ComfyUI Node: Hybrid Loading of Minimax H3 Models for Better Quality
ThatsALovelyShirt · reddit · 2026-08-11
A developer noticed that Minimax H3's ref2va model had worse audio and visual quality compared to fl2va, despite sharing the same architecture.
By comparing them, the divergence was found mostly in the .adalnproj. tensors. The author created a custom ComfyUI model loader that selectively overlays specific tensors from ref2va onto an fl2va base at load time, with no memory overhead. Experiments showed that overlaying tensors for blocks 30-49 gives good reference capability while preserving the high quality of fl2va.
Related event: MiniMax H3 Local Long Video Generation and Optimization Practices(7 posts)→
More from Multimodal
- Kling 4.0 fully available, shows markedly better motion coherence in fast-action scenes — alifcoder · 2026-10-16
- MJev-Spatial: Single-Pass Classification Beats Autoregressive VLMs on Spatial Reasoning — mindlab-research · 2026-10-16
- H-DDCM: Hadamard Codebooks Make Diffusion-Based Image and Video Compression Scalable — Technion · 2026-10-16
- ComfyUI v0.41.0 ships with Flux3 image model and MiniMax-H3 LightVAE support — Fcking_Chuck · 2026-10-16
- AI-generated short film 'The Dragon Rider' glides cinematically above the clouds — jijo6686 · 2026-10-16
- Realistic candid Midjourney portrait prompt: natural skin, lived-in bedroom light — tisch_eins · 2026-10-16