Open cduguet opened 6 months ago
Great job with it! I am trying to make it run on a custom dataset, that I've made with colmap. So far the relevant files are only :
$ tree . . ├── distorted │ ├── images ... │ └── sparse ... ├── images │ ├── test53_sax_cam01_1_1.jpg │ ├── test53_sax_cam01_1_10.jpg │ ├── test53_sax_cam01_1_11.jpg ... ├── sparse │ └── 0 │ ├── cameras.bin │ ├── images.bin │ ├── points3D.bin │ └── points3D.ply
The images look:
Images and dataset has been undistorted More details about the model:
$ colmap model_analyzer --path sparse/0/ Cameras: 1 Images: 224 Registered images: 224 Points: 11203 Observations: 114430 Mean track length: 10.214228 Mean observations per image: 510.848214 Mean reprojection error: 1.562057px
Executed:
python train.py -s /data/datasets/sax/test-jpg/ --eval -m /data/datasets/sax/test-jpg/gshader/ -w --brdf_dim 0 --sh_degree -1
but all renders end up being black after the render script, and even the files, and the file sizes are conspicuously small:
$ ls -lashrt input.ply 296K -rw-r--r-- 1 root root 296K Feb 28 19:57 input.ply
$ ls -lashrt brdf_mlp/iteration_30000/brdf_mlp.hdr 40K -rw-r--r-- 1 root root 39K Feb 28 21:32 brdf_mlp/iteration_30000/brdf_mlp.hdr
Am I missing something fundamental?
Great job with it! I am trying to make it run on a custom dataset, that I've made with colmap. So far the relevant files are only :
The images look:
Images and dataset has been undistorted More details about the model:
Executed:
but all renders end up being black after the render script, and even the files, and the file sizes are conspicuously small:
Am I missing something fundamental?