Mingqiao Ye

Hi, you can add an argument `--skip_test` in the command. This problem occurs since here all images are used for training and there is no test image or label. Your...

Hi, thank you very much for that problem! Yes we need curly brackets here and I have corrected the code.

Hi, it looks like you do not load the correct object mask path. You can add a breakpoint [here](https://github.com/lkeab/gaussian-grouping/blob/main/scene/dataset_readers.py#L98) and check `os.path.exists(object_path)`.

Hi, Since the error mentions CUDA (cudart.lib), you can check if you have the correct version of the CUDA Toolkit installed, as required by GroundingDINO. Compatibility between CUDA, PyTorch, and...

Hi, this problem comes from too much shared data for rasterization. You can reduce the [block size](https://github.com/lkeab/gaussian-grouping/blob/main/submodules/diff-gaussian-rasterization/cuda_rasterizer/config.h#L15) to avoid this problem. Note that you also need to change the corresponding...

Hi, convert.py is used in [original Gaussian Splatting](https://github.com/graphdeco-inria/gaussian-splatting) for converting custom dataset. Its usage of COLMAP only supports undistorted datasets (PINHOLE or SIMPLE_PINHOLE cameras). You may refer to some of...

Hi, for illegal memory access problem, maybe you can refer to [this issue](https://github.com/graphdeco-inria/gaussian-splatting/issues/41).

You can also use other online viewer like [this one](https://github.com/antimatter15/splat).

Hi, thanks for watching our work! We expect to release the code for inpainting around January 5th.

Hi, we get this object id from the training label in `data/bear/object_mask` folder. You can check the pixel value of the label map for corresponding object. Another way for text-prompt,...