Licon VBVR LoRA adapts MiniMax-H3 for image-to-video
Original title: LiconStudio/Minimax-H3-VBVR — new model trending #30 on Hugging Face
The adapter, H3-VBVR-I2V.safetensors, is trained on rebuilt VBVR data and loads in a MiniMax-H3 ComfyUI workflow.
LiconStudio published Licon VBVR, an image-to-video LoRA for MiniMax-H3, listed as trending number 30 on Hugging Face. The author says it reduces unintended object duplication, clipping, and interpenetration and improves character action logic versus MiniMax-H3-FL2VA. Training data comes from VBVR-Pro-SFT-Video, rebuilt for H3. Users download H3-VBVR-I2V.safetensors and load it in a compatible MiniMax-H3 image-to-video workflow, including ComfyUI's LoRA folder.
Unconfirmed: Improvement claims come from the author and side-by-side examples, not an independent benchmark.
- Weights file is named H3-VBVR-I2V.safetensors.
- Source dataset is VBVR-Pro-SFT-Video, rebuilt for H3.
- Comparisons are shown against original MiniMax-H3-FL2VA.
- For ComfyUI, place the file in ComfyUI/models/loras/.
- MiniMax is credited for the base model.
- Vendors
- MiniMaxLiconStudio
- AI models
- MiniMax-H3Licon VBVRMiniMax-H3-FL2VA
From the original article
1 min read at huggingface.co**Licon VBVR** is an image-to-video LoRA for **MiniMax-H3**, trained on VBVR data rebuilt and reprocessed specifically for H3. It improves object consistency and character action logic, reducing unintended object duplication, clipping, and interpenetration during motion and interactions. - **Reduced object duplication:** fewer unintended copies of objects during generation.
By LiconStudio · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.