[test]: Perform final test to visualize rendered spiral path

[a-lemus96]

Run a test using any scene from LLFF dataset and visualize results.

[a-lemus96]

Will run a quick test, hoping that GPU devices have enough memory.

[a-lemus96]

Again got CUDA error. Not sure if it is bc someone is using the device as I am not able to use nvidia-smi command. Will try by reducing batch_size.

[a-lemus96]

Got the following error when using 5k validation rate:

[NeRF]:  10%|██████████▍                                                                                             | 5000/50000 [14:45<2:12:45,  5.65it/s]
Traceback (most recent call last):
  File "/home/lemus/projects/fs-nerf/src/run-nerf.py", line 590, in <module>
    main()
  File "/home/lemus/projects/fs-nerf/src/run-nerf.py", line 484, in main
    train(
  File "/home/lemus/projects/fs-nerf/src/run-nerf.py", line 350, in train
    rgb, depth = R.render_frame(
                 ^^^^^^^^^^^^^^^

[a-lemus96]

Rendered video for spiral path consists of completely dark frames. Running a test to see if validation images during training exhibit this behavior.

[a-lemus96]

This is the result for 5k iterations

so the problem must be in path computation

[a-lemus96]

Path poses shape is incorrect: Path poses shape: torch.Size([1, 3, 5]). Create an issue to fix this semantic error.

[a-lemus96]

Now that bug has been handled, tests are being performed to evaluate final rendered result.

[a-lemus96]

Finally, rendered videos for 5k iterations using NeRF

RGB:

https://github.com/a-lemus96/fs-nerf/assets/95151624/30ed41e9-2ef0-4b05-8114-4158f203d5eb

[a-lemus96]

Finally, rendered videos for 5k iterations using NeRF

Depth:

https://github.com/a-lemus96/fs-nerf/assets/95151624/3f5bcda2-010a-4813-89b7-adc1e7127eb0