bebeal / mipnerf-pytorch

A re-implementation of mip-NeRF in PyTorch
146 stars 15 forks source link

Modifications might be needed to run the code #6

Open asjmasjm opened 1 year ago

asjmasjm commented 1 year ago

I followed the instructions from the README.md, ran the code and it turned out that the compiler complained about some kind of python grammatical errors as "list indices must be integers or slices, not tuple". The compiler also pinned the problem source to the function generate_rays() defined at the 202th line in the datasets.py.

I changed camera_directions = [v @ p2c[:3, :3].T for v, p2c in zip(pixel_directions, self.pix2cam)] directions = [v @ c2w[:3, :3].T for v, c2w in zip(camera_directions, self.cam_to_world)] origins = [np.broadcast_to(c2w[:3, -1], v.shape) for v, c2w in zip(directions, self.cam_to_world)] to camera_directions = [v @ np.array(p2c)[:3, :3].T for v, p2c in zip(pixel_directions, self.pix2cam)] directions = [v @ np.array(c2w)[:3, :3].T for v, c2w in zip(camera_directions, self.cam_to_world)] origins = [np.broadcast_to(np.array(c2w)[:3, -1], v.shape) for v, c2w in zip(directions, self.cam_to_world)] then it seemed everything was alright. I am not sure whether this is caused by the python version mismatch or something like this.

DrawingProcess commented 7 months ago

I have a same issue. Thank you for issue taking.