ComfyUI Node
TS H3 Regenerate
Latent upscale and 2/3-step Euler refinement with NVIDIA's Open-H3-Regenerate recipe. Audio is held clean. Optional first/last frame guidance uses a normal H3 video VAE at final resolution.
TS H3 Regenerate
- model
- conditioning
- latent
- audio_vae
- first_frame
- video_vae
- last_frame
- latent
- report
◄upscale_model▾►
◄width1920►
◄height1088►
◄denoise1.00►
◄steps3►
◄seed0►
◄audio_modekeep source audio (lipsync)►
◄precisionbf16►
◄chunk_length0►
◄temporal_overlap17►
◄reference_modeoff►
◄chunk_modeindependent (crossfade)►
CategoryTS/Video
Inputs (19)
| Name | Type | Default | Description |
|---|---|---|---|
| model | MODEL | Output of TS H3 Regenerate Model. | |
| conditioning | CONDITIONING | Use the original positive conditioning. Full Reference keeps native image, video and audio references from MiniMax H3 Reference to Video. | |
| latent | LATENT | Denoised H3 video or audio+video latent. | |
| upscale_model | COMBO | 1 options: (no upscale models found in: /tmp/ComfyUI/models/latent_upscale_models) | |
| width | INT | 192064–4096 | — |
| height | INT | 108864–4096 | — |
| denoise | FLOAT | 1.000–1 | 1 = release sigma schedule; lower = less redraw, shift-6 grid from a lower starting sigma; 0 = upscale only. |
| steps | COMBO | 3 | 2 options: 3, 2 |
| seed | INT | 00–18446744073709550000 | — |
| audio_mode | COMBO | keep source audio (lipsync) | Source audio is held clean for lipsync. Reference mode encodes silence; mux the original soundtrack after decoding. |
| precision | COMBO | bf16 | Latent upscaler precision. NVIDIA reference uses bf16. |
| chunk_length | INT | 00–100000 | 0 = whole clip, release recipe. Nonzero = lower-memory chunks, in multiples of 17 frames. The stitching mode controls how chunks share context. |
| temporal_overlap | INT | 170–100000 | — |
| audio_vaeopt | VAE | H3 audio VAE; required only for silence reference mode. | |
| reference_modeopt | COMBO | off | First/last uses the normal H3 video VAE and keyframe images at final size. Full Reference uses encoded media from the original Ref2VA conditioning with a clean Ref2VA Regenerate model. Reference resolution and length affect VRAM use. |
| first_frameopt | IMAGE | Start image prepared directly from the source at final output size, before strong draft downscaling. Only the first image is used; it is fitted to width/height without cropping. | |
| video_vaeopt | VAE | Normal H3 video VAE for encoding the reference, including its compatible INT8 version. The fast decode-only VAE cannot be used here. | |
| chunk_modeopt | COMBO | independent (crossfade) | Lock overlap fixes the whole overlap. Blend overlap fixes its early part and gradually transitions to the redrawn tail. Both continuation modes require nonzero overlap for multiple chunks. |
| last_frameopt | IMAGE | End image for first and last frames mode. The first image in the batch is center-cropped to the final canvas if necessary, then encoded as a guide at the clip's last frame. |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| latent | LATENT | — |
| report | STRING | — |