GitLab

* multres grid

* prop

* benchmark with prop and occ

* benchmark blender with weight_decay

* docs

* bump version

* Reduce conversion. Assume that ray_indices should always be long and pack_info always int

* Remove some more casting

* More casting fixes

* Remove unecessary long casting

* Fix changes in precision at python level, ray_pdf_query test is already broken since it was reverted

* Support long when rendering transmittance

* Support long

* Woops

- Faster rendering function via nvidia-cub, shipped with cuda >= 11.0 (Require >=11.6 for out use). ~10% speedup
- Expose transmittance computation.

* support alpha for both marching and rendering; relax pytorch requirement

* bump version

* zeros -> emtpy in cuda

* disable cuda hint if is been built

* update examples with better perf

* bump version to 0.2.0

* update readme

* update index and readm

* clean up doc

* more organized version

* add copyrights

* proper versioning

* proper versioning

* cleanup project; read version from pyproject

* read version from pyproject

* update weblink

* cleanup nerfacc file structure

* propoer test contraction

* proper test for intersection

* proper tests for pack, grid, and intersection

* proper testing for rendering

* bug fix

* proper testing for marching

* run check reformat

* add hints on nvcc not found

* add doc to readme

* resume github check

* update readthedocs env

* rm tool.setuptools.packages.find

* seems working

* contraction func in cuda

* Update type

* More type updates

* disable DDA for contraction

* update contraction perfom in readme

* 360 data: Garden

* eval at max_steps

* add perform of 360 to readme

* fix contraction scaling

* tiny hot fix

* new volrend

* cleanup ray_marching.cu

* cleanup backend

* tests

* cleaning up Grid

* fix doc for grid base class

* check and fix for contraction

* test grid

* rendering and marching

* transmittance_compress verified

* rendering is indeed faster

* pipeline is working

* lego example

* cleanup

* cuda folder is cleaned up! finally!

* cuda formatting

* contraction verify

* upgrade grid

* test for ray marching

* pipeline

* ngp with contraction

* train_ngp runs but slow

* trasmittance seperate to two. Now NGP is as fast as before

* verified faster than before

* bug fix for contraction

* ngp contraction fix

* tiny cleanup

* contraction works! yay!

* contraction with tanh seems working

* minor update

* support alpha rendering

* absorb visibility to ray marching

* tiny import update

* get rid of contraction temperture;

* doc for ContractionType

* doc for Grid

* doc for grid.py is done

* doc for ray marching

* rendering function

* fix doc for rendering

* doc for vol rend

* autosummary for utils

* fix autosummary line break

* utils docs

* api doc is done

* starting work on examples

* contraction for npg is in python now

* further clean up examples

* mlp nerf is running

* dnerf is in

* update readme command

* merge

* disable pylint error for now

* reformatting and skip tests without cuda

* fix the type issue for contractiontype

* fix cuda attribute issue

* bump to 0.1.0
Co-authored-by: Matt Tancik <tancik@berkeley.edu>