tink/nerfstudio
4
0 Fork 0
mirror of https://github.com/nerfstudio-project/nerfstudio.git synced 2026-11-12 02:45:34 +08:00
Code Activity
1,875 Commits 85 Branches 40 Tags
a8b3f71eeb3dca02012e9d2bd1722a708b43c59f
Commit Graph
1875 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Sacha LewinandBrent Yi a8b3f71eeb Full implementation of features for dynamic scenes into the Viser viewer. (#2936) 
* Add time buttons & export render_time

* Added preview of dynamic scenes, time interpolation for rendering, etc. (Fixes #2934)

* Fixes to comply with Ruff

* Fixed UnboundLocalError in static mode

* Formatting

* Fixed pyrights compliance

* Nit remove unused sleep

* Fixed JSON export/import for dynamic camera paths

* Fixed fov override check for loading json camera paths

---------

Co-authored-by: Brent Yi <yibrenth@gmail.com>
 2024-02-21 00:02:05 +00:00
Justin Kerr 36c48c7177 Update project, rasterize with block_width interface (#2937) 
* new interface for block width in gsplat
 2024-02-20 22:09:55 +00:00
Pierre-Alain Langlois 59d1097962 Fixed ns_render dataset bug in splatfacto (#2917) 
* Fixed ns_render dataset bug in splatfacto
 2024-02-20 09:19:05 -08:00
Sacha Lewin 6c2bfad931 Make time slider effective in viewer (Fixes #2928) (#2931) 2024-02-19 14:42:38 -08:00
Piero ToffaninandChung Min Kim 45db2bcfab Call rescale_output_resolution before early return (#2897) 
* Call rescale_output_resolution before early return

Co-authored-by: Chung Min Kim <chungmin99@gmail.com>
 2024-02-13 23:12:18 +00:00
Justin Kerr 7f9def1ff0 Switch default matching method for videos to 'sequential' in COLMAP (#2912) 
default matching method for video to sequential for COLMAP
 2024-02-13 20:26:17 +00:00
J.Y eb9bcb577d Implement new rasterization mode using opacity compensation factor (#2888) 
Implement new rasterization mode for tiny gaussians
 2024-02-13 11:43:04 -08:00
Sebastiaan 4f798b23f6 Fix an issue where gradient accumulation could not be passed as argument due to a type error. (#2910) 2024-02-13 18:09:59 +00:00
Pierre-Alain Langlois 57fbc07bf1 Add transparency carving to splatfacto (#2889) 
Implemented alpha training in splatfacto
 2024-02-12 15:31:42 -08:00
Ling Jing 1b4b682bce add zipnerf-pytorch as external method (#2850) 
* add zipnerf-pytorch as external method
 2024-02-12 22:10:14 +00:00
Gunjan Chhablani ce5adbb7d7 Fix benchmarking script camera optimizer argument (#2899) 2024-02-10 16:00:59 -08:00
Tobias Fischer efa91e5df7 [Fix] convert camera_idx to tensor (#2890) 
convert camera_idx to tensor
 2024-02-09 17:53:34 +02:00
hturki 1d070f5625 support generic TDataset in paralleldatamanager (#2884) 
* support generic TDataset in paralleldatamanager

* run black

* run ruff

* run ruff
 2024-02-07 16:50:15 -08:00
Frédéric DevernayandFrédéric Devernay 66e119485d Remove useless call to .with_suffix(""), which causes a bug if filename contains dots (#2882) 
* Remove useless call to .with_suffix(""), which causes a bug if filename contains dots
Fixes https://github.com/nerfstudio-project/nerfstudio/issues/2881

---------

Co-authored-by: Frédéric Devernay <f-dy@users.noreply.github.com>
 2024-02-07 11:16:40 -08:00
Chung Min Kim dc53517565 Add default depth output when no gaussians visible (#2874) 
* Add default depth output when no gaussians visible

* Add accumulation to default output return

* Change to using new_ones/zeros
 2024-02-07 00:21:21 +00:00
Brent Yi 6127ec0f72 Update viewer legacy docs (#2875) 2024-02-06 16:07:37 -08:00
Ethan Weber af2072bf7c Enable always returning depth for Gaussian Splatting (#2873) 
enable always returning depth, if specified in the config
 2024-02-06 10:48:20 +02:00
Congrong Xu 52cda21186 Add multi-processing support for cache_image. Fixed the issue of insufficient precision leading to a completely black image when converting from RGBA to RGB. (#2867) 2024-02-05 17:04:14 +00:00
Too Yuen Min 2046cfe9fb fix: pixelated downscaled images (#2866) 2024-02-05 08:57:14 -08:00
Matthew Tancik ffe3c31c13 v1.0.1 (#2865) v1.0.1 2024-02-01 17:00:02 -08:00
Brent Yi 6436742c34 Upper-bound torch version for dev installs only (#2864) 2024-02-02 00:19:54 +00:00
Piero Toffanin c5a400c8c9 Fix ODM data processor coordinate system (#2847) 
* Fix ODM data processor coordinate system
 2024-02-01 10:34:19 -08:00
Simon Bethke 10c6f29451 Changed torch version to <2.2 (#2861) 
* Changed torch version to <=2.2, as <=2.1 will uninstall torch 2.1.2+cu118 and leave users without cuda in torch

* changed <=2.2 to <2.2 to prevent the issue at to prevent the type error referenced here https://github.com/nerfstudio-project/nerfstudio/pull/2857

* Changed torch version to <=2.2, as <=2.1 will uninstall torch 2.1.2+cu118 and leave users without cuda in torch
 2024-02-01 11:06:17 +00:00
Justin Kerr 0e01a90bba Replace depth alpha mask logic with torch.where for gradients in splatfacto (#2856) 
* replace alpha depth logic with torch.where for differentiability
 2024-01-31 14:03:34 -08:00
J.YandJustin Kerr de36210ae5 Avoid OOM for caching undistorted images of splatfacto (#2855) 
Co-authored-by: Justin Kerr <justin.g.kerr@gmail.com>
 2024-01-31 13:41:20 -08:00
mattstrong-stanford ebbebdab4c Have predicted depth as expected_depth so gradients can be computed (#2852) 
Update depth_nerfacto.py
 2024-01-31 19:48:08 +00:00
Justin Kerr 17748a8309 limit torch to <=2.1 to avoid type errors (#2857) 2024-01-31 19:34:00 +00:00
Jonáš Kulhánek 3a90cb529f Fix #2843 - safer typing (#2848) 2024-01-30 11:06:17 -08:00
Jonáš Kulhánek 4779f816d6 Add flag to disable apperance embedding (#2843) 2024-01-29 23:07:37 +00:00
Ethan WeberandJustin Kerr c397588300 Mask flexibility for methods than need to sample in the masked region (#2647) 
* minor changes to make masking more flexible

* formatting

* updating ignore mask

---------

Co-authored-by: Justin Kerr <justin.g.kerr@gmail.com>
 2024-01-29 11:47:49 -08:00
J.Y d8b517d6a5 Add average_init_density to improve robustness of nerfacto training (#2834) 2024-01-28 14:09:27 -08:00
Piero Toffanin fb1fee1d8f Add support for loading 3D points with ODM data processor (#2838) 
* Add support for loading 3D points with OpenDroneMap data processor
 2024-01-28 11:09:31 -08:00
psetinek f9a5227828 External Method: seathru_nerf (#2829) 
* add seathru_nerf as external method
 2024-01-28 11:01:46 -08:00
Congrong Xu 3f23e6efb2 Fix the background color bug when exporting video with the "random" background option (#2831) 
patch #2815
 2024-01-27 11:27:49 -08:00
Brent Yi 1bc5b424b4 Update viewer docs (#2826) 
* Update viewer docs

* Typo
 2024-01-26 22:30:50 -08:00
Anthony Zhang 1aba4ea7a2 fixing colmap installation to use apt-get since gdown on the URL unable to download colmap (#2823) 
fixing colmap installation to use apt-get since gdown on the URL was failing
 2024-01-25 23:09:23 +00:00
Piotr Sokólski 67b1005784 Fix "mutable default" error in render.py (#2825) 
This fixes 

```
ValueError: mutable default <class 'nerfstudio.data.scene_box.OrientedBox'> for field obb is not allowed: use default_factory
```
 2024-01-25 15:04:14 -08:00
Justin Kerr f2f4bf55b2 Add alpha rendering to splatfacto, fix depth bug (#2824) 
* Add alpha rendering to splatfacto, output as accumulation
 2024-01-25 12:51:16 -08:00
Brent Yi 3bade3a9dc Remove viewer build badge (#2819) 2024-01-24 23:11:22 -08:00
Chung Min Kim 3db08019db Use packaging for colmap version (#2814) 
* Use packaging.version for colmap version

* Run ruff

* Change default_version to str type
 2024-01-24 16:02:05 -08:00
Brent Yi fb4f9f0060 Fix mask shape for Fisheye624 cameras (#2813) 2024-01-23 16:08:04 -08:00
hturki 365592036e Have Mill19 dataset use official splits by default (#2745) 
add splits to mill19
 2024-01-23 23:51:58 +00:00
VasuAgrawal 7cc27da7ff Make downloading more windows-friendly (#2810) 
* Make downloading more windows-friendly

* Remove unnecessary f-strings
 2024-01-23 15:37:05 -08:00
Ikko Eltociear Ashimine 0dfefbf922 Fix typo in parallel_datamanager.py (#2812) 
occured -> occurred
 2024-01-23 20:56:01 +02:00
Huang Binxiao b669f51f02 Update splatfacto.py (#2804) 
Shape error: The shape of batch['mask'] is [H, W, 1], not[H, W]
 2024-01-23 10:16:22 -08:00
VasuAgrawal 5988c7ee7c Add documentation on existing datasets (#2806) 2024-01-23 09:54:13 -08:00
J.YandJianbo Ye 8952fb531e important fix typo (#2800) 
fix typo

Co-authored-by: Jianbo Ye <jianboye@amazon.com>
 2024-01-19 21:15:00 -08:00
Brent Yi 19a21d7773 Bump version to 1.0.0 (#2799) 
Update pyproject.toml
v1.0.0 		2024-01-19 18:15:02 -08:00
Congrong XuVasu AgrawalEthan WeberBrent YiJustin Kerr
 		c83ed7dcaf fix the bugs associated with blender datas (#2704) 
* fix the bug when camera.distortion_params is None

* Handle background color override when using blender.

* fix bare except

* format

* Update background color override in Blender dataparser

* Add ability to download EyefulTower dataset

* wip before I copy linning's stuff in

* Generate per-resolution cameras.xml

* Generate transforms.json at download

* Fix a couple of quotes

* Use official EyefulTower splits for train and val

* Disable projectaria-tools on windows

* Fix extra imports

* Add a new nerfacto method tund for EyefulTower

* Split eyefultower download into a separate file

* Fix typo

* Add some fisheye support for eyeful data

* Reformatted imports to not be dumb

* Apparently this file was missed when formatting originally

* Added 1k resolution scenes

* revert method_configs.py to original values

* Also add 1k exrs

* Add option to modify bg color in gaussian splatting

* fix back the config, bg color should work now

* removed camera optimizer for gs to align with main

* Address feedback

* Revert changes to pyproject.toml, to be added in a later PR

* Oops, probably shouldn't have gotten rid of awscli ...

* adding support for bg color, tested and reformatted now

* formatted

* formatted

* changed glob variable name

* Refactor background color variable name

* prevent viser overriding

---------

Co-authored-by: Vasu Agrawal <vasuagrawal@meta.com>
Co-authored-by: Ethan Weber <ethanweber@berkeley.edu>
Co-authored-by: Brent Yi <yibrenth@gmail.com>
Co-authored-by: Justin Kerr <justin.g.kerr@gmail.com>
 2024-01-20 02:10:06 +00:00
Brent YiandJustin Kerr ff4002d343 Clean up ns-train {method} --help for not-yet-installed external methods (#2760) 
* Clean up `ns-train {method} --help` for not-yet-installed external methods

* Ruff

* Ruff

* add clearer print statement

* Types? Not sure if this fixes it

---------

Co-authored-by: Justin Kerr <justin.g.kerr@gmail.com>
 2024-01-19 18:02:56 -08:00