search

eckertzhang / Text2NeRF

Official implementation of 'Text2NeRF: Text-Driven 3D Scene Generation with Neural Radiance Fields'
MIT License
123 stars 6 forks source link

360 degree scene #5

Open NikitaVasilevN opened 9 months ago

NikitaVasilevN commented 9 months ago

Hello, I tried to run your model to generate a 360 degree scene, but I faced with a problem. After 850 epochs, the following error appears Epoch 851 lr0 = 0.02 lr1 = 0.001 1%|█▍ | 1/103 [00:00<01:27, 1.16it/s] Traceback (most recent call last): File "/workspace/t2n_docker_conda/Text2NeRF/text2nerf_main.py", line 672, in reconstruction(args) File "/workspace/t2n_docker_conda/Text2NeRF/text2nerf_main.py", line 556, in reconstruction rgb_map, alphas_map, depth_map, weights, z_vals = renderer(rays_train, tensorf, chunk=args.batch_size, File "/workspace/t2n_docker_conda/Text2NeRF/renderer.py", line 42, in OctreeRender_trilinear_fast return torch.cat(rgbs), None, torch.cat(depth_maps), torch.cat(weights), torch.cat(z_val) RuntimeError: torch.cat(): expected a non-empty list of Tensors

yezifeiafei commented 9 months ago

I got the same error.

eckertzhang commented 9 months ago

Sorry for the inconvenience. Please change the 'pose_traj=circle0' to 'pose_traj=circle'. Here, 'circle0' was used for early code testing and has been discarded.