{"id":4240,"date":"2023-12-21T01:47:04","date_gmt":"2023-12-21T07:47:04","guid":{"rendered":"https:\/\/cocoalopez.com\/blog\/?p=4240"},"modified":"2023-12-22T18:34:39","modified_gmt":"2023-12-23T00:34:39","slug":"3d-gaussian-splatting-1-click","status":"publish","type":"post","link":"https:\/\/cocoalopez.com\/blog\/?p=4240","title":{"rendered":"3D GAUSSIAN SPLATTING TUTORIAL (ONE CLICK INSTALL)"},"content":{"rendered":"\n

For those looking for a one click (almost solution) there’s this nifty little tool called pinokio that auto installs dependencies and environments (yeah including conda) for running web apps and python based scripts with one click install and run.<\/p>\n\n\n\n

Summary:<\/h2>\n\n\n\n
    \n
  1. Install Pinokio:
    https:\/\/docs.pinokio.computer\/download\/windows.html<\/a><\/li>\n\n\n\n
  2. Install NVIDIA CUDA 11.8:
    https:\/\/developer.download.nvidia.com\/compute\/cuda\/11.8.0\/local_installers\/cuda_11.8.0_522.06_windows.exe<\/a>
    (Or newer)
    https:\/\/developer.nvidia.com\/cuda-11-8-0-download-archive?target_os=Windows&target_arch=x86_64&target_version=10&target_type=exe_local<\/a><\/li>\n\n\n\n
  3. Open Pinokio and click on DISCOVER <\/li>\n\n\n\n
  4. Paste the Gaussian Splatting GIT URL:
    https:\/\/github.com\/bycloud-AI\/gaussian-splatting-Windows.git<\/li>\n\n\n\n
  5. Click on “One click Install with Pinokio”<\/li>\n\n\n\n
  6. After Downloading everything, click on “INSTALL”, this will run the python scripts, setup the conda environment, and install almost everything needed for it.<\/li>\n\n\n\n
  7. Once its done, it will pop a Command Window, before you can use it you need to run the next two commands<\/li>\n\n\n\n
  8. pip install submodules\\diff-gaussian-rasterization<\/code><\/li>\n\n\n\n
  9. pip install submodules\\simple-knn<\/code><\/li>\n\n\n\n
  10. Once they are installed, you can run the normal python commands for COLMAP, Training and Viewing<\/li>\n<\/ol>\n\n\n\n

    There are 3 steps to follow in order to get a gaussian splat:<\/h4>\n\n\n\n
      \n
    1. Place your image sequence on input_data\/YOURPROJECT then run the following python command:
      python convert.py -s input_data\/YOURPROJECT --colmap_executable COLMAP-3.8-windows-cuda\\COLMAP.bat<\/code><\/li>\n\n\n\n
    2. After the COLMAP has been extracted, you gotta create the Gaussian Splat by training it with the following command:
      python train.py -s input_data\/YOURPROJECT<\/code><\/li>\n\n\n\n
    3. Once the training is finished, the Gaussian Splat will be located at the folder “Output”, from there you gotta rename the last created folder to the name of YOURPROJECT, eg.
      Created folder was named “de225777-e” at “output”, so you gotta rename it to “YOURPROJECT”<\/li>\n\n\n\n
    4. Once you renamed the folder, you can view your Gaussian Splat by running this python command:
      start viewers\/bin\/SIBR_GaussianViewer_app.exe . -m output\\YOURPROJECT<\/code><\/li>\n<\/ol>\n\n\n\n

      TUTORIAL<\/h2>\n\n\n\n
      \n\n
      \n\t
      \n\t\t
      \n\t\t\t
      <\/div>\n\t\t\t