Open-weight kung fu short film made locally on one RTX 4090 with zero regenerations
Important-Apple7422 · reddit · 2026-11-18
Missing old-school 80s/90s kung fu movies, a Redditor made one — "DRUNKEN TOFU FIST" — entirely with open-weight, locally run models, achieving the result with zero regenerations:
- Keyframes: Qwen Image 2.1 (full, int8) with the Qwen3-VL 8B text encoder, using reference images to keep characters consistent across shots
- Video + audio: MiniMax H3 (image-to-video with native synced audio, int8) plus the Pika H3 Flex LoRA and Qwen3-VL 32B text encoder — one model produces the visuals, Mandarin lip-sync, dub voices, and music
- QA: Gemma 4 E2B transcribes every generated clip and dub take to verify the dialogue came out right
- Runtime: ComfyUI 0.45.5 on a single RTX 4090 (24 GB), no cloud, no API credits
A full tutorial is linked. Demonstrates what the current open-source video stack can do within 24 GB of local VRAM.
More from Multimodal
- Former voice actor launches Voyager 1, an AI voice you can direct line by line — Scobleizer · 2026-11-18
- Recreating Whiplash's tense rehearsal room with MiniMax H3 and ComfyUI — hucun16 · 2026-11-18
- Claude Plus Unreal Engine 5.9 MCP Builds a Game Prototype in One Day — TomLikesRobots · 2026-11-18
- Suno's new Visuals feature pairs AI-generated music with AI-generated video — The Decoder · 2026-11-18
- Osmo claims human-level AI smell: 10M labels trained world's first olfaction foundation model — MIT Tech Review AI · 2026-11-18
- AURA 2.0: a streaming video LLM that speaks first, calls tools, and grounds visually — Shuhan Li · 2026-11-18