
2025-06-01 05:39:54 -04:00
comfyanonymous
19e45e9b0e
Make it easier to pass lists of tensors to models. ( #8358 )
2025-05-31 20:00:20 -04:00
drhead
08b7cc7506
use fused multiply-add pointwise ops in chroma ( #8279 )
2025-05-30 18:09:54 -04:00
comfyanonymous
704fc78854
Put ROCm version in tuple to make it easier to enable stuff based on it. ( #8348 )
2025-05-30 15:41:02 -04:00
comfyanonymous
f2289a1f59
Delete useless file. ( #8327 )
2025-05-29 08:29:37 -04:00
comfyanonymous
5e5e46d40c
Not really tested WAN Phantom Support. ( #8321 )
2025-05-28 23:46:15 -04:00
comfyanonymous
1c1687ab1c
Support HiDream SimpleTuner loras. ( #8318 )
2025-05-28 18:47:15 -04:00
comfyanonymous
06c661004e
Memory estimation code can now take into account conds. ( #8307 )
2025-05-27 15:09:05 -04:00
comfyanonymous
89a84e32d2
Disable initial GPU load when novram is used. ( #8294 )
2025-05-26 16:39:27 -04:00
comfyanonymous
e5799c4899
Enable pytorch attention by default on AMD gfx1151 ( #8282 )
2025-05-26 04:29:25 -04:00
comfyanonymous
a0651359d7
Return proper error if diffusion model not detected properly. ( #8272 )
2025-05-25 05:28:11 -04:00
comfyanonymous
5a87757ef9
Better error if sageattention is installed but a dependency is missing. ( #8264 )
2025-05-24 06:43:12 -04:00
comfyanonymous
0b50d4c0db
Add argument to explicitly enable fp8 compute support. ( #8257 )
...
This can be used to test if your current GPU/pytorch version supports fp8 matrix mult in combination with --fast or the fp8_e4m3fn_fast dtype.
2025-05-23 17:43:50 -04:00
drhead
30b2eb8a93
create arange on-device ( #8255 )
2025-05-23 16:15:06 -04:00
comfyanonymous
f85c08df06
Make VACE conditionings stackable. ( #8240 )
2025-05-22 19:22:26 -04:00
