TiledWan Video VACE Pipeline
Wan2.1 VACE video processing node that applies WanVideo VACE pipeline to large videos through tiling. Enables processing of arbitrarily large videos that would otherwise exceed memory or models limits. Uses dimension-wise tiling with temporal and spatial overlap to maintain quality and consistency. The complete WanVideo VACE pipeline is applied to each tile with overwriting on overlapping regions for consistency. Processing Algorithm: 1. Temporal tiling: Split video into chunks with overlap (default: 81 frames, 10-frame overlap) 2. Spatial tiling: Split each chunk into tiles with overlap (default: 832×480, 20-pixel overlap) 3. Temporal consistency: Previous chunks provide reference frames for upcoming chunks 4. Spatial consistency: Already-processed neighboring tiles overwrite overlapping regions 5. WanVideo VACE processing: Each tile processed through complete pipeline 6. Dimension-wise stitching: Column-wise → Line-wise → Temporal stitching 7. Final cropping: Output matches exact input dimensions Key Features: - Handles any video size through tiling - Temporal consistency across chunks via frame reference chaining - Spatial consistency through neighbor tile overwriting - Memory-efficient with model offloading between tiles - Complete WanVideo VACE pipeline integration - Fade blending for seamless stitching - Tensor safety with hard copies to prevent data contamination Consistency Mechanisms: - Temporal: Last frames from previous chunks overwrite first frames of current chunks - Spatial: Left/top neighbors overwrite overlapping edges in current tiles - All overwritten regions have masks zeroed for proper processing - Reference frames are used to maintain temporal coherence across temporal chunks
- video
- mask
- model
- vae
- text_embeds
- samples
- feta_args
- context_options
- loop_args
- sigmas
- unianimate_poses
- fantasytalking_embeds
- uni3c_embeds
- multitalk_embeds
- freeinit_args
- cache_args
- slg_args
- experimental_args
- vace_ref_images
- processed_video
- processing_info
Inputs (47)
| Name | Type | Default | Description |
|---|---|---|---|
| video | IMAGE | Input video frames to process. Can be very large videos that will be automatically tiled for memory efficiency. | |
| mask | MASK | Mask defining areas to be processed. Should match the video dimensions and frame count. | |
| model | WANVIDEOMODEL | WanVideo model for video generation/processing. | |
| vae | WANVAE | Video VAE for encoding/decoding frames. | |
| target_frames | INT | 8116–200 | Target number of frames per temporal chunk. The model works best with 81 frames. If different, might generate artefact or frame drops |
| target_width | INT | 83264–2048 | Target width for spatial tiles. Works best around 832. |
| target_height | INT | 48064–2048 | Target height for spatial tiles. Works best around 480. |
| frame_overlap | INT | 100–40 | Number of overlapping frames between temporal chunks. |
| spatial_overlap | INT | 200–100 | Pixel overlap between spatial tiles. |
| steps | INT | 30 | Number of diffusion sampling steps. |
| cfg | FLOAT | 6.000–30 | Classifier-free guidance strength. Higher values follow the conditioning more closely. |
| shift | FLOAT | 5.000–1000 | Shift parameter for the sampling schedule. |
| seed | INT | 00–18446744073709550000 | Random seed for reproducible results. Use the same seed to get identical outputs. |
| scheduler | COMBO | unipc | Sampling scheduler algorithm. |
| vace_strength | FLOAT | 1.0000–10 | VACE encoding strength. Controls how much the video is modified during processing. 1.0 mean total reconstruction, 0.0 means no change. |
| vace_start_percent | FLOAT | 0.000–1 | — |
| vace_end_percent | FLOAT | 1.000–1 | — |
| decode_enable_vae_tiling | BOOLEAN | false | — |
| decode_tile_x | INT | 27240–2048 | — |
| decode_tile_y | INT | 27240–2048 | — |
| decode_tile_stride_x | INT | 14432–2040 | — |
| decode_tile_stride_y | INT | 12832–2040 | — |
| debug_mode | BOOLEAN | true | — |
| debug_color_shift | BOOLEAN | false | Enable color shift debugging to visualize tile boundaries. Useful for debugging tiling artifacts. |
| force_offload_between_tiles | BOOLEAN | true | — |
| text_embedsopt | WANVIDEOTEXTEMBEDS | — | |
| samplesopt | LATENT | — | |
| riflex_freq_indexopt | INT | 00–1000 | — |
| denoise_strengthopt | FLOAT | 1.000–1 | — |
| force_offloadopt | BOOLEAN | true | — |
| batched_cfgopt | BOOLEAN | false | — |
| rope_functionopt | COMBO | comfy | 2 options: default, comfy |
| feta_argsopt | FETAARGS | — | |
| context_optionsopt | CONTEXTOPTIONS | — | |
| loop_argsopt | LOOPARGS | — | |
| sigmasopt | SIGMAS | — | |
| unianimate_posesopt | UNIANIMATE_POSES | — | |
| fantasytalking_embedsopt | FANTASYTALKING_EMBEDS | — | |
| uni3c_embedsopt | UNI3C_EMBEDS | — | |
| multitalk_embedsopt | MULTITALK_EMBEDS | — | |
| freeinit_argsopt | FREEINIT_ARGS | — | |
| cache_argsopt | CACHEARGS | — | |
| slg_argsopt | SLGARGS | — | |
| experimental_argsopt | EXPERIMENTALARGS | — | |
| vace_ref_imagesopt | IMAGE | — | |
| vace_tiled_vaeopt | BOOLEAN | false | — |
| decode_normalizationopt | COMBO | default | 2 options: default, minmax |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| processed_video | IMAGE | Processed video with WanVideo VACE pipeline applied through tiled processing. Maintains original dimensions and quality. |
| processing_info | STRING | Detailed information about the processing including tile counts, timing, and processing summary. |