Instructions to use unfundedResearcher/TRELLIS2-Minecraft-Texture-LoRA-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- TRELLIS.2
How to use unfundedResearcher/TRELLIS2-Minecraft-Texture-LoRA-v1 with TRELLIS.2:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
TRELLIS.2 Minecraft Texture LoRA (512, v1)
LoRA adapter for the texture stage (tex_slat_flow_model_512) of microsoft/TRELLIS.2, trained on 11,981 Minecraft build pairs from Minecraft-Fable-ImageGLB-v1.
- Rank 16 / alpha 32 / dropout 0.05, attention linears (to_qkv, to_kv, to_q, to_out)
- 10k steps, AdamW lr 1e-4, flow-matching loss (official trainer math)
- Trained on AMD MI300X (ROCm 7.14)
Recommended inference scale: 0.3 (full strength overshoots/darkens). Pair with the shape-stage LoRA: TRELLIS2-Minecraft-LoRA-v1.
Texture latents used for training: Minecraft-TRELLIS-Train-v2.
- Downloads last month
- 10
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for unfundedResearcher/TRELLIS2-Minecraft-Texture-LoRA-v1
Base model
microsoft/TRELLIS.2-4B