883 Commits

Author	SHA1	Message	Date
John Pollock	76f72385fa	Merge 093d17c587 into 985fb9d6ad	2026-07-05 13:55:34 +03:00
Alexis Rolland	7f287b705e	fix: Bug when setting transparency in color picker (#14764)	2026-07-04 19:13:38 -04:00
John Pollock	093d17c587	Add SeedVR2 temporal chunk and merge nodes ... ## Summary Sampling a long SeedVR2 video in one pass runs out of VRAM: the DiT working set grows linearly with latent frames times pixel area, so a 100 frame clip at 4x upscale needs more memory than any consumer card has. This PR adds two workflow-level nodes that split the latent into overlapping temporal chunks and recombine the sampled chunks with a Hann crossfade. The executor's list handling runs the stock KSampler once per chunk, so the sampler itself is untouched. - **Chunk SeedVR2 Latent** splits the latent on the temporal axis. `frames_per_chunk` is in pixel frames on the 4n+1 grid, `temporal_overlap` sets how many latent frames adjacent chunks share, and `chunking_mode=auto` solves the chunk size from measured free VRAM and the latent's own dimensions. The node outputs the effective overlap so the merge is wired, not typed. - **Merge SeedVR2 Latent Chunks** recombines the sampled chunks in order, crossfading each shared region with a Hann window (flat shoulders on the outer thirds, fade across the middle third). Zero overlap is a plain concatenation, bit-identical to `torch.cat`. ## Changes - Added 2 nodes and 1 crossfade helper to `comfy_extras/nodes_seedvr.py` (+201 lines). - Added 4 chunk-law constants to `comfy/ldm/seedvr/constants.py` (+10 lines). - Added pytest unit tests in `tests-unit/comfy_extras_test/test_seedvr2_temporal_chunk.py` (+66 lines): chunk geometry, 4n+1 and mode validation, overlap clamping, temporal noise mask slicing (5-D and 4-D), the auto law including batch scaling, crossfade weights and blend direction, round trip, and metadata handling. Everything outside the two new nodes is byte-identical to the base branch. The new constants are read only by the chunk node, and workflows that do not use these nodes take no new code path. ## Auto mode calibration The auto law is `max_latent_frames = (free_GiB - reserved - margin) / (0.30 * megapixels)`, calibrated on an RTX 5090 (32 GB) with the 3b fp16 model: a 17-cell resolution sweep plus temporal bisection located the activation wall and confirmed it is linear in latent frames times pixel area (the same total-voxel budget holds from 1.5:1 through 24:1 aspect ratios and under transposition). The margin is four standard deviations of the measured run-to-run spread, which costs about one latent frame of chunk size and makes an out-of-memory failure a lottery ticket rather than a coin flip. Manual mode bypasses the law entirely. On a 32 GB card, a 640x480x100 input at 4x upscale sampled as a single chunk allocates past 31 GiB and dies; auto mode picks 49 frame chunks (three chunks with overlap) and the full pipeline completes in about 240 seconds. The same law stands down on small inputs: a 320x240x100 clip runs as a single chunk because it fits. ## Example workflow Load a video, 4x upscale, auto chunking, temporal overlap 3. Expected output for a 640x480x100 input: 2560x1920, 100 frames, seams invisible at the default overlap. <details> <summary>API workflow JSON</summary> ```json { "14": {"inputs": {"vae_name": "ema_vae_fp16.safetensors"}, "class_type": "VAELoader"}, "17": {"inputs": {"video": ["24", 0]}, "class_type": "GetVideoComponents"}, "20": {"inputs": {"fps": ["17", 2], "images": ["30", 0], "audio": ["17", 1]}, "class_type": "CreateVideo"}, "21": {"inputs": {"tile_size": 192, "overlap": 64, "temporal_size": 64, "temporal_overlap": 8, "pixels": ["27", 0], "vae": ["14", 0]}, "class_type": "VAEEncodeTiled"}, "22": {"inputs": {"tile_size": 256, "overlap": 32, "temporal_size": 64, "temporal_overlap": 8, "samples": ["33", 0], "vae": ["14", 0]}, "class_type": "VAEDecodeTiled"}, "23": {"inputs": {"unet_name": "seedvr2_3b_fp16.safetensors", "weight_dtype": "default"}, "class_type": "UNETLoader"}, "24": {"inputs": {"file": "input.mp4", "video-preview": ""}, "class_type": "LoadVideo"}, "25": {"inputs": {"filename_prefix": "video/seedvr2_upscale", "format": "auto", "codec": "auto", "video": ["20", 0]}, "class_type": "SaveVideo"}, "26": {"inputs": {"resize_type": "scale by multiplier", "resize_type.multiplier": 4, "scale_method": "bicubic", "input": ["17", 0]}, "class_type": "ResizeImageMaskNode"}, "27": {"inputs": {"resized_images": ["26", 0]}, "class_type": "SeedVR2Preprocess"}, "28": {"inputs": {"model": ["23", 0], "vae_conditioning": ["32", 0]}, "class_type": "SeedVR2Conditioning"}, "29": {"inputs": {"seed": 5770521, "steps": 1, "cfg": 1, "sampler_name": "euler", "scheduler": "simple", "denoise": 1, "model": ["23", 0], "positive": ["28", 0], "negative": ["28", 1], "latent_image": ["32", 0]}, "class_type": "KSampler"}, "30": {"inputs": {"color_correction_method": "lab", "images": ["22", 0], "original_resized_images": ["26", 0]}, "class_type": "SeedVR2PostProcessing"}, "32": {"inputs": {"frames_per_chunk": 21, "temporal_overlap": 3, "chunking_mode": "auto", "latent": ["21", 0]}, "class_type": "SeedVR2TemporalChunk"}, "33": {"inputs": {"temporal_overlap": ["32", 1], "latent_chunks": ["29", 0]}, "class_type": "SeedVR2TemporalMerge"} } ``` </details> ## Prior art - Reference implementation: https://github.com/ByteDance-Seed/SeedVR - Community precedent for temporal chunking with blended reassembly: https://github.com/numz/ComfyUI-SeedVR2_VideoUpscaler Chunk boundaries are a mathematical compromise: the model attends within a chunk, so different chunkings produce different outputs. The overlap crossfade hides the seam; power users can widen or zero the overlap from the workflow.	2026-07-04 13:05:19 -05:00
comfyanonymous	62001efd4f	Update comfy_extras/nodes_seedvr.py ... Co-authored-by: Alexis Rolland <alexis@comfy.org>	2026-07-03 19:01:25 -07:00
comfyanonymous	d8506849f1	Update comfy_extras/nodes_seedvr.py ... Co-authored-by: Alexis Rolland <alexis@comfy.org>	2026-07-03 19:01:15 -07:00
comfyanonymous	06dc642763	Update comfy_extras/nodes_seedvr.py ... Co-authored-by: Alexis Rolland <alexis@comfy.org>	2026-07-03 19:01:02 -07:00
Alexis Rolland	e06033fb97	Merge branch 'master' into seedvr2-native-support-v5	2026-07-03 12:36:42 +08:00
comfyanonymous	c7b2c3b569	Refactors and cleanups.	2026-07-02 22:59:38 -04:00
comfyanonymous	77d42ed7e9	Remove SeedVR2ProgressiveSampler.	2026-07-01 22:19:37 -04:00
comfyanonymous	f437d87155	Cleanups using AGENTS.md	2026-07-01 22:17:51 -04:00
Alexis Rolland	6e11828d10	chore: Update nodes categories (#14674)	2026-07-01 05:20:20 +08:00
pythongosssss	603d891eaf	Update GLSL node to use ANGLE library (CORE-162) (#13195)	2026-06-27 08:40:31 +08:00
Terry Jia	dac4ea3a80	feat: Bounding boxes canvas and Ideogram JSON prompt (#14537)	2026-06-25 22:34:09 +08:00
comfyanonymous	64e1d740b8	Add advanced krea 2 model merging node. (#14621)	2026-06-24 20:37:30 -07:00
Yousef R. Gamaleldin	b22d0fb9c0	feat: Add Support For Simple Seed (CORE-295) (#14616)	2026-06-25 09:39:10 +08:00
Alexis Rolland	b0f9e326af	Add output socket to save nodes (#13866)	2026-06-22 10:15:28 +08:00
Alexis Rolland	69d34f2654	Rename a bunch of nodes (#14547)	2026-06-20 08:01:28 +08:00
Barish Ozbay	cd77c551d6	feat: Context Windows sampling with LTX2 models and IC-LoRa guides (CORE-3) (#13325)	2026-06-20 07:47:31 +08:00
Terry Jia	2ab3816dcf	feat: add Load3DAdvanced node (#14316)	2026-06-20 07:06:55 +08:00
Matt Miller	94ee49b161	harden: load training-dataset shards with weights_only=True (#14543) ... LoadTrainingDataset was the only torch.load call in the codebase without weights_only=True; comfy/utils.py and comfy/sd1_clip.py already pass it. Recent PyTorch defaults to weights_only=True, so this is defense-in-depth for installs pinned to older PyTorch. Verified a typical shard (latents + standard conditioning) round-trips cleanly under weights_only=True.	2026-06-18 15:30:57 -04:00
John Pollock	ad04a6199e	Merge branch 'master' into seedvr2-native-support-v5	2026-06-18 09:58:04 -05:00
comfyanonymous	52257bb435	Add negative prompt to boogu edit node and set min images to 0. (#14529)	2026-06-17 15:42:29 -07:00
Jukka Seppänen	e25c391888	feat: Support Boogu-Image (CORE-308) (#14523)	2026-06-17 14:22:36 -07:00
Jukka Seppänen	a590d60bb1	feat: SCAIL-2 multireference (CORE-310) (#14509) ... * SCAIl-2: support multiref	2026-06-17 16:21:23 +03:00
Alexis Rolland	d202707ff2	Update TripoSplat categories (#14512)	2026-06-17 21:02:45 +08:00
Alexis Rolland	ca1622ca24	chore: Update nodes categories (CORE-263) (#14460)	2026-06-17 08:33:09 +08:00
Jukka Seppänen	fc964047e7	feat: Support text generation with Qwen3-VL (CORE-276) (#14298)	2026-06-17 08:12:44 +08:00
Maksim	d38ea29d62	Add the checkbox to disable head drawing in node SDPoseDrawKeypoints (#14446)	2026-06-16 16:21:04 +08:00
Alexis Rolland	a439dcae07	Update nodes titles (#14417)	2026-06-16 11:42:00 +08:00
Alexander Piskun	fe54b5e955	Add 10-bit video support (#14452) ... Create Video gets a bit_depth option (8-bit/10-bit); the selected depth is carried by the video and applied when it gets encoded. Save Video and Video Slice now keep the source bit depth instead of always quantizing to 8-bit, so 10-bit videos stay 10-bit. 10-bit uses h264 with the yuv420p10le pixel format,so there's no new codec or container. Signed-off-by: bigcat88 <bigcat88@icloud.com>	2026-06-13 16:05:25 +03:00
John Pollock	d54ce3d781	Add SeedVR2 workflow nodes	2026-06-12 11:19:35 -05:00
comfyanonymous	10d466b0e3	Don't crash when using flux kv cache with split batches. (#14422)	2026-06-11 16:38:06 -07:00
Jukka Seppänen	b97e60fc6b	Fix SCAIL-2 reference mask background convention (#14415)	2026-06-11 16:17:04 +08:00
Kohaku-Blueleaf	f350acdf21	[Trainer/bug] Ensure model is not inference mode (CORE-72) (#13400) ... * Ensure model is not inference mode * force clone inside training mode to avoid inference tensor * Allow force deepcopy for model patcher	2026-06-09 23:07:47 -04:00
Talmaj	5ece24e73c	Depth anything 3 (Core-135) (#13853) ... Co-authored-by: Alexis Rolland <alexisrolland@hotmail.com>	2026-06-10 09:28:24 +08:00
Jukka Seppänen	f8e51b674c	feat: Add Bernini-R model support (Wan video) (CORE-279) (#14216)	2026-06-10 07:47:34 +08:00
Kohaku-Blueleaf	ad564899d3	Ensure conditions are not trainable to avoid bugs (#14368)	2026-06-09 15:55:29 -04:00
Alexis Rolland	9fc6f5f6dd	Move bg_removal_model input socket to first position for nicer display (#14353)	2026-06-09 23:36:56 +08:00
Jukka Seppänen	184009c2f6	feat: Add model support for SCAIL-2 (#14373) ... * initial SCAIL2 support	2026-06-09 18:24:09 +03:00
comfyanonymous	00b633f368	Revert "Add SeedVR2 support (CORE-6) (#14110)" (#14359) ... This reverts commit 7863cf0e53.	2026-06-08 18:00:20 -04:00
Silver	a1c434eb65	Improve ResolutionSelector (#14309)	2026-06-09 01:05:10 +08:00
Jukka Seppänen	fc258b10e5	Add Color primitive (#14260)	2026-06-09 00:30:28 +08:00
John Pollock	7863cf0e53	Add SeedVR2 support (CORE-6) (#14110)	2026-06-08 18:15:05 +08:00
Terry Jia	ea36cb16d6	feat(3d): reorder Preview3DAdvanced / PreviewGaussianSplat / PreviewPointCloud inputs and outputs (#14308)	2026-06-05 19:01:57 -07:00
Terry Jia	a65a5464c7	BE-1172 fix(3d): save Preview3DAdvanced / PreviewGaussianSplat / PreviewPointCloud to temp/, rename viewport input (#14294)	2026-06-05 14:18:41 -07:00
Terry Jia	2ef2cf1a7c	feat: add PreviewGaussianSplat + PreviewPointCloud nodes (#14194)	2026-06-05 12:30:58 -07:00
Jukka Seppänen	5aa71b9bc2	Enable cfg1 optimization for DualModelGuider with CFGGuider (#14290) ... * Enable cfg1 optimization for DualModelGuider * Fix CFG Override tooltip	2026-06-05 10:04:10 +03:00
Alexis Rolland	ab0d8a9203	Consolidate audio nodes into SaveAudioAdvanced node (CORE-202) (#13871)	2026-06-04 19:29:41 -07:00
Alexander Piskun	4f99ce0f8c	[Partner Nodes] fix SaveWEBM node to save alpha channel; add BriaTransparentVideoBackground Partner node (#14257)	2026-06-03 16:05:48 -07:00
Yousef R. Gamaleldin	7758b9b321	fix: Image grid bug fix (CORE-215) (#14100)	2026-06-03 16:03:32 -07:00