84867067ea
Python 3.14 instructions. ( #10337 )
2025-10-14 02:09:12 -04:00
3374e900d0
Faster workflow cancelling. ( #10301 )
2025-10-13 23:43:53 -04:00
51696e3fdc
ComfyUI version 0.3.65
2025-10-13 23:39:55 -04:00
dfff7e5332
Better memory estimation for the SD/Flux VAE on AMD. ( #10334 )
2025-10-13 22:37:19 -04:00
e4ea393666
Fix loading old stable diffusion ckpt files on newer numpy. ( #10333 )
2025-10-13 22:18:58 -04:00
c8674bc6e9
Enable RDNA4 pytorch attention on ROCm 7.0 and up. ( #10332 )
2025-10-13 21:19:03 -04:00
3dfdcf66b6
convert nodes_hunyuan.py to V3 schema ( #10136 )
2025-10-13 12:36:26 -07:00
95ca2e56c8
WAN2.2: Fix cache VRAM leak on error ( #10308 )
...
Same change pattern as 7e8dd275c243ad460ed5015d2e13611d81d2a569
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
27ffd12c45
add indent=4 kwarg to json.dumps() ( #10307 )
2025-10-13 12:14:52 -07:00
e693e4db6a
Always set diffusion model to eval() mode. ( #10331 )
2025-10-13 14:57:27 -04:00
d68ece7301
Update the extra_model_paths.yaml.example ( #10319 )
2025-10-12 23:54:41 -04:00
894837de9a
update extra models paths example ( #10316 )
2025-10-12 23:35:33 -04:00
fdc92863b6
Update node docs to 0.3.0 ( #10318 )
2025-10-12 23:32:02 -04:00
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
84e9ce32c6
Implement the mmaudio VAE. ( #10300 )
2025-10-11 22:57:23 -04:00
f43b8ab2a2
Update template to 0.1.95 ( #10294 )
2025-10-11 10:27:22 -07:00
14d642acd6
feat(api-nodes): add price extractor feature; small fixes to Kling & Pika nodes ( #10284 )
2025-10-10 16:21:40 -07:00
aa895db7e8
feat(GeminiImage-ApiNode): add aspect_ratio and release version of model ( #10255 )
2025-10-10 16:17:20 -07:00
cdfc25a160
Fix save audio nodes saving mono audio as stereo. ( #10289 )
2025-10-10 17:33:51 -04:00
81e4dac107
convert nodes_upscale_model.py to V3 schema ( #10149 )
2025-10-09 16:08:40 -07:00
90853fb9cd
convert nodes_flux to V3 schema ( #10122 )
2025-10-09 16:07:17 -07:00
f1dd6e50f8
Fix bug with applying loras on fp8 scaled without fp8 ops. ( #10279 )
2025-10-09 19:02:40 -04:00
fc0fbf141c
convert nodes_sd3.py and nodes_slg.py to V3 schema ( #10162 )
2025-10-09 15:18:23 -07:00
f3d5d328a3
fix(v3,api-nodes): V3 schema typing; corrected Pika API nodes ( #10265 )
2025-10-09 15:15:03 -07:00
139addd53c
More surgical fix for #10267 ( #10276 )
2025-10-09 16:37:35 -04:00
cbee7d3390
convert nodes_latent.py to V3 schema ( #10160 )
2025-10-08 23:14:00 -07:00
6732014a0a
convert nodes_compositing.py to V3 schema ( #10174 )
2025-10-08 23:13:15 -07:00
989f715d92
convert nodes_lora_extract.py to V3 schema ( #10182 )
2025-10-08 23:11:45 -07:00
2ba8d7cce8
convert nodes_model_downscale.py to V3 schema ( #10199 )
2025-10-08 23:10:23 -07:00
51fb505ffa
feat(api-nodes, pylint): use lazy formatting in logging functions ( #10248 )
2025-10-08 23:06:56 -07:00
72c2071972
Mvly/node update ( #10042 )
...
* updated V2V node to allow for control image input
exposing steps in v2v
fixing guidance_scale as input parameter
TODO: allow for motion_intensity as input param.
* refactor: comment out unsupported resolution and adjust default values in video nodes
* set control_after_generate
* adding new defaults
* fixes
* changed control_after_generate back to True
* changed control_after_generate back to False
---------
Co-authored-by: thorsten <thorsten@tripod-digital.co.nz >
2025-10-08 20:30:41 -04:00
6e59934089
Refactor model sampling sigmas code. ( #10250 )
2025-10-08 17:49:02 -04:00
3e0eb8d33f
feat(V3-io): allow Enum classes for Combo options ( #10237 )
2025-10-08 00:14:04 -07:00
637221995f
ComfyUI version 0.3.64
2025-10-08 00:53:43 -04:00
51697d50dc
update template to 0.1.94 ( #10253 )
2025-10-07 19:48:51 -07:00
19f595b788
Bump frontend to 1.27.10 ( #10252 )
2025-10-07 17:54:00 -07:00
8a15568f10
Temp fix for LTXV custom nodes. ( #10251 )
2025-10-07 19:55:23 -04:00
9e984c48bc
feat(api-nodes): add Sora2 API node ( #10249 )
2025-10-07 14:11:37 -07:00
fc34c3d112
fix(ReCraft-API-node): allow custom multipart parser to return FormData ( #10244 )
2025-10-07 13:15:32 -07:00
8aea746212
Implement gemma 3 as a text encoder. ( #10241 )
...
Not useful yet.
2025-10-06 22:08:08 -04:00
8c19910427
convert nodes_kling.py to V3 schema ( #10236 )
2025-10-06 16:26:52 -07:00
e77e0a8f8f
convert nodes_pika.py to V3 schema ( #10216 )
2025-10-06 16:20:26 -07:00
a49007a7b0
fix(api-nodes): allow negative_prompt PixVerse to be multiline ( #10196 )
2025-10-06 16:13:43 -07:00
6ae3515801
fix(api-nodes): enable more pylint rules ( #10213 )
2025-10-06 16:05:57 -07:00
6bd3f8eb9f
ComfyUI version 0.3.63
2025-10-06 14:49:04 -04:00
7326e46dee
Update template to 0.1.93 ( #10235 )
...
* Update template to 0.1.92
* Update template to 0.1.93
2025-10-06 10:57:00 -07:00
195e0b0639
Remove useless code. ( #10223 )
2025-10-05 15:41:19 -04:00
187f43696d
fix(api-nodes): disable "std" mode for Kling2.5-turbo ( #10212 )
2025-10-04 23:34:18 -07:00
caf07331ff
Remove soundfile dependency. No more torchaudio load or save. ( #10210 )
2025-10-04 22:05:05 -04:00
b1fa1922df
convert nodes_stable3d.py to V3 schema ( #10204 )
2025-10-04 12:33:48 -07:00
comfyanonymous
comfyanonymous
Alexander Piskun
rattus128
Daniel Harte
Christian Byrne
ComfyUI Wiki
Jedrzej Kosinski
filtered