{"id":15029688,"url":"https://github.com/thu-ml/prolificdreamer","last_synced_at":"2025-05-16T03:03:01.974Z","repository":{"id":169396382,"uuid":"645368117","full_name":"thu-ml/prolificdreamer","owner":"thu-ml","description":"ProlificDreamer: High-Fidelity and Diverse Text-to-3D Generation with Variational Score Distillation (NeurIPS 2023 Spotlight)","archived":false,"fork":false,"pushed_at":"2023-11-22T04:42:56.000Z","size":52848,"stargazers_count":1527,"open_issues_count":19,"forks_count":45,"subscribers_count":108,"default_branch":"main","last_synced_at":"2025-04-08T13:07:52.376Z","etag":null,"topics":["diffusion-model","dreamfusion","nerf","prolificdreamer","stablediffusion","text-to-3d"],"latest_commit_sha":null,"homepage":"https://ml.cs.tsinghua.edu.cn/prolificdreamer/","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/thu-ml.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null}},"created_at":"2023-05-25T13:50:00.000Z","updated_at":"2025-04-08T12:14:51.000Z","dependencies_parsed_at":"2023-07-06T16:31:44.025Z","dependency_job_id":null,"html_url":"https://github.com/thu-ml/prolificdreamer","commit_stats":null,"previous_names":["thu-ml/prolificdreamer"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/thu-ml%2Fprolificdreamer","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/thu-ml%2Fprolificdreamer/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/thu-ml%2Fprolificdreamer/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/thu-ml%2Fprolificdreamer/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/thu-ml","download_url":"https://codeload.github.com/thu-ml/prolificdreamer/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":254459083,"owners_count":22074604,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["diffusion-model","dreamfusion","nerf","prolificdreamer","stablediffusion","text-to-3d"],"created_at":"2024-09-24T20:11:22.375Z","updated_at":"2025-05-16T03:02:56.960Z","avatar_url":"https://github.com/thu-ml.png","language":"Python","funding_links":[],"categories":["Python"],"sub_categories":[],"readme":"# ProlificDreamer\n\nOfficial implementation of *[ProlificDreamer: High-Fidelity and Diverse Text-to-3D Generation with Variational Score Distillation](https://arxiv.org/abs/2305.16213)*, published in NeurIPS 2023 (Spotlight).\n\n\n\u003cp align=\"center\"\u003e\n    \u003cimg src=\"teaser.png\"\u003e\n\u003c/p\u003e\n\n## Installation\n\nThe codebase is built on [stable-dreamfusion](https://github.com/ashawkey/stable-dreamfusion). For installation, \n```\npip install -r requirements.txt\n```\n\n## Training\nProlificDreamer includes 3 stages for high-fidelity text-to-3d generation.\n```\n# --------- Stage 1 (NeRF, VSD guidance) --------- #\n# This costs approximately 27GB GPU memory, with rendering resolution of 512x512\nCUDA_VISIBLE_DEVICES=0 python main.py --text \"A pineapple.\" --iters 25000 --lambda_entropy 10 --scale 7.5 --n_particles 1 --h 512  --w 512 --workspace exp-nerf-stage1/\n# If you find the result is foggy, you can increase the --lambda_entropy. For example\nCUDA_VISIBLE_DEVICES=0 python main.py --text \"A pineapple.\" --iters 25000 --lambda_entropy 100 --scale 7.5 --n_particles 1 --h 512  --w 512 --workspace exp-nerf-stage1/\n# Generate with multiple particles. Notice that generating with multiple particles is only supported in Stage 1.\nCUDA_VISIBLE_DEVICES=0 python main.py --text \"A pineapple.\" --iters 100000 --lambda_entropy 10 --scale 7.5 --n_particles 4 --h 512  --w 512 --t5_iters 20000 --workspace exp-nerf-stage1/\n\n# --------- Stage 2 (Geometry Refinement) --------- #\n# This costs \u003c20GB GPU memory\nCUDA_VISIBLE_DEVICES=0 python main.py --text \"A pineapple.\" --iters 15000 --scale 100 --dmtet --mesh_idx 0  --init_ckpt /path/to/stage1/ckpt --normal True --sds True --density_thresh 0.1 --lambda_normal 5000 --workspace exp-dmtet-stage2/\n# If the results are with maney floaters, you can increase --density_thresh. Notice that the value of --density_thresh must be consistent in stage2 and stage3.\nCUDA_VISIBLE_DEVICES=0 python main.py --text \"A pineapple.\" --iters 15000 --scale 100 --dmtet --mesh_idx 0  --init_ckpt /path/to/stage1/ckpt --normal True --sds True --density_thresh 0.4 --lambda_normal 5000 --workspace exp-dmtet-stage2/\n\n# --------- Stage 3 (Texturing, VSD guidance) --------- #\n# texturing with 512x512 rasterization\nCUDA_VISIBLE_DEVICES=0 python main.py --text \"A pineapple.\" --iters 30000 --scale 7.5 --dmtet --mesh_idx 0  --init_ckpt /path/to/stage2/ckpt --density_thresh 0.1 --finetune True --workspace exp-dmtet-stage3/\n```\n\nWe also provide a script that can automatically run these 3 stages.\n```\nbash run.sh gpu_id text_prompt\n```\n\nFor example, \n```\nbash run.sh 0 \"A pineapple.\"\n```\n\n**Limitations:** (1) Our work ultilizes the original Stable Diffusion without any 3D data, thus the multi-face Janus problem is prevalent in the results. Ultilizing text-to-image diffusion which has been finetuned on multi-view images will alleviate this problem.\n(2) If the results are not satisfactory, try different seeds. This is helpful if the results have a good quality but suffer from the multi-face Janus problem.\n\n## TODO List\n- [x] Release our code.\n- [ ] Combine MVDream with VSD to alleviate the multi-face problem.\n\n## Related Links\n- ProlificDreamer is also integrated in [Threestudio](https://github.com/threestudio-project/threestudio) library ❤️.\n- [DreamCraft3D](https://mrtornado24.github.io/DreamCraft3D/)\n- [Fantasia3D](https://fantasia3d.github.io/)\n- [Magic3D](https://research.nvidia.com/labs/dir/magic3d/)\n- [DreamFusion](https://dreamfusion3d.github.io/)\n- [SJC](https://pals.ttic.edu/p/score-jacobian-chaining)\n- [Latent-NeRF](https://github.com/eladrich/latent-nerf)\n\n## BibTeX\nIf you find our work useful for your project, please consider citing the following paper.\n\n```\n@inproceedings{wang2023prolificdreamer,\n  title={ProlificDreamer: High-Fidelity and Diverse Text-to-3D Generation with Variational Score Distillation},\n  author={Zhengyi Wang and Cheng Lu and Yikai Wang and Fan Bao and Chongxuan Li and Hang Su and Jun Zhu},\n  booktitle={Advances in Neural Information Processing Systems (NeurIPS)},\n  year={2023}\n}\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fthu-ml%2Fprolificdreamer","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fthu-ml%2Fprolificdreamer","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fthu-ml%2Fprolificdreamer/lists"}