Open Michaelwhite34 opened 2 years ago
Also, after hybrid_optimization should I just run the code "python render_video_from_img.py --path=/PATH_TO/checkpoint_train.pth --output_dir=/PATH_TO_WRITE_TO/ --img_path=/PATH_TO_IMAGE/ --curriculum="celeba" or "carla" or "srnchairs"" ?
Hi,
I tested it with a 12GB GPU. Perhaps try decreasing max_batch_size
? The rendered images should be saved to output_dir
, so there is no need to run any additional script.
Best.
Hi,
I tested it with a 12GB GPU. Perhaps try decreasing
max_batch_size
? The rendered images should be saved tooutput_dir
, so there is no need to run any additional script.Best.
The resolution of output image is very low, can we have higher resolution output? Also it only gives a image, how to get the interpolation video ?
Hi,
The output resolution can be changed by specifying a higher "img_size" in "render_options": https://github.com/primecai/Pix2NeRF/blob/ee036fe5592c65dd6cf57ee8486a242ce4e25c08/hybrid_optimization.py#L63
Note that we train the model with 64x64 images, so naively setting a resolution very high may not work very well.
Regarding interpolation video, some coding is required to merge hybrid_optimization.py
and render_video_from_img.py
. I will find a chance to merge these two scripts.
Best.
RuntimeError: CUDA out of memory. Tried to allocate 768.00 MiB (GPU 0; 6.00 GiB total capacity; 3.51 GiB already allocated; 391.91 MiB free; 3.57 GiB reserved in total by PyTorch)