v0.3.3

Highlights

• Faster downscaling during image processing
• VR180 rendering
• nerfacto alpha transparency training
• Sparse-NeRF depth losses
• New template repo for implementing new methods
• Bug fixes

What's Changed

• Add examps to generfacto docs by @terrancewang in #2120
• Fix get_num_frames_in_video by @olegpokhilchenko in #2090
• Update nerfacto.md by @tancik in #2126
• Fix pickling of VanillaDataManager by @jkulhanek in #2129
• Jkulhanek/colmap dataparser by @jkulhanek in #2115
• Update generfacto.md by @tancik in #2132
• Fix to fall back to torch if tcnn not found by @ethanweber in #2124
• allow ns-process-data to use CR2 raw images by @machenmusik in #2119
• Fix unnecessary encodings calculation in tensorf field by @kst179 in #2145
• Remove breakpoint from colmap_dataparser by @jkulhanek in #2141
• Add option to render evaluated images by @hturki in #2135
• Fix typo in nerfacto.py by @MvWouden in #2149
• add build-arg cuda version to dockerfile by @Ilyabasharov in #2139
• Fix use_appearance_embedding for instant-ngp by @brentyi in #1999
• Default to tcnn by @tancik in #2159
• make all SDFFieldConfig params configurable via constructor by @decrispell in #2167
• Correct path notation for Windows in documentation by @KaerMorh in #2171
• Fix image saving format issue in render by @colin-de in #2172
• fix the ZeroDivisionError occured in int(image_width / aspect_ratio) when aspect_ratio equals zero by @WebGao in #2177
• Remove pytorch hash table if tcnn implementation is used by @georghess in #2175
• Feature/run_beta_viewer by @Ilyabasharov in #2157
• Avoid zero duration crash by @machenmusik in #2193
• Fixing Colmap VocabTreeMatching command to support spaces in Filename by @JDihlmann in #2195
• Default fied_heads with instances will cause shared parameters by @hyperzy in #2197
• Pin versions of pyceres+pixsfm, up pycolmap by @skydes in #2199
• Adding input handling for RFFEncoding's get_out_dim(self) by @CardiacMangoes in #2201
• Fix colmap vocab_tree_path (#2195) to work with Windows by @machenmusik in #2204
• Add saving/loading scheduler state_dict by @hyperzy in #2203
• Omnidirectional Stereo Update - Alignment and Compositing by @cvachha in #2217
• Reorient point cloud normals for better surface reconstruction by @NotAGoose in #2220
• Try non-strict model loading on RuntimeError. ... by @sjuxax in #2214
• Add documatations for TensoRF by @Xzzit in #2209
• Fast downscaling for video and image processing (PR2168 plus images) by @machenmusik in #2213
• refactor: get_normalized_directions by @SauravMaheshkar in #2216
• Update torchmetrics by @tancik in #2223
• Remove superfluous -r1 (fix ffmpeg/colmap video) by @SharkWipf in #2231
• Fix arguments that are not passed from config to NerfactoField by @liu115 in #2186
• Minor kwarg fix to be future-compatible with viser version by @kerrj in #2235
• Add std for eval by @Khoa-NT in #2239
• Implement nerfacto alpha transparency training by @nepfaff in #2165
• Load schedulers from ckpt in a backward-compatible manner by @hyperzy in #2236
• support packed samples in semantic renderer by @hturki in #2247
• Render VR180 video by @cvachha in #2261
• nerfbusters doc by @FrederikWarburg in #2208
• Use correct pixel area in volumetric sampler by @hturki in #2268
• ns-render with depth colormap support by @ethanweber in #2125
• Update index.md by @tancik in #2279
• Moved Volinga as an external method by @Frivas97 in #2282
• Add depth_unit_scale_factor to sdfstudio_dataparser by @alexsax in #2285
• Option to process images on GPU for speed at the expense of memory by @AX-I in #2288
• More robust install instructions by @brentyi in #2277
• Remove system-level CUDA install from install instructions by @brentyi in #2289
• Monodepth Regularization by @ethanweber in #2284
• Allow same dimensions optimizations to be disabled by @machenmusik in #2292
• convert raw to JPG not TIFF, to avoid fast-downscale ffmpeg error by @machenmusik in #2295
• Train and eval data splits (Nerfbusters) by @FrederikWarburg in #2207
• PixelSampler Fixes by @AdamRashid96 in #2298
• Add sponsors by @tancik in #2307
• Adding support for PairPixelSampler and PatchPixelSampler masks by @ethanweber in #2305
• template repository by @maturk in #2319
• When generating depth images, use torch_compile with ZoeDepth for compatibility by @machenmusik in #2303
• Rename num_cameras to camera_indices by @maturk in #2326
• Ddp docs by @yurkovak in #2332
• fix: Allow custom user ID by @yurkovak in #2334
• fix: ColmapDataParser downscaling images error. by @onpix in #2337
• fix #2318: add train/test/eval split for instant ngp dataparser by @elenacliu in #2335
• Circle crop to images processing pipeline by @kerrj in #2341
• viser 0.1.0, fix type checker errors by @brentyi in #2343
• Update pixel_samplers.py by @alexsax in #2327
• v0.3.3 by @tancik in #2349

New Contributors

• @olegpokhilchenko made their first contribution in #2090
• @kst179 made their first contribution in #2145
• @Ilyabasharov made their first contribution in #2139
• @KaerMorh made their first contribution in #2171
• @WebGao made their first contribution in #2177
• @georghess made their first contribution in #2175
• @JDihlmann made their first contribution in #2195
• @hyperzy made their first contribution in #2197
• @skydes made their first contribution in #2199
• @CardiacMangoes made their first contribution in #2201
• @NotAGoose made their first contribution in #2220
• @sjuxax made their first contribution in https://github.com/nerfstudio-project/nerfstudio/pu...

v0.3.2

Highlights

• Add nerfacto-huge model
• Add text-to-3d model generfacto
• Support for CPU and M1/M2
• Support for rendering omni-directional stereo video
• Bug fixes

What's Changed

• Update installation docs by @tancik in #1970
• generalize mask collation to other image-like data by @decrispell in #1959
• fix large number of typos in repository by @MvWouden in #1977
• Basic viser frontend by @tancik in #1973
• Add warnings for (slower) Torch 1.x and Windows not supporting torch.compile by @jkulhanek in #1982
• Added at_train_end logic to base pipeline by @maturk in #1932
• Bump socket.io-parser from 4.2.1 to 4.2.3 in /nerfstudio/viewer/app by @dependabot in #1983
• fix visualize encoding bug by @SophusDavid in #1984
• Fix Viewer Recursion: StackOverflow in Server Setup by @lukasHoel in #1987
• Fix custom dataparser docs by @tancik in #1991
• style(ruff): provide better integration with ruff vscode extension by @SauravMaheshkar in #1994
• Fix viewer decorator by @tancik in #1995
• SIGGRAPH 2023 Citation Update by @ethanweber in #1997
• add deepfloyd model by @terrancewang in #1993
• pass grad scaler from trainer to pipeline by @terrancewang in #1996
• Fix exponential decay warmup by @tancik in #2002
• Ability to use vanilla pytorch instead of tcnn for Nerfacto by @ethanweber in #1936
• Fixes wandb runs resuming using WANDB_RUN_ID by @jkulhanek in #2004
• add controls to the beta viewer by @tancik in #2005
• Render Omni-directional stereo video (360 3D VR) by @cvachha in #1986
• Bump tyro by @brentyi in #2008
• Add Metadata Handling to Cameras Collation Function by @MingwuZheng in #2001
• [WIP] Fix model loading collate_fn non-static by @jkulhanek in #1990
• Colab update to Python 3.10 by @akristoffersen in #2017
• Update metashape_utils.py by @henrypearce4D in #2011
• Add antialias parameter as default None by @maturk in #2033
• installation.md: Use https instead of ssh for git clone. by @SharkWipf in #2030
• Add option for model normals on ply export by @CameronFraser in #1965
• Fix Typo on colab file by @Bishadkoju in #2039
• Fix bug when copying images when input_dir = output_dir by @maturk in #2041
• Reduce GPU memory usage in CacheDataloader by @jkulhanek in #1730
• Quick fix for eval script failing for depth-nerfacto (and other methods with a non-standard dataset) by @jMysliwiec in #2047
• Fix metashape multiple camera models error message by @henrypearce4D in #2057
• Add kwargs to dataparser outputs by @maturk in #2053
• nerfstudio dataparser: avoid loading images at inference time by @f-dy in #2051
• fix for scannet dataset camera pose loading by @colin-de in #2056
• colormaps: add "gray" by @f-dy in #2058
• Move out NeRFPlayer: clean files and add method link by @lsongx in #2072
• Update data_utils doc by @tancik in #2076
• nerfacto-huge by @kerrj in #2003
• reduce nerfacto-huge memory by @kerrj in #2082
• Fix data downscale paths by @tancik in #2071
• don't force downscale factor to 1 for nerfacto-huge by @machenmusik in #2088
• Fix resolution rendering logic in statemachine for viewer beta by @kerrj in #2085
• [WIP] dreamfusion implementation by @terrancewang in #1008
• add gradient accumulation support by @ksnzh in #2049
• Add non-trainable camera indices to CameraOptimizer by @maturk in #2080
• Update nerfacto.md by @Shiva-Gantha in #2107
• Add option to process masks on GPU for speed at the expense of memory by @machenmusik in #2110
• Fixing the image_dir if using skip_image_processing by @Khoa-NT in #2111
• Allow newer pycolmap for newer python by @jkulhanek in #2113
• Fix spherical harmonics #2081 by @jkulhanek in #2112
• Fix quickstart doc for remote host viewing setup by @jlightfoot in #2101
• v0.3.2 by @tancik in #2118

New Contributors

• @MingwuZheng made their first contribution in #2001
• @SharkWipf made their first contribution in #2030
• @CameronFraser made their first contribution in #1965
• @Bishadkoju made their first contribution in #2039
• @jMysliwiec made their first contribution in #2047
• @colin-de made their first contribution in #2056
• @Shiva-Gantha made their first contribution in #2107
• @Khoa-NT made their first contribution in #2111
• @jlightfoot made their first contribution in #2101

Full Changelog: v0.3.1...v0.3.2

Code release for SIGGRAPH 2023 paper.

This branch exists for historical purposes. It was used to generate the numbers in the SIGGRAPH paper. We don't plan on integrating this code into the main branch, but it can exist here as a reference or for reproducing our results. We view nerfstudio as an active repo with continual contributions, and some of these changes are undesirable for the main branch (e.g., how we deal with appearance embeddings for evaluation). This is done in an effort to keep our repo extensible for further changes.

Update Aug 29, 2023 - We are deprecating the URL used to host the data from processing MipNeRF-360 with ns-process-data. You can now find it on Hugging Face here. Or, you can replicate the steps described in our paper or use a different MipNeRF-360-specific data parser and use their COLMAP reconstructions directly.

v0.3.1

##Highlights

• Fix compatibility with Windows

What's Changed

• Remove hloc from pyproejct by @tancik in #1955
• Update render docstring by @tancik in #1960
• Backward compatibility for torch 1.x and fix for windows by @jkulhanek in #1962
• Report tcnn import errors by @tancik in #1969
• Fix mask check in PatchPixelSampler by @lukasHoel in #1971
• v0.3.1 by @tancik in #1972

New Contributors

• @lukasHoel made their first contribution in #1971

Full Changelog: v0.3.0...v0.3.1

v0.3.0

Highlights

• Transition to pytorch 2.0 (reinstallation of pytorch and tinycudann is necessary)
• Python 3.8 now minimum version
• Split screen option in viewer
• Camera Controls exposed in python
• Repo now type checked, new PRs will not be merged with type errors
• Pylint -> Ruff

What's Changed

• Wandb/experiment improvements by @ethanweber in #1879
• Make VanillaDataManager generic by @jkulhanek in #1868
• fixing wandb envs by @ethanweber in #1880
• Avoids running test_train.py on external methods by @ethanweber in #1882
• Bump minimum Python version to 3.8 by @brentyi in #1876
• Fix viewer quit-on-train-completion by @tancik in #1885
• Skip scheduler step is opt is not performed by @tancik in #1854
• CropParamsMessage by @takeru in #1886
• Fix mask bug by @Tao-11-chen in #1891
• implementing l1 and tv reg in the original tensorf paper by @AlstonXiao in #1855
• Update docs for ns-viewer by @tancik in #1892
• Fix viewer when opened before train setup finishes by @jkulhanek in #1888
• Render with colormaps by @tancik in #1893
• Fix misc warnings by @tancik in #1894
• Move scripts folder into nerfstudio by @tancik in #1897
• FIx random indices bug by @Tao-11-chen in #1900
• Add split screen by @tancik in #1898
• fix metashape_to_json() if there is no 'components' tag in .xml by @salykovaa in #1906
• #1732 fix sdf dataparser no mono paths by @MvWouden in #1910
• Fix split options by @tancik in #1904
• Implement gradient scaling by @jkulhanek in #1901
• Make rich console global by @tancik in #1916
• Add INGP Docs by @ayaanzhaque in #1774
• extract dataparser configs to a stand-alone file, support custom dataparsers by @wuzirui in #1909
• Improved interpolate render by @Frivas97 in #1911
• Update copyright by @ethanweber in #1917
• fix pytorch 2.0 optimizer no inf check error by @ksnzh in #1919
• Split up render commands by @tancik in #1923
• Update method docs by @tancik in #1925
• Fix static type for dataparser configs by @brentyi in #1924
• Fix bug related to distributed data parallel training by @jb-ye in #1856
• Add button groups to viewer by @tancik in #1905
• Add hints to viewer by @tancik in #1928
• Update contributing docs by @tancik in #1926
• Create PCA colormap for images with channels > 3 by @chungmin99 in #1929
• Discover external methods from CLI by @jkulhanek in #1934
• torchtyping => jaxtyping by @brentyi in #1838
• Type fixes and cleanup by @tancik in #1937
• More typing fixes by @brentyi in #1941
• Switch to ruff by @brentyi in #1939
• Documentation for k-planes add-on by @Giodiro in #1935
• Switching to Torch 2.0 by default. by @jkulhanek in #1922
• bump pyproject torch by @tancik in #1942
• Fix some typing problems by @jkulhanek in #1940
• Update SD for pytorch 2 by @tancik in #1943
• Bump torchvision to match PyTorch by @jkulhanek in #1947
• Drop backward compatibility for Torch 1.13 by @jkulhanek in #1948
• Final typing fixes + add Pyright to CI by @tancik in #1944
• Add pre-commit by @jkulhanek in #1949
• Don't run temporal tv loss test by @tancik in #1952
• output jpg option and progress bar by @takeru in #1896
• Expose viewer controls to pipeline by @kerrj in #1804
• v0.3.0 by @tancik in #1954

New Contributors

• @AlstonXiao made their first contribution in #1855
• @salykovaa made their first contribution in #1906
• @jb-ye made their first contribution in #1856

Full Changelog: v0.2.2...v0.3.0

v0.2.2

TLDR: Mainly Bugfixes

What's Changed

• Docs fixes for viewer elements by @brentyi in #1770
• Add documentation for LERF by @chungmin99 in #1771
• add individual grad_scaler config by @wuzirui in #1762
• Configurable ip address by @tancik in #1779
• Added docs wording change and also --crop-bottom by @THE-COB in #1767
• Update lerf docs by @tancik in #1782
• Improve mask generation script for nuScenes data by @ZiYang-xie in #1679
• Bump webpack from 5.74.0 to 5.80.0 in /nerfstudio/viewer/app by @dependabot in #1783
• Bump decode-uri-component from 0.2.0 to 0.2.2 in /nerfstudio/viewer/app by @dependabot in #1784
• Bump @sideway/formula from 3.0.0 to 3.0.1 in /nerfstudio/viewer/app by @dependabot in #1785
• Automatically reconnect to server by @tancik in #1790
• Use current host in landing page by @machenmusik in #1792
• force resolutions to newer versions for security by @machenmusik in #1799
• Rerender order swap by @kerrj in #1802
• Add trainer parsing by @kerrj in #1801
• Fix element parsing by @kerrj in #1803
• Fix folder labels type annotation by @brentyi in #1805
• Update NGP related models with nerfacc 0.5.2 by @liruilong940607 in #1809
• Render in threejs viewer by @machenmusik in #1794
• Update renderer logs by @tancik in #1817
• Fix profiler for PyTorch 2.0 by @jkulhanek in #1812
• Allow masks with instant ngp dataparser by @nepfaff in #1741
• force resolution of yaml for security by @machenmusik in #1818
• Revert loading entire dataset to GPU if using masks by @nepfaff in #1821
• fix_mask by @Tao-11-chen in #1819
• Add missing .cu files to pyproject by @tancik in #1824
• Refactor of image and video dataset conversion by @MalcolmMielle in #1690
• Update README.md by @akanazawa in #1828
• Auto update viewer version by @tancik in #1825
• Fix typo in readme by @liruilong940607 in #1833
• Reduce the GPU memory required in viewer for NGP methods by @liruilong940607 in #1837
• fix crop_center by @takeru in #1844
• Fix viewer meshing options by @tancik in #1846
• Support equirectangular sensor with metashape by @f-dy in #1841
• Fix typos in Ray Samplers docs by @MvWouden in #1851
• Fix typos in SDF Field by @MvWouden in #1860
• Remove unecessary assert in pixel sampler on image dimensions by @maturk in #1861
• Update logging docs by @tancik in #1863
• Add Tetra-NeRF to the docs by @jkulhanek in #1862
• Tetra-NeRF docs - fix versions and table formatting. by @jkulhanek in #1864
• Fix flag name in tetra-nerf by @jkulhanek in #1870
• Fix torchfield #1749 #1539 by @SophusDavid in #1859
• SceneContraction maps to range [-2, 2], not [-1, 1]. by @decrispell in #1872
• Quit when not using viewer by @tancik in #1873
• Fix images to nerfstudio filename by @MvWouden in #1853
• Remove nerfacc aabb by @tancik in #1874
• Enable method config registration from env variable by @jkulhanek in #1869
• v0.2.2 by @tancik in #1875

New Contributors

• @chungmin99 made their first contribution in #1771
• @wuzirui made their first contribution in #1762
• @ZiYang-xie made their first contribution in #1679
• @nepfaff made their first contribution in #1741
• @MalcolmMielle made their first contribution in #1690
• @takeru made their first contribution in #1844
• @MvWouden made their first contribution in #1851
• @maturk made their first contribution in #1861
• @SophusDavid made their first contribution in #1859

Full Changelog: v0.2.1...v0.2.2

v0.2.1

TLDR

• Bug fixes
• Time controls re-added to viewer

What's Changed

• return unscaled input in positional encoding by @hturki in #1736
• Fix ns-install-cli by @brentyi in #1754
• Fix divide by zero when train util set to 1 by @dylanhu7 in #1757
• Update disconnected state UI by @tancik in #1758
• Delete useless lists of tensors put to CUDA which may cause OOM. by @Tao-11-chen in #1748
• Colab Fix: change localhost to 0.0.0.0 by @akristoffersen in #1759
• Add typing extensions to pyproject by @tancik in #1763
• Add PyTorch profiler to support tracing of forward/backward calls by @jkulhanek in #1761
• Add time controls to viewer by @tancik in #1764
• Add completed training state by @tancik in #1768
• Update host address by @tancik in #1766
• v0.2.1 by @tancik in #1769

New Contributors

• @dylanhu7 made their first contribution in #1757
• @Tao-11-chen made their first contribution in #1748

Full Changelog: v0.2.0...v0.2.1

v0.2.0

Highlights

• Support Custom GUI elements in models. More info here
• Complete rewrite of viewer backend to use Viser
• Add Neus and Neusfacto
• Add Instruct-NeRF2NeRF
• TensoRF speedup (~5x)
• Bug fixes

Known regressions

• Support for modifying time in viewer was removed during the viewer refactor. This functionality will return in in the next version of nerfstudio.

What's Changed

• bug fix for model loading from checkpoint by @terrancewang in #1567
• Updated Colab to Python3.9 and added colmap install speedup. by @akristoffersen in #1587
• Fixed: MinimalDataParser add options, allow Nerfstudio data format to specify test indices by @jkulhanek in #1586
• Adds ns-process-data images tests and fixes #1586 by @jkulhanek in #1591
• Update 360 docs by @tancik in #1593
• NeuS Implementation by @pablovela5620 in #1444
• Add option to use Pixel Perfect SFM by @pablovela5620 in #1554
• feat(fields): create common get_normalized_directions fn by @SauravMaheshkar in #1599
• Adding the crop code made for videos to images by @THE-COB in #1598
• Fix the typeguard version by @tancik in #1612
• TensoRF go brrrrrrr by @mcallisterdavid in #1609
• Adding explicit pytorch version for Cuda 11.7 installs by @julianrendell in #1611
• Prevent colmap from opening by @tancik in #1627
• fix missing param refine-pixsfm by @ksnzh in #1618
• Use the run timestamp as run name on WandB by @iSach in #1619
• Removed deprecated insta360 by @THE-COB in #1621
• Adding NeRF-OSR Dataset by @JADGardner in #1626
• Add option to auto select a websocket port by @ethanweber in #1628
• Adds util function to convert transformed poses back + tests by @jkulhanek in #1594
• improve metashape data import by @f-dy in #1605
• Fixed incorrect ffmpeg command construction by @jackson-waschura in #1645
• Typo in nerfosr validation set by @JADGardner in #1631
• add quality setting; allow png for lossless compression by @machenmusik in #1639
• Check port before loading data by @alvin-xu-5745 in #1633
• realitycapture: allow periods within filename, not only for file extension by @machenmusik in #1649
• misc viewer cleanup by @tancik in #1652
• Add SDFStudio links by @tancik in #1653
• realitycapture nerfacto variable resolution by @machenmusik in #1651
• Render Cell Colab Fix by @akristoffersen in #1655
• Add KPlanesEncoding encoding by @Giodiro in #1658
• Update README.md by @krahets in #1657
• Fix render window size by @tancik in #1659
• Update pixsfm with better param by @HrsPythonix in #1661
• Update realitycapture_utils.py by @henrypearce4D in #1663
• Fix bug: sh render is not correct by @WANGSSSSSSS in #1667
• Update scheduler for nerfacto by @HrsPythonix in #1662
• Update installation.md by @pixel-alex in #1675
• add SO_REUSEADDR option to reuse the port in TIME-WAIT status by @yzslab in #1673
• fix checkpoint generated by ddp training cannot be loaded by @yzslab in #1682
• Exposed number of actual frames to extract with ffmpeg for more transparency by @jackson-waschura in #1683
• Optimize viewer UI by @tancik in #1692
• Update sphinx docs by @tancik in #1695
• Docs for adding new methods by @tancik in #1697
• Update dataparser docs by @tancik in #1698
• Create nerfacto-big config for nerfacto with sampling, network size, resolution, and train steps cranked up by @kerrj in #1668
• Update nerfstudio data defaults by @tancik in #1701
• Select render panel in render view by @tancik in #1693
• Monitor actual gpu memory allocation by @liruilong940607 in #1715
• Remove useless code in camera_paths by @97littleleaf11 in #1717
• Add instruct-nerf2nerf to docs by @tancik in #1711
• add sdf mesh export and texturing by @pablovela5620 in #1620
• update in2n docs based by @ayaanzhaque in #1721
• Add support for torch v2.0.0 by @jkulhanek in #1713
• NeusFacto by @pablovela5620 in #1654
• Respect wandb environment variables by @jkulhanek in #1738
• Update blender web docs by @tancik in #1740
• Viewer backend refactor by @brentyi in #1702
• Update method docs by @tancik in #1742
• Metadata for cameras by @jake-austin in #1623
• Improve viewer element types by @brentyi in #1743
• Update sdfstudio.md by @LownyCGI in #1744
• Adding Blender tutorial video and documentation on NeRF in NeRF composite by @cvachha in #1745
• v0.2.0 by @tancik in #1751

New Contributors

• @julianrendell made their first contribution in #1611
• @JADGardner made their first contribution in #1626
• @jackson-waschura made their first contribution in #1645
• @alvin-xu-5745 made their first contribution in #1633
• @Giodiro made their first contribution in #1658
• @HrsPythonix made their first contribution in #1661
• @henrypearce4D made their first contribution in #1663
• @WANGSSSSSSS made their first contribution in #1667
• @pixel-alex made their first contribution in #1675
• @yzslab made their first contribution in #1673
• @ayaanzhaque made their first contribution in #1721
• @LownyCGI made their first contribution in #1744

Full Changelog: v0.1.19...v0.2.0

v0.1.19

Highlights

• No longer using webrtc, many existing viewer connection issues should now be resolved
• Support for 360 cameras (equirectangular video)
• Add patch sampling
• Add Triplanes
• Bug fixes

What's Changed

• Patch Sampling by @akristoffersen in #1458
• Support multiple versions of stable diffusion by @tancik in #1461
• Equirectangular input by @THE-COB in #1446
• Uniform Sampler in Nerfacto by @iSach in #1464
• Add Windows installation tips to quickstart. by @DreekFire in #1427
• Refactor schedulers by @tancik in #1466
• Update exponential scheduler by @tancik in #1468
• Fix eval dataloader to use eval ray generator by @ethanweber in #1473
• Enable viewer with eval by @tancik in #1474
• [Update] Add hloc to Dockerfile, bump version of docker image in dev-container, improve docs by @Zunhammer in #1478
• Fix datamanager by @tancik in #1481
• Triplanes by @jake-austin in #1456
• Datamanager eval fix by @tancik in #1482
• semantic_nerfw_configurability by @nikmo33 in #1467
• Fix relative argument by @tancik in #1487
• Add extensions tab by @tancik in #1488
• Rename train_split_percentage with train_split_fraction by @97littleleaf11 in #1497
• bump nerfacc to 0.3.5 by @liruilong940607 in #1492
• chore: delete setup.cfg by @SauravMaheshkar in #1495
• ns-process-data: fix image renaming for the --skip-colmap case by @jkulhanek in #1459
• Call load_state_dict on model classes by @jkulhanek in #1499
• Fixes bug introduced in #1499 preventing load_from_checkpoint to work correctly by @jkulhanek in #1501
• feat(docker): Add opencontainers image-spec to Dockerfile by @SauravMaheshkar in #1502
• fixes for nerf synthetic dataset by @liruilong940607 in #1507
• Rename friends dataset to sitcoms3d by @97littleleaf11 in #1503
• Volinga model and NVOL documentation by @Frivas97 in #1486
• Update ffmpeg comment by @tancik in #1505
• Fix logic when using filename datasets by @tancik in #1510
• feat(engine): improve type hints by @SauravMaheshkar in #1509
• Colab tinycudann drive link updated. by @akristoffersen in #1512
• SFM: sfm-based depth, some breadcrumbs for pycolmap by @pwais in #1457
• Check that video path is correct by @tancik in #1513
• Extending ns-render to support loop rendering using interpolated cameras by @ponimatkin in #1508
• Add support for more colmap camera models in ns-process-data by @jkulhanek in #1500
• feat: improve type hints by @SauravMaheshkar in #1515
• Dataparsers for ScanNet and ARKitScenes by @ponimatkin in #1514
• Add weight-decay option to RAdam optimizer by @jkulhanek in #1525
• Bug fix for SHRenderer() by @KokeCacao in #1526
• feat(fields): Improve Type Hints by @SauravMaheshkar in #1524
• Fix pose centering flag for ARKitScenes dataparser by @ponimatkin in #1528
• Fixing data processing script when depth is not used by @ponimatkin in #1527
• Remove dataparser dependency by @tancik in #1529
• Adding conversion to torch.float32 in get_interpolated_poses_many. by @ponimatkin in #1532
• Fix crash when having no eval dataset and not using the viewer by @iSach in #1537
• Update nerf.md by @misza222 in #1546
• Fix camera transformations in ARKitScenes and ScanNet by @ponimatkin in #1544
• Speed up point cloud export by @tancik in #1547
• Update datamager docs by @tancik in #1548
• Avoid loading entire dataset on GPU by @tancik in #1551
• Instant ngp args by @ponimatkin in #1549
• Added video cropping for equirectangular and made it independent of mask in normal videos by @THE-COB in #1542
• Fix typo in docs Nerfology/Model components/Ray samplers by @loicmagne in #1553
• Fix minor scannet bug. by @ponimatkin in #1555
• Added documentation for equirectangular input by @THE-COB in #1561
• Correct instructions for Docker build by @JRombouts in #1557
• Fixed percent crop typo by @THE-COB in #1573
• MinimalDataParser add options, allow Nerfstudio data format to specify test indices by @jkulhanek in #1541
• Changed percent-crop to crop-factor so it's less confusing by @THE-COB in #1575
• Replace default colormap option with turbo for single-channel output by @DreekFire in #1568
• Remove regex for export path by @tancik in #1577
• Patch Sampling bug fix by @akristoffersen in #1578
• scripts/exporter.py: export PLY compatible with PCL by @f-dy in #1572
• Revert "MinimalDataParser add options, allow Nerfstudio data format to specify test indices" by @tancik in #1583
• camera_utils: add "vertical" orientation and "focus" center methods by @f-dy in #1543
• Remove webrtc by @tancik in #1579
• v0.1.19 by @tancik in #1585

New Contributors

• @97littleleaf11 made their first contribution in #1497
• @SauravMaheshkar made their first contribution in #1495
• @Frivas97 made their first contribution in #1486
• @pwais made their first contribution in #1457
• @ponimatkin made their first contribution in #1508
• @KokeCacao made their first contribution in #1526
• @misza222 made their first contribution in #1546
• @loicmagne made their first contribution in #1553
• @JRombouts made their first contribution in #1557
• @f-dy made their first contribution in #1572

Full Changelog: v0.1.18...v0.1.19

v0.1.18

What's Changed

• Update to allow pytorch (and functorch) 1.13.1 by @machenmusik in #1390
• Update SDS loss by @tancik in #1426
• Fix LPIPS data range by @iamNCJ in #1425
• devcontainer for VSCode development by @pablovela5620 in #1428
• Improve ns-install-cli by @jkulhanek in #1396
• Add arxiv link by @tancik in #1434
• add monosdf normal loss and midas scale invariant depth loss by @pablovela5620 in #1410
• Proposal network initial sampler by @tancik in #1435
• Allow use of black or white background in Nerfacto models by @iSach in #1441
• Add ns-viewer docs by @tancik in #1450
• Bump Dockerfile versions, add user to sudoers list. by @Zunhammer in #1448
• Fix mesh export normals by @tancik in #1452
• Fix depth nerfacto exporting by @tancik in #1453
• Fix colmap bug when not all images are matched by @tancik in #1454
• v0.1.18 by @tancik in #1455

New Contributors

• @iSach made their first contribution in #1441

Full Changelog: v0.1.17...v0.1.18