MiniMax I2V LoRAs seem to do nothing: ComfyUI user reports weights make no difference
SnooMachines1543 · reddit · 2026-09-14
A Reddit user reports that MiniMax LoRAs used in the ComfyUI H3 I2V template appear to have little to no effect: weights tuned from low to high, placed before/after the template's existing LoRAs, or bypassing those LoRAs entirely all produced virtually identical outputs.
They also found that I2V generation defaults to a "fixed" seed — and even after switching to randomize, outputs stay identical unless the prompt or reference image changes. No resolution yet; the post is seeking explanations.
More from Multimodal
- HyperFrames + HeyGen: decent video from a basic prompt in ~10 minutes — toolstelegraph · 2026-10-30
- Hailuo MCP + Claude: chaining motion graphics generation through an agent — aziz4ai · 2026-10-30
- AI short film Nexus ep 7 cost $28,775 to make, creator shares full breakdown — NightsRadiant · 2026-10-30
- LayerShop: open-source Photoshop-style editor with ComfyUI local AI on layers — HonestTicket6013 · 2026-10-30
- AI video prompt goes viral: rally car racing a collapsing volcanic mountainside — umesh_ai · 2026-10-30
- Redditor finishes a short dinosaur movie made with Veo 3 and Seedance — HeadOpen4823 · 2026-10-30