torch_efficient_distloss

torch_efficient_distloss copied to clipboard

I want to train on custom data. I have the poses, but I am unable to get the disp directory for training

i-nkrumah

RuntimeError: Error building extension 'segment_cumsum_cuda'

2

Hi ! I'm trying to run [Block-NeRF](https://github.com/sjtuytc/UnboundedNeRFPytorch) and I faced this error: ```bash Using /root/.cache/torch_extensions/py310_cu116 as PyTorch extensions root... Detected CUDA files, patching ldflags Emitting ninja build file /root/.cache/torch_extensions/py310_cu116/segment_cumsum_cuda/build.ninja... Building...

IaroslavS

It is really helpful and such a great work! Wondering the principle behind the optimization :)

2

I've read the code about `eff_distloss_native()`, but I could not really understand how it works :( Can you please put on some derivations so that we could more understand the...

TIMESTICKING


lhc991025

Thank you for this great implementation of distortion loss. I ran into this issue when compiling your extension. ```shell \lib\site-packages\torch\include\pybind11\cast.h(624): error: too few arguments for template template parameter "Tuple" detected...

tobias-kirschstein

interval value for last point sampled on ray?

2

Hi, What do you set the interval value to for the last sampled point on each ray. Using a large term (1e10) as a substitute for infinity make the loss...

hturki

I fond a core dumped issue when using this library:  .

sjtuytc

add an einsum_distloss implementation

4

Spark001