TQTQliu
commented
3 months ago Hi, thanks for your interest.
I just checked this problem and found that it is because the Gaussian splatting code was updated a few days ago. The updated rasterizer function will return three values instead of two, so this error ValueError: too many values to unpack (expected 2) occurred.
There are two solutions:
1) You can modify this line, from rendered_image, _ = rasterizer( to rendered_image, _, _ = rasterizer( to avoid this error.
2) To prevent gaussian splatting from being updated further and resulting in incompatibility with our code, we upload the submodules folder. Therefore, you can also re-clone our repository and then install the diff-gaussian-rasterization and simple-knn with the following commands:
pip install lib/submodules/diff-gaussian-rasterization
pip install lib/submodules/simple-knnYou can also find updated instructions in readme.
kgkgkgjb
Hi, thank you very much for your contribution!
python run.py --type evaluate --cfg_file configs/mvsgs/colmap_eval.yaml test_dataset.dataroot examples/scene1 When I run this command, an error occurs： Traceback (most recent call last): File "run.py", line 98, in
globals()['run ' + args.type]()
File "run.py", line 78, in run_evaluate
output = network(batch)
File "/opt/conda/envs/mvsgs/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1194, in _call_impl
return forward_call(*input, **kwargs)
File "lib/networks/mvsgs/network.py", line 111, in forward
render_novel_i_0 = render(batch[f'novel_view{i}'], b_i, world_xyz[b_i], color_out[b_i], rot_out[b_i], scale_out[b_i], opacity_out[b_i], bg_color=cfg.mvsgs.bg_color)
File "/workspace/MVSGaussian/lib/gaussian_renderer/init.py", line 61, in render
cov3D_precomp=None)
ValueError: too many values to unpack (expected 2)