Jeff / Jianfei Guo

Yes! And ngp is already being noticed the day it's released. A new CUDA-optimized and better-designed version of nerf-- & neurecon is already on its way! 😄

Hi @DeXtmL, The clip model is in the implementation of OPENAI-CLIP, which is located at https://github.com/openai/CLIP You can install it via their intructions. Sorry that I haven't got time to...

I believe it is claimed in section 3.1 in the original pi-gan paper that the gammas and betas should be different for different SIREN blocks. They use the i-th subscripts...

I had the same issue; Intersections when ray_o inside AABBs is needed and useful: 1. consider dealing with a large wild scene (e.g. streets of cities) with an observer (e.g....

Hi @Caenorst , Cooool! I've done some testing, In most of the case, it works; also, the time consumed by `unbatched_raytrace` dropped from 7.1 ms to 6.1 ms in my...

Hi @ootts , Apologies for the delayed response. This is simply a hyperparameter that amplifies the gradients of the learnable parameter. The paper merely asserts that this parameter is learnable,...

Hi @akshatdave , It might be very late, but I have developed a new and simpler normalization method for the BMVS dataset, which comes with some descriptions: https://github.com/PJLab-ADG/neuralsim/tree/main/dataio/bmvs#normalization I have...

This is also what I'm looking for! Especially for the scaling parameters/settings. Done some tedious research to align shapenet v2 models with original SRN datasets, but there are some weird...

@psguo As for the sphere radius, it can be infered from the poses. For example, for chairs_train, `np.linalg.norm(c2w[...,:3,3], axis=-1)` is constant at `1.3`, so the sphere radius is `1.3`. And...

UPDATE: I've found how the shapenet models are scaled to render the original SRN datasets!!!!!!! - Quick answer: after the models are imported, I believe Dr. Sitzmann scales them again...