VDN-H3 video generation running on 24GB VRAM: 30-second clips on an RTX 3090 Ti
NoMouse9610 · reddit · 2026-09-08
The author released a ComfyUI implementation of VDN-H3 optimized for the RTX 3090 Ti's 24GB VRAM, using INT8 ConvRot weights. It was tested with both FL2VA and Ref2VA INT8 ConvRot models, working from 0.4MP to 0.8MP, and longer videos up to 30 seconds generate fine, albeit slowly.
Reference speed: 10 seconds of 0.4MP video at 8 steps takes about 2:08 on a 3090 Ti. The implementation and INT8 ConvRot weights are open-sourced on GitHub and Hugging Face, crediting OpenVDN for VDN-H3 and Saganaki22 for the original ComfyUI port.
More from Infra
- Anthropic S-1: training infra edge cut per-task cost 40% and doubled revenue growth — BenBajarin · 2026-10-26
- Rewriting scaling law study: optimal synthetic data share rises from 10% to 32% with compute — heghbalz · 2026-10-26
- US states halt new data centers as 72% of Americans oppose local builds — YvesMulkers · 2026-10-26
- $5M Seed Round Can't Afford One DeepSeek Training Run: GPU-less AI Startup's Three Paths — FinanceYF5 · 2026-10-26
- HBM demand shifts: low-stack inference plans shelved, Rubin to go 12-high only — BenBajarin · 2026-10-26
- CircuitHub unveils The Grid, a factory-scale robot making PCBs in the US twice as fast as Shenzhen — garrytan · 2026-10-26