MiniMax H3 OpenVDN DMD8 Universal Video Generation 4K

RunningHub workflow page: https://www.runninghub.ai/zh-cn/post/2098775148949520386?inviteCode=rh-v1111

Bilibili tutorial video: https://www.bilibili.com/video/BV1vFYz6HECU/

# OpenVDN — 4K-target post-processing

## Purpose and preparation

Resolve the matching custom-node/runtime versions and assets. Load authorized media and inspect active inputs, actual schedules, effective dimensions and every save destination before a small test.

Only the main example is enabled; two complete additional examples and their saves are bypassed. Main image 15 connects first_frame and image 56 last_frame, not generic reference-image sockets. In the larger-output variants, prompt prose says references rather than first/last frames, but actual wiring takes precedence. Inspect intended roles without assuming endpoint fidelity.

The composer selects OpenVDN/vdn-minimax-h3 and stage_dmd_8nfe, with verify_hashes=false and allow_structural_base=true. Its execution plan supplies model/sampler/sigmas; no explicit numeric sigma array is serialized and the report is unused. DMD8 labels do not prove executed NFE. JSON integrity checks are not model-weight hash verification.

The conditioner stores 1344×768 and124 frames without external size/time links, with AUTO/native settings. CreateVideo stores 24fps/8-bit. If124 frames are produced, nominal duration is about 5.167 seconds, not an observed file length. Output audio is decoded generated audio, not source passthrough.

Decoded images pass through SeedVR2, linked target2160, then a resize request3840×2160, Lanczos, keep_proportion=resize, divisible_by2, CPU. Audio bypasses spatial processing. These are post-decoding targets, not native generation dimensions or measured results. Cached resize diagnostics are not render evidence.

Interpolation requests 2x,motion auto,threshold0.24,keep_audio on. This is temporal, not spatial scaling. For generation presets,24→48fps is configured intent, not measured cadence/frame count/duration. Combined VIDEO and explicit images/audio are connected; precedence requires implementation review. VIDEO output feeds SaveVideo with auto format/codec.

## Files and selected assets

- [Workflow JSON]()

- [Selected cover]()

- `qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors`

- `minimax_h3_hybrid_fl2va_ref2va_b25-49-int8.safetensors`

- `minimax_h3_video_vae_int8_convrot.safetensors`

- `minimax_h3_audio_vae_fp32.safetensors`

- `seedvr2_ema_3b_fp8_e4m3fn.safetensors`

## Version, evidence and rights

This is a local descriptive draft, not a published upgrade. Files01–08 are byte-identical to originals; only the standalone ninth file's historical video-preview field is cleared in its new copy. Originals remain unchanged.

No model inference, visual/audio acceptance, wall-clock benchmark or license verification was performed. Covers are selected assets, not accepted render evidence. Weights and external input media are not bundled. Check workflow, nodes, models, LoRAs, source media and cover rights separately; no blanket commercial license is asserted.

[Source-provided tutorial](https://www.youtube.com/watch?v=R3W2ab1_TzY) · [Source-provided workflow entry](https://www.runninghub.cn/post/2098414520787103746/?inviteCode=rh-v1111)

Links are not live-verified. The workflow is supplied through AI-KSK.