Home Creators Posts Import Register Favorites Logout

Downloads

Content

Meta SAM 3D Objects Gradio app reconstructs full 3D shape, texture and layout from one image; a 30 MP image takes under 1 minute on RTX 5090.

Premium SECourses app for Meta Research SAM 3D Objects with 1-click installers for Windows, RunPod and Massed Compute, pre-compiled flash_attn, xformers, sageattention, gsplat, moge and pytorch3d, PLY output and support for RTX 2000 to RTX 5000 and cloud GPUs.

This is a premium Gradio app made for https://github.com/facebookresearch/sam-3d-objects - with pre-compiled libraries and 1-click installers for Windows, RunPod and Massed Compute

Please visit Patreon exclusive posts index to find all of our Patreon scripts, posts, apps and more.

Join discord to get help, chat, discuss and also tell me your discord username to get your special rank : SECourses Discord

Please also Star, Watch and Fork our Stable Diffusion & Generative AI  GitHub repository and join our Reddit subreddit and follow me on LinkedIn (my real profile)

=======

Latest installer zip file : SAM_3D_Objects_v3.zip

  • Hopefully I will add more features just let me know what you need this is V1.0 release

    • For exmaple I plan to add inpainting masking instead of mask upload

  • I had to pre-compile flash_attn, xformers, sageattention, gsplat, moge, pytorch3d for Windows and Linux to make 1-click installers

  • Still you have to strictly follow requirements tutorial and have CUDA 12.9, C++, MSVC, FFmpeg and others

  • SAM 3D Objects is a foundation model that reconstructs full 3D shape geometry, texture, and layout from a single image

  • The app process 30 megapixel image under 1 minute on RTX 5090 ultra fast and generated ply (Polygon File Format or the Stanford Triangle Format) files are saved inside outputs folder

  • We support all the GPUs starting from RTX 2000 series to RTX 5000 series including cloud GPUs like H100, B200, RTX PRO 6000, L40S etc as usual

Windows Requirements

  • Python 3.10, FFmpeg, CUDA 12.9, cuDNN 9.12 , C++ tools and Git

  • If it doesn't work make sure to below tutorial and install everything exactly as shown in this below tutorial

  • https://youtu.be/DrhUHnYfwC0

  • This is a very very requirements dependend app so follow tutorial exactly with its updated post

How To Install and Use:

Windows:

  • First watch above requirements tutorial and follow every step

    • Extract zip archive into a drive like c:/sam_3d_objects_v1

    • Longer directory path causes issues and don't use space characters in folder names

    • Double click and install with Windows_Install_or_Update.bat and that is it

    • Then run with Windows_Start_App.bat double click

    • Don't run anything ever as administrators

    • And follow above and below attached screenshot

Massed Compute (Recommend Cloud) :

RunPod (Cloud):

Related posts

All posts in this topic: Image to 3D Generation Apps · New to SECourses? Start Here: Best and Most Popular SECourses Posts

Files

Comments

Dazzastrous

Installing collected packages: jupyter_client Attempting uninstall: jupyter_client Found existing installation: jupyter_client 8.6.3 Uninstalling jupyter_client-8.6.3: Successfully uninstalled jupyter_client-8.6.3 ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. ipykernel 7.1.0 requires jupyter-client>=8.0.0, but you have jupyter-client 7.4.9 which is incompatible. Successfully installed jupyter_client-7.4.9

Han Vu

Collecting nvidia-pyindex==1.0.9 Using cached nvidia-pyindex-1.0.9.tar.gz (10 kB) Preparing metadata (pyproject.toml) ... error error: subprocess-exited-with-error × Preparing metadata (pyproject.toml) did not run successfully. │ exit code: 1 ╰─> [13 lines of output] running dist_info creating C:\Users\vubiq\AppData\Local\Temp\pip-modern-metadata-djucvekd\nvidia_pyindex.egg-info writing C:\Users\vubiq\AppData\Local\Temp\pip-modern-metadata-djucvekd\nvidia_pyindex.egg-info\PKG-INFO writing dependency_links to C:\Users\vubiq\AppData\Local\Temp\pip-modern-metadata-djucvekd\nvidia_pyindex.egg-info\dependency_links.txt writing entry points to C:\Users\vubiq\AppData\Local\Temp\pip-modern-metadata-djucvekd\nvidia_pyindex.egg-info\entry_points.txt writing top-level names to C:\Users\vubiq\AppData\Local\Temp\pip-modern-metadata-djucvekd\nvidia_pyindex.egg-info\top_level.txt writing manifest file 'C:\Users\vubiq\AppData\Local\Temp\pip-modern-metadata-djucvekd\nvidia_pyindex.egg-info\SOURCES.txt' reading manifest file 'C:\Users\vubiq\AppData\Local\Temp\pip-modern-metadata-djucvekd\nvidia_pyindex.egg-info\SOURCES.txt' reading manifest template 'MANIFEST.in' adding license file 'LICENSE.md' writing manifest file 'C:\Users\vubiq\AppData\Local\Temp\pip-modern-metadata-djucvekd\nvidia_pyindex.egg-info\SOURCES.txt' creating 'C:\Users\vubiq\AppData\Local\Temp\pip-modern-metadata-djucvekd\nvidia_pyindex-1.0.9.dist-info' error: invalid command 'bdist_wheel' [end of output] note: This error originates from a subprocess, and is likely not a problem with pip. error: metadata-generation-failed × Encountered error while generating package metadata. ╰─> nvidia-pyindex note: This is an issue with the package mentioned above, not pip. hint: See above for details. Traceback (most recent call last): File "E:\Programs\SAM_3D_Objects_v1\HF_model_downloader.py", line 10, in from huggingface_hub import HfApi, list_repo_files, hf_hub_url, get_hf_file_metadata ModuleNotFoundError: No module named 'huggingface_hub' Press any key to continue . . .

SECourses

Ye I get this too no issues. It is due to its dependencies

SECourses

made a fix with v2 zip file. use it to install

Roy F

============================================================ Downloading: SAM 3D Objects Checkpoints Description: Main checkpoints for SAM 3D Objects model Repository: MonsterMMORPG/Wan_GGUF Target: C:\AI\Sam3D\SAM_3D_Objects_v1\sam-3d-objects\checkpoints\hf ============================================================ [1/15] pipeline.yaml [ERROR] Cannot get size for sam-3d-objects/pipeline.yaml [FAILED] pipeline.yaml [2/15] slat_decoder_gs.ckpt [SKIP] slat_decoder_gs.ckpt already complete and verified (cached) [3/15] slat_decoder_gs.yaml [SKIP] slat_decoder_gs.yaml already complete (576.0 B) [4/15] slat_decoder_gs_4.ckpt [SKIP] slat_decoder_gs_4.ckpt already complete and verified (cached) [5/15] slat_decoder_gs_4.yaml [SKIP] slat_decoder_gs_4.yaml already complete (575.0 B) [6/15] slat_decoder_mesh.ckpt [SKIP] slat_decoder_mesh.ckpt already complete and verified (cached) [7/15] slat_decoder_mesh.yaml [SKIP] slat_decoder_mesh.yaml already complete (300.0 B) [8/15] slat_generator.ckpt [SKIP] slat_generator.ckpt already complete and verified (cached) [9/15] slat_generator.yaml [ERROR] Cannot get size for sam-3d-objects/slat_generator.yaml [FAILED] slat_generator.yaml [10/15] ss_decoder.ckpt [SKIP] ss_decoder.ckpt already complete and verified (cached) [11/15] ss_decoder.yaml [SKIP] ss_decoder.yaml already complete (244.0 B) [12/15] ss_encoder.safetensors [OK] ss_encoder.safetensors already complete [13/15] ss_encoder.yaml [OK] ss_encoder.yaml already complete [14/15] ss_generator.ckpt [SKIP] ss_generator.ckpt already complete and verified (cached) [15/15] ss_generator.yaml [ERROR] Cannot get size for sam-3d-objects/ss_generator.yaml [FAILED] ss_generator.yaml ============================================================ Downloading: MoGe Depth Model Description: MoGe ViT-L depth estimation model for pointmap generation Repository: Ruicheng/moge-vitl Target: C:\AI\Sam3D\SAM_3D_Objects_v1\sam-3d-objects\models\moge ============================================================ [1/1] model.pt [SKIP] model.pt already complete and verified (cached) ============================================================ Downloading DINOv2 Backbone (via torch.hub) ============================================================ C:\AI\Sam3D\SAM_3D_Objects_v1\sam-3d-objects\venv\lib\site-packages\torch\cuda\__init__.py:63: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you. import pynvml # type: ignore[import] [INFO] Loading DINOv2 via torch.hub (this will download the model)... Using cache found in C:\AI\Sam3D\SAM_3D_Objects_v1\sam-3d-objects\models\dinov2\hub\facebookresearch_dinov2_main C:\AI\Sam3D\SAM_3D_Objects_v1\sam-3d-objects\models\dinov2\hub\facebookresearch_dinov2_main\dinov2\layers\swiglu_ffn.py:43: UserWarning: xFormers is available (SwiGLU) warnings.warn("xFormers is available (SwiGLU)") C:\AI\Sam3D\SAM_3D_Objects_v1\sam-3d-objects\models\dinov2\hub\facebookresearch_dinov2_main\dinov2\layers\attention.py:27: UserWarning: xFormers is available (Attention) warnings.warn("xFormers is available (Attention)") C:\AI\Sam3D\SAM_3D_Objects_v1\sam-3d-objects\models\dinov2\hub\facebookresearch_dinov2_main\dinov2\layers\block.py:33: UserWarning: xFormers is available (Block) warnings.warn("xFormers is available (Block)") [OK] DINOv2 model downloaded successfully ============================================================ Patching pipeline.yaml for local model loading... ============================================================ [WARNING] pipeline.yaml not found, skipping patch ============================================================ Download Summary ============================================================ Successful: 14 Failed: 3 ⚠ Some downloads failed. Please re-run the script to retry.

SECourses

thanks downloader fixed with v3 file

SECourses

please use v3 zip file and let me know

Roy F

Thank you! It's all working now :)

Jose Garcia

What’s the best use for this? Making a character in blendr or something? Thanks in advance

SECourses

This is for 3D working people. It generates a ply (Polygon File Format or the Stanford Triangle Format). I am not expert either :D but from single image it generates a good 3d object with texture

Roy F

This is an awesome tool. On the Meta AI Demos site, they have a great tool to generate the mask by just clicking on objects in the uploaded image: https://aidemos.meta.com/segment-anything/editor/convert-image-to-3d Any way to integrate this functionality into the Gradio APP ? This would improve the workflow of not having to generate the mask externally.

SECourses

do they have github repo for this? i was planning to add gradio brush feature to mask yourself on image

Roy F

I don't think they published a repo for the auto-masking functionality

Roy F

Thanks! That will work