Gsplat

Highlights
* support 2DGS
* Camera model: perspective, orthogonal, fish-eye
* Start to support pre-built wheels (python 10)

What's Changed
* Update build workflow by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/369
* fix (apparant) typos in compression docs by christian-helms in https://github.com/nerfstudio-project/gsplat/pull/370
* eval supports multi-GPU by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/371
* correct a typo to fix 373 by jckhng in https://github.com/nerfstudio-project/gsplat/pull/383
* Add option to render camera trajectory as an ellipse. by jefequien in https://github.com/nerfstudio-project/gsplat/pull/380
* Add an option to disable world space normalization in examples/simple_trainer.py by JojiJoseph in https://github.com/nerfstudio-project/gsplat/pull/379
* Fix points projection when sampling patches for training by nnop in https://github.com/nerfstudio-project/gsplat/pull/386
* fix reshape error when multi GPU bachsize > 1 not packed by Ultized in https://github.com/nerfstudio-project/gsplat/pull/382
* Fix typo by cr333 in https://github.com/nerfstudio-project/gsplat/pull/392
* Make download script work on Windows, too by cr333 in https://github.com/nerfstudio-project/gsplat/pull/397
* Fused differentiable SSIM by rahul-goel in https://github.com/nerfstudio-project/gsplat/pull/396
* 2DGS by FantasticOven2 in https://github.com/nerfstudio-project/gsplat/pull/208
* citation.bib by maturk in https://github.com/nerfstudio-project/gsplat/pull/400
* Implement Bilateral Guided Gaussian Splatting. by jefequien in https://github.com/nerfstudio-project/gsplat/pull/291
* Bump fused-ssim commit by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/405
* Update rendering.py by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/407
* Update CITATION.bib by cr333 in https://github.com/nerfstudio-project/gsplat/pull/410
* Make commands easier in README by MrNeRF in https://github.com/nerfstudio-project/gsplat/pull/420
* Implement Fisheye-GS. by jefequien in https://github.com/nerfstudio-project/gsplat/pull/398
* Publish precompiled wheels in Github release and add wheel index by martinResearch in https://github.com/nerfstudio-project/gsplat/pull/365
* 1.4.0 by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/421
* fix workflow by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/422
* Workflow fix by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/423

New Contributors
* christian-helms made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/370
* jckhng made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/383
* JojiJoseph made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/379
* nnop made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/386
* Ultized made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/382
* cr333 made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/392
* rahul-goel made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/396
* FantasticOven2 made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/208
* MrNeRF made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/420
* martinResearch made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/365

**Full Changelog**: https://github.com/nerfstudio-project/gsplat/compare/v1.3.0...v1.4.0

Highlight
- Support 3DGS Compression
- Support orthographic projection
- Support TanksAndTemples Dataset
- Various Bug Fix


What's Changed
* Make sure the build directory exists. by keunhong in https://github.com/nerfstudio-project/gsplat/pull/319
* docs: update README.md by eltociear in https://github.com/nerfstudio-project/gsplat/pull/322
* Fix typo in examples/simple_viewer.py by simo23 in https://github.com/nerfstudio-project/gsplat/pull/321
* Consolidate simple_trainer.py and simple_trainer_mcmc.py by brentyi in https://github.com/nerfstudio-project/gsplat/pull/325
* Compressing gaussian splats by jefequien in https://github.com/nerfstudio-project/gsplat/pull/309
* Simplify `tyro` usage in `simple_trainer.py` by brentyi in https://github.com/nerfstudio-project/gsplat/pull/332
* Cleanup C/CUDA code for C users by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/333
* Patch to make Strategy compatible with splatfacto by kerrj in https://github.com/nerfstudio-project/gsplat/pull/334
* Fix minor typo by zhiguotang5 in https://github.com/nerfstudio-project/gsplat/pull/342
* correct distCoeffs of SIMPLE_RADIAL by Master-cai in https://github.com/nerfstudio-project/gsplat/pull/337
* Add np.random.choice fallback for many Gaussians exceeding torch.multinomial limits by soskek in https://github.com/nerfstudio-project/gsplat/pull/338
* data parser for t&t dataset by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/347
* Added orthographic projection by VladislavZavadskyy in https://github.com/nerfstudio-project/gsplat/pull/349
* Fix bug related to refine_scale2d and add pause_refine_after_reset to default strategy by jb-ye in https://github.com/nerfstudio-project/gsplat/pull/354
* tt compression script by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/357
* Update to 1.3.0 by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/366

New Contributors
* keunhong made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/319
* simo23 made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/321
* zhiguotang5 made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/342
* Master-cai made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/337
* soskek made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/338
* VladislavZavadskyy made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/349

**Full Changelog**: https://github.com/nerfstudio-project/gsplat/compare/v1.2.0...v1.3.0

Highlight

Multi-GPU distributed rasterization is supported! E.g. 4 GPUs could lead to more than 3x speedup as well as 3x less memory usage at the same time.

What's Changed
* [Docs] Evaluation metrics for mcmc, antialiased, and absgrad. by maturk in https://github.com/nerfstudio-project/gsplat/pull/300
* Update requirements.txt by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/308
* Fix projection for images with non-centered camera (e.g. crops) by JonathonLuiten in https://github.com/nerfstudio-project/gsplat/pull/305
* prevent race condition when JIT in multiprocess by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/312
* Support Multi-GPU training based on the paper "On Scaling Up 3D Gaussian Splatting Training" by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/253
* Update version.py by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/313

New Contributors
* JonathonLuiten made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/305

**Full Changelog**: https://github.com/nerfstudio-project/gsplat/compare/v1.1.1...v1.2.0

What's Changed
hot fix


**Full Changelog**: https://github.com/nerfstudio-project/gsplat/compare/v1.1.0...v1.1.1

Highlight
Multiple densification strategies are now supported in gsplat, including
- `DefaultStrategy()`: The original 3DGS densification process.
- `DefaultStrategy(absgrad=True)`: Densification process from ["AbsGS: Recovering Fine Details for 3D Gaussian Splatting"](https://arxiv.org/abs/2404.10484)
- `MCMCStrategy()`. Densification process from ["3D Gaussian Splatting as Markov Chain Monte Carlo"](https://arxiv.org/abs/2404.09591)

What's Changed
* Update 3DGS examples to use the latest nerfview from pypi by hangg7 in https://github.com/nerfstudio-project/gsplat/pull/206
* docs: update README.md by eltociear in https://github.com/nerfstudio-project/gsplat/pull/210
* Added a guidance for installing gsplat on Windows by KevinXu02 in https://github.com/nerfstudio-project/gsplat/pull/213
* Minor fix for _QuatScaleToCovarPreci back with sparse grad by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/214
* Fix wrong description in rasterization API doc by yifanlu0227 in https://github.com/nerfstudio-project/gsplat/pull/216
* support 2^n+1 channels for rasterize_to_pixels by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/218
* workaround for n-dim rendering with n>32 by wuzirui in https://github.com/nerfstudio-project/gsplat/pull/217
* Minor fix for _RasterizeToPixels back to avoid NaNs by opipari in https://github.com/nerfstudio-project/gsplat/pull/235
* clean up the shape of the colors by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/228
* static shared mem -> dynamic shared mem by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/229
* updating background for different rendering modes [D, ED, RGB+D, RGB+… by vye16 in https://github.com/nerfstudio-project/gsplat/pull/240
* Fix `max_range` in `_torch_impl._rasterize_to_pixels()` by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/242
* Implement 3DGS-MCMC in gsplat. by jefequien in https://github.com/nerfstudio-project/gsplat/pull/238
* sqrt int to float by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/246
* Clamp `ratios` in MCMC's `compute_relocation` by yzslab in https://github.com/nerfstudio-project/gsplat/pull/247
* add poetry files, and fix numpy error and no data error. by MasahiroOgawa in https://github.com/nerfstudio-project/gsplat/pull/251
* Separate C source into many translation units by fwilliams in https://github.com/nerfstudio-project/gsplat/pull/255
* Fwilliams/fix headers and ninja by fwilliams in https://github.com/nerfstudio-project/gsplat/pull/256
* template everything but don't actually support multiple precisions yet by fwilliams in https://github.com/nerfstudio-project/gsplat/pull/259
* fix glm path in JIT building by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/260
* Fwilliams/atomics by fwilliams in https://github.com/nerfstudio-project/gsplat/pull/261
* set the MAX_JOBS to 10 if not set. by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/262
* float3 -> vec3<float>; float2 -> vec2<float>; by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/267
* Start adding support for double, float16, and bfloat16 by fwilliams in https://github.com/nerfstudio-project/gsplat/pull/270
* fix test_fully_fused_projection_packed test by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/289
* Add optional masks in `rasterize_to_pixels()` to support Grendel by alexis-mmm in https://github.com/nerfstudio-project/gsplat/pull/284
* Abstracting out the GS heuristics into a Strategy class by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/278
* Bump to 1.1.0 by liruilong940607 in https://github.com/nerfstudio-project/gsplat/pull/290

New Contributors
* yifanlu0227 made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/216
* wuzirui made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/217
* opipari made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/235
* jefequien made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/238
* yzslab made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/247
* MasahiroOgawa made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/251
* fwilliams made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/255
* alexis-mmm made their first contribution in https://github.com/nerfstudio-project/gsplat/pull/284

**Full Changelog**: https://github.com/nerfstudio-project/gsplat/compare/v1.0.0...v1.1.0

- Comparing to the [official implementation](https://github.com/graphdeco-inria/gaussian-splatting), gsplat enables up to **4x less training memory footprint**, and up to **2x less training time** on Mip-NeRF 360 captures, and potential more on larger scenes.

- Support extremely large scene rendering, which is magnitudes faster than the official CUDA backend [diff-gaussian-rasterization](https://github.com/graphdeco-inria/diff-gaussian-rasterization).

- Extra features, including batch rasterization, N-D feature rendering (faster), depth rendering, sparse gradient etc.