Hugging Face Publishes TRELLIS.2 LoRA Fine-Tuning Guide and Scripts
Start here
Hugging Face released fine-tuning scripts and containers for TRELLIS.2, a 3D generation model, which can run locally or on Hugging Face Jobs. The guide applies LoRA adapters through the PEFT library only to the sparse-structure flow DiT and shape flow DiT, while leaving the material stage frozen. The release targets a concrete failure: on a 188-object footwear subset of the GSO dataset, the pretrained model often rendered incorrect soles, and the LoRA fine-tune is meant to correct this. Users need an HF_TOKEN and an account with access to DINOv3, which is a gated model; those using their own dataset must upload its meshes to the Hub first, and setting --push-to makes the container upload adapter checkpoints to the Hub.
Generated from reports · updated 22 h ago
Report timeline
Follow the reports to see the event from different sides.
- Hugging Face BlogHugging Face Releases Fine-Tuning Guide and Scripts for TRELLIS.2 3D Generation
Hugging Face published fine-tuning scripts and containers for TRELLIS.2, a 3D generation model, runnable locally or on Hugging Face Jobs. The guide uses LoRA adapters via the PEFT library on only the sparse-structure flow DiT and shape flow DiT, leaving the material stage frozen. On a 188-object footwear subset of the GSO dataset, the pretrained model often rendered incorrect soles, which the LoRA fine-tune targets.
Heat for this event
Not enough continuous observations yet to draw a trend.