{"name":"PartCrafter: Structured 3D Mesh Generation via Compositional Latent Diffusion Transformers","description":"PartCrafter is an innovative structured 3D generative model that enables the creation of complex 3D meshes and scenes from a single RGB image. Accepted at NeurIPS 2025, this project leverages compositional latent diffusion transformers to jointly generate multiple parts and objects in one shot. It offers powerful capabilities for both 3D object and scene generation, making it a valuable tool for researchers and developers in the field.","github":"https://github.com/wgsxm/PartCrafter","url":"https://osrepos.com/repo/wgsxm-partcrafter","source":"osrepos.com","sourceDescription":"This repository profile is provided by osrepos.com, an open source repository discovery platform.","repositoryProfile":"https://osrepos.com/repo/wgsxm-partcrafter","generatedFor":"open source discovery and AI-assisted research","markdown":"https://osrepos.com/repo/wgsxm-partcrafter.md","json":"https://osrepos.com/repo/wgsxm-partcrafter.json","topics":["3d generation","3d mesh","latent diffusion","transformers","image to 3d","python","generative AI","neurips"],"keywords":["3d generation","3d mesh","latent diffusion","transformers","image to 3d","python","generative AI","neurips"],"stars":null,"summary":"PartCrafter is an innovative structured 3D generative model that enables the creation of complex 3D meshes and scenes from a single RGB image. Accepted at NeurIPS 2025, this project leverages compositional latent diffusion transformers to jointly generate multiple parts and objects in one shot. It offers powerful capabilities for both 3D object and scene generation, making it a valuable tool for researchers and developers in the field.","content":"## Introduction\n\nPartCrafter is a groundbreaking open-source project that introduces a structured 3D generative model capable of creating intricate 3D meshes and entire scenes from a single 2D RGB image. This research, accepted at NeurIPS 2025, utilizes Compositional Latent Diffusion Transformers to achieve its impressive results. Unlike traditional methods, PartCrafter can jointly generate multiple parts and objects in a single pass, offering a more holistic and efficient approach to 3D content creation. It provides robust solutions for both part-level 3D object generation and comprehensive 3D scene generation.\n\n## Installation\n\nTo get PartCrafter up and running, follow these steps. The project recommends `torch-2.5.1+cu124` and `python-3.11`, though other versions might also work. A CUDA-enabled GPU with at least 8GB VRAM is required.\n\nFirst, create a conda environment (optional):\n\nbash\nconda create -n partcrafter python=3.11.13\nconda activate partcrafter\npip install torch==2.5.1 torchvision==0.20.1 torchaudio==2.5.1 --index-url https://download.pytorch.org/whl/cu124\n\n\nNext, clone the repository and install dependencies:\n\nbash\ngit clone https://github.com/wgsxm/PartCrafter.git\ncd PartCrafter\nbash settings/setup.sh\n\n\nIf you do not have root access and are using a conda environment, you can install required graphics libraries with:\n\nbash\nconda install -c conda-forge libegl libglu pyopengl\n\n\nFor Windows users, specific installation guides are available in [this pull request](https://github.com/wgsxm/PartCrafter/pull/24) and [this fork](https://github.com/JackDainzh/PartCrafter-Windows/tree/windows-main).\n\n## Examples\n\nPartCrafter offers quick start scripts for both 3D object and scene generation. The necessary model weights are automatically downloaded upon first use.\n\n### 3D Part-Level Object Generation\n\nGenerate a 3D part-level object from an image:\n\nbash\npython scripts/inference_partcrafter.py \\\n  --image_path assets/images/np3_2f6ab901c5a84ed6bbdf85a67b22a2ee.png \\\n  --num_parts 3 --tag robot --render\n\n\nSpecify `--rmbg` if you use custom images. This will remove the background of the input image and resize it appropriately.\n\n### 3D Scene Generation\n\nGenerate a 3D scene from an image:\n\nbash\npython scripts/inference_partcrafter_scene.py \\\n  --image_path assets/images_scene/np6_0192a842-531c-419a-923e-28db4add8656_DiningRoom-31158.png \\\n  --num_parts 6 --tag dining_room --render\n\n\n## Why Use PartCrafter?\n\nPartCrafter stands out for several compelling reasons:\n\n*   **Cutting-edge Research**: Recognized and accepted at NeurIPS 2025, showcasing its significant contributions to the field of generative AI.\n*   **Structured 3D Generation**: It uniquely generates 3D objects and scenes with distinct, compositionally structured parts, offering more control and realism.\n*   **Image-to-3D Capability**: Seamlessly transforms 2D RGB images into complex 3D models and environments, streamlining the content creation pipeline.\n*   **Advanced AI Architecture**: Leverages Compositional Latent Diffusion Transformers, a sophisticated approach for high-quality and coherent 3D outputs.\n*   **Versatility**: Supports both detailed part-level object generation and expansive 3D scene construction, catering to diverse application needs.\n*   **Open-Source and Accessible**: Fully open-sourced with pre-trained models available on HuggingFace, including an interactive demo for easy experimentation.\n\n## Links\n\n*   **GitHub Repository**: [https://github.com/wgsxm/PartCrafter](https://github.com/wgsxm/PartCrafter){:target=\"_blank\"}\n*   **Project Page**: [https://wgsxm.github.io/projects/partcrafter](https://wgsxm.github.io/projects/partcrafter){:target=\"_blank\"}\n*   **arXiv Paper**: [https://arxiv.org/abs/2506.05573](https://arxiv.org/abs/2506.05573){:target=\"_blank\"}\n*   **HuggingFace Demo**: [https://huggingface.co/spaces/alexnasa/PartCrafter](https://huggingface.co/spaces/alexnasa/PartCrafter){:target=\"_blank\"}\n*   **HuggingFace Model (Object)**: [https://huggingface.co/wgsxm/PartCrafter](https://huggingface.co/wgsxm/PartCrafter){:target=\"_blank\"}\n*   **HuggingFace Model (Scene)**: [https://huggingface.co/wgsxm/PartCrafter-Scene](https://huggingface.co/wgsxm/PartCrafter-Scene){:target=\"_blank\"}\n*   **YouTube Video**: [https://www.youtube.com/watch?v=ZaZHbkkPtXY](https://www.youtube.com/watch?v=ZaZHbkkPtXY){:target=\"_blank\"}","metrics":{"detailViews":7,"githubClicks":4},"dates":{"published":null,"modified":"2026-01-14T00:01:01.000Z"}}