TRELLIS.2-Text-to-3D-RERUN

A Gradio with Rerun Embedded demonstration for Microsoft’s TRELLIS.2-4B model with integrated Rerun visualization. It converts text prompts or uploaded images into high-quality, textured 3D assets (GLB) through a two-stage workflow: Text to Image (Z-Image-Turbo) → Image to 3D (TRELLIS.2). The demo features interactive 3D viewing powered by the Rerun SDK, with proper coordinate system setup, axes helpers, and downloadable GLB files.

Important

Device Support and Testing (Hopper Architecture): NVIDIA H200, Kernel (Memory-Efficient): FlashAttention 3

Features

• Text-to-Image-to-3D: Generate base images from prompts using Z-Image-Turbo, then lift to 3D.
• Direct Image-to-3D: Upload RGBA/PNG images; auto-preprocesses with background removal (BRIA-RMBG-2.0) and cropping.
• Rerun 3D Viewer: Interactive visualization with correct RIGHT_HAND_Y_UP coordinates, colored axes (X=red, Y=green, Z=blue), and clean 3D view blueprint.
• Advanced Controls: Resolutions (512/1024/1536), detailed sampler settings for sparse structure, shape, and material stages, face decimation, texture size.
• Robust Export: GLB with PNG textures (extension_webp=False for compatibility); fallback remeshing if high-quality fails.
• Session Management: Per-user temp directories; auto-cleanup on unload.
• Custom Theme: OrangeRedTheme with responsive layout.
• Rich Examples: 70+ image inputs and 60+ text prompts (cats, planes, cars, furniture, etc.).

---

---

ImageToStl.com_trellis_output_2025-12-28T182443.glb.mp4

ImageToStl.com_trellis_output_2025-12-28T190445.glb.mp4

---

Prerequisites

• Python 3.10 or higher.
• CUDA-compatible GPU (required for bfloat16 and optimizations).
• pip >= 23.0.0 (see pre-requirements.txt).
• Stable internet for initial model downloads.

Installation

1. Clone the repository:

   git clone https://github.com/PRITHIVSAKTHIUR/TRELLIS.2-Text-to-3D-RERUN-FA3.git
   cd TRELLIS.2-Text-to-3D-RERUN-FA3
   

2. Install pre-requirements: Create a pre-requirements.txt file with the following content, then run:

   pip install -r pre-requirements.txt
   

   pre-requirements.txt content:

   pip>=23.0.0
   

3. Install dependencies: Create a requirements.txt file with the following content, then run:

   pip install -r requirements.txt
   

   requirements.txt content:

   --extra-index-url https://download.pytorch.org/whl/cu124
   git+https://github.com/huggingface/diffusers.git@refs/pull/12790/head
   torch==2.6.0
   torchvision==0.21.0
   triton==3.2.0
   pillow==12.0.0
   matplotlib
   rembg
   imageio==2.37.2
   imageio-ffmpeg==0.6.0
   tqdm==4.67.1
   easydict==1.13
   opencv-python-headless==4.12.0.88
   trimesh==4.10.1
   zstandard==0.25.0
   kornia==0.8.2
   timm==1.0.22
   git+https://github.com/huggingface/transformers.git@v4.57.3
   git+https://github.com/EasternJournalist/utils3d.git@9a4eb15e4021b67b12c460c7057d642626897ec8
   https://github.com/JeffreyXiang/Storages/releases/download/Space_Wheels_251210/flash_attn_3-3.0.0b1-cp39-abi3-linux_x86_64.whl
   https://github.com/JeffreyXiang/Storages/releases/download/Space_Wheels_251210/cumesh-0.0.1-cp310-cp310-linux_x86_64.whl
   https://github.com/JeffreyXiang/Storages/releases/download/Space_Wheels_251210/flex_gemm-0.0.1-cp310-cp310-linux_x86_64.whl
   https://github.com/JeffreyXiang/Storages/releases/download/Space_Wheels_251210/o_voxel-0.0.1-cp310-cp310-linux_x86_64.whl
   https://github.com/JeffreyXiang/Storages/releases/download/Space_Wheels_251210/nvdiffrast-0.4.0-cp310-cp310-linux_x86_64.whl
   https://github.com/JeffreyXiang/Storages/releases/download/Space_Wheels_251210/nvdiffrec_render-0.0.0-cp310-cp310-linux_x86_64.whl
   omegaconf
   termcolor
   icecream
   pyserde
   gradio
   rerun-sdk
   gradio_rerun
   scipy
   jax
   jaxtyping
   monopriors
   braceexpand
   

4. Start the application:

   python app.py
   

   The demo launches at http://localhost:7860.

Usage

1. Text-to-Image-to-3D:

   • Enter prompt (e.g., "A cyberpunk Cat 3D").
   • Click "1.Generate Image".
   • Proceed to 3D.

2. Image-to-3D:

   • Upload image directly.

3. Configure:

   • Resolution, sampler params, faces/texture size.

4. Generate 3D: Click "2.Generate 3D".

5. Output:

   • Interactive Rerun viewer with proper 3D orientation.
   • Download GLB button.

Rerun Viewer

• Correct coordinate system (RIGHT_HAND_Y_UP).
• Axes helpers for orientation.
• Clean blueprint view.
• Recordings saved in tmp/ as .rrd.

Troubleshooting

• Rerun Issues: Ensure gradio_rerun and rerun-sdk; blueprint optional.
• Export Fails: Fallback remesh=False; aggressive simplification to 1M faces.
• OOM: Reduce resolution/steps; clear cache.
• Preprocessing: BRIA-RMBG requires internet.

Repository: https://github.com/PRITHIVSAKTHIUR/TRELLIS.2-Text-to-3D-RERUN-FA3.git

Contributing

Contributions welcome! Enhance Rerun blueprints, add examples, or optimize post-processing.

License

Apache License 2.0. See LICENSE for details.

Built by Prithiv Sakthi. Report issues via the repository.