Commit Graph
2344 Commits
Author SHA1 Message Date
patientxandGitHub 8590e1f713 Merge branch 'comfyanonymous:master' into master 2025-10-26 14:29:29 +03:00
comfyanonymousandGitHub f6bbc1ac84 Fix mistake. (#10484) 2025-10-25 23:07:29 -04:00
comfyanonymousandGitHub 098a352f13 Add warning for torch-directml usage (#10482)
Added a warning message about the state of torch-directml.
2025-10-25 20:05:22 -04:00
comfyanonymousandGitHub 426cde37f1 Remove useless function (#10472) 2025-10-24 19:56:51 -04:00
patientxandGitHub 0b5b11787f Merge branch 'comfyanonymous:master' into master 2025-10-24 13:17:09 +03:00
comfyanonymousandGitHub 1bcda6df98 WIP way to support multi multi dimensional latents. (#10456) 2025-10-23 21:21:14 -04:00
patientxandGitHub 043932cb67 Merge pull request #347 from nota-rudveld/patch-2
fix: check for version compatibility before calling PyTorch method
2025-10-23 02:28:35 +03:00
patientxandGitHub d4bcb93575 Merge branch 'comfyanonymous:master' into master 2025-10-22 11:34:33 +03:00
comfyanonymousandGitHub 9cdc64998f Only disable cudnn on newer AMD GPUs. (#10437) 2025-10-21 19:15:23 -04:00
patientxandGitHub 5bf1c8be44 Merge branch 'comfyanonymous:master' into master 2025-10-21 03:49:14 +03:00
nota-rudveldandGitHub c1c044c8cb fix: check for version compatibility before calling PyTorch method 2025-10-20 17:28:14 -04:00
comfyanonymousandGitHub 2c2aa409b0 Log message for cudnn disable on AMD. (#10418) 2025-10-20 15:43:24 -04:00
patientxandGitHub 657a7872ab Merge branch 'comfyanonymous:master' into master 2025-10-19 15:20:17 +03:00
comfyanonymousandGitHub b4f30bd408 Pytorch is stupid. (#10398) 2025-10-19 01:25:35 -04:00
comfyanonymousandGitHub dad076aee6 Speed up chroma radiance. (#10395) 2025-10-18 23:19:52 -04:00
comfyanonymousandGitHub 0cf33953a7 Fix batch size above 1 giving bad output in chroma radiance. (#10394) 2025-10-18 23:15:34 -04:00
comfyanonymousandGitHub 5b80addafd Turn off cuda malloc by default when --fast autotune is turned on. (#10393) 2025-10-18 22:35:46 -04:00
comfyanonymousandGitHub 9da397ea2f Disable torch compiler for cast_bias_weight function (#10384)
* Disable torch compiler for cast_bias_weight function

* Fix torch compile.
2025-10-17 20:03:28 -04:00
patientxandGitHub 76dde47dbb Merge branch 'comfyanonymous:master' into master 2025-10-18 00:05:02 +03:00
comfyanonymousandGitHub b1293d50ef workaround also works on cudnn 91200 (#10375) 2025-10-16 19:59:56 -04:00
comfyanonymousandGitHub 19b466160c Workaround for nvidia issue where VAE uses 3x more memory on torch 2.9 (#10373) 2025-10-16 18:16:03 -04:00
patientxandGitHub 7b0643ada1 Merge branch 'comfyanonymous:master' into master 2025-10-16 16:40:50 +03:00
FaychandGitHub afa8a24fe1 refactor: Replace manual patches merging with merge_nested_dicts (#10360) 2025-10-15 17:16:09 -07:00
Jedrzej KosinskiandGitHub 493b81e48f Fix order of inputs nested merge_nested_dicts (#10362) 2025-10-15 16:47:26 -07:00
patientxandGitHub 26589a3a0b Merge branch 'comfyanonymous:master' into master 2025-10-15 12:18:21 +03:00
comfyanonymousandGitHub 1c10b33f9b gfx942 doesn't support fp8 operations. (#10348) 2025-10-15 00:21:11 -04:00
comfyanonymousandGitHub 3374e900d0 Faster workflow cancelling. (#10301) 2025-10-13 23:43:53 -04:00
comfyanonymousandGitHub dfff7e5332 Better memory estimation for the SD/Flux VAE on AMD. (#10334) 2025-10-13 22:37:19 -04:00
comfyanonymousandGitHub e4ea393666 Fix loading old stable diffusion ckpt files on newer numpy. (#10333) 2025-10-13 22:18:58 -04:00
comfyanonymousandGitHub c8674bc6e9 Enable RDNA4 pytorch attention on ROCm 7.0 and up. (#10332) 2025-10-13 21:19:03 -04:00
patientxandGitHub eae7a58e60 Merge branch 'comfyanonymous:master' into master 2025-10-14 02:07:30 +03:00
rattus128andGitHub 95ca2e56c8 WAN2.2: Fix cache VRAM leak on error (#10308)
Same change pattern as 7e8dd275c2
applied to WAN2.2

If this suffers an exception (such as a VRAM oom) it will leave the
encode() and decode() methods which skips the cleanup of the WAN
feature cache. The comfy node cache then ultimately keeps a reference
this object which is in turn reffing large tensors from the failed
execution.

The feature cache is currently setup at a class variable on the
encoder/decoder however, the encode and decode functions always clear
it on both entry and exit of normal execution.

Its likely the design intent is this is usable as a streaming encoder
where the input comes in batches, however the functions as they are
today don't support that.

So simplify by bringing the cache back to local variable, so that if
it does VRAM OOM the cache itself is properly garbage when the
encode()/decode() functions dissappear from the stack.
2025-10-13 15:23:11 -04:00
comfyanonymousandGitHub e693e4db6a Always set diffusion model to eval() mode. (#10331) 2025-10-13 14:57:27 -04:00
patientxandGitHub fa7942933b Merge branch 'comfyanonymous:master' into master 2025-10-12 13:56:39 +03:00
comfyanonymousandGitHub a125cd84b0 Improve AMD performance. (#10302)
I honestly have no idea why this improves things but it does.
2025-10-12 00:28:01 -04:00
comfyanonymousandGitHub 84e9ce32c6 Implement the mmaudio VAE. (#10300) 2025-10-11 22:57:23 -04:00
patientxandGitHub aa6afacc01 Merge branch 'comfyanonymous:master' into master 2025-10-10 02:25:35 +03:00
comfyanonymousandGitHub f1dd6e50f8 Fix bug with applying loras on fp8 scaled without fp8 ops. (#10279) 2025-10-09 19:02:40 -04:00
patientxandGitHub 3553ce45e5 Merge branch 'comfyanonymous:master' into master 2025-10-09 23:40:21 +03:00
comfyanonymousandGitHub 139addd53c More surgical fix for #10267 (#10276) 2025-10-09 16:37:35 -04:00
patientxandGitHub 77fc639ed2 Merge branch 'comfyanonymous:master' into master 2025-10-09 15:55:12 +03:00
comfyanonymousandGitHub 6e59934089 Refactor model sampling sigmas code. (#10250) 2025-10-08 17:49:02 -04:00
patientxandGitHub 2502069447 Merge branch 'comfyanonymous:master' into master 2025-10-07 14:02:12 +03:00
comfyanonymousandGitHub 8aea746212 Implement gemma 3 as a text encoder. (#10241)
Not useful yet.
2025-10-06 22:08:08 -04:00
patientxandGitHub e5c08bb5c2 Merge branch 'comfyanonymous:master' into master 2025-10-06 00:08:39 +03:00
comfyanonymousandGitHub 195e0b0639 Remove useless code. (#10223) 2025-10-05 15:41:19 -04:00
patientxandGitHub c3a59c8e40 Merge branch 'comfyanonymous:master' into master 2025-10-04 00:51:31 +03:00
Finn-HeckerandGitHub 93d859cfaa Fix type annotation syntax in MotionEncoder_tc __init__ (#10186)
## Summary
Fixed incorrect type hint syntax in `MotionEncoder_tc.__init__()` parameter list.

## Changes
- Line 647: Changed `num_heads=int` to `num_heads: int` 
- This corrects the parameter annotation from a default value assignment to proper type hint syntax

## Details
The parameter was using assignment syntax (`=`) instead of type annotation syntax (`:`), which would incorrectly set the default value to the `int` class itself rather than annotating the expected type.
2025-10-03 14:32:19 -07:00
patientxandGitHub 603dfa1a65 Merge branch 'comfyanonymous:master' into master 2025-10-02 14:06:08 +03:00
rattus128andGitHub 4965c0e2ac WAN: Fix cache VRAM leak on error (#10141)
If this suffers an exception (such as a VRAM oom) it will leave the
encode() and decode() methods which skips the cleanup of the WAN
feature cache. The comfy node cache then ultimately keeps a reference
this object which is in turn reffing large tensors from the failed
execution.

The feature cache is currently setup at a class variable on the
encoder/decoder however, the encode and decode functions always clear
it on both entry and exit of normal execution.

Its likely the design intent is this is usable as a streaming encoder
where the input comes in batches, however the functions as they are
today don't support that.

So simplify by bringing the cache back to local variable, so that if
it does VRAM OOM the cache itself is properly garbage when the
encode()/decode() functions dissappear from the stack.
2025-10-01 18:42:16 -04:00