Commit Graph
28 Commits
Author SHA1 Message Date
Ruilong Li 39e6c97d42 set pose_optimizer to None if it's mode is "off" 2023-04-07 19:09:10 +00:00
Ruilong Li 07698dfb71 profiling 2023-04-06 20:30:56 +00:00
Matthew Tancik 2e777e4577 Update dataparser docs (#1698)
* Update dataparser docs

* Add missing files
2023-04-04 18:30:52 -07:00
machenmusik 83d0d75635 realitycapture nerfacto variable resolution (#1651)
* allow periods within filename, not only for file extension

* don't assume all images are the same resolution

* black reformatting

* Move variable resolution collate logic into vanilla datamanager

* good catch pylint

* now that vanilla supports variable resolution, remove VariableResDataManager

* remove comment

* CONSOLE.print

* avoid `IndexError: tensors used as indices must be long, byte or bool tensors`

* use large dataset parameters for phototourism
2023-03-23 21:45:40 -04:00
James Gardner 9a01738b0e Adding NeRF-OSR Dataset (#1626)
* Added NeRFOSR Dataparser

* updating nerfosr dataparser

* fix import order in datamanager

* updating nerfosr dataparser

* update nerfosr dataparser
2023-03-17 15:53:19 -07:00
Matthew Tancik a36f33c5ee Update datamager docs (#1548) 2023-03-02 10:27:30 -08:00
Matthew Tancik bfaf11cd72 Remove dataparser dependency (#1529) 2023-02-27 18:32:40 -08:00
Georgy Ponimatkin 5d640bfcfd Dataparsers for ScanNet and ARKitScenes (#1514)
* Adding dataparsers for ScanNet and ARKitScenes datasets.

* Fixing automatic width and height determination in ARKitScenes.

* Making arkitscenes_dataparser.py to pass pylint checks.

* Fixing docstring in arkitscenes_dataparser.py.

* Fixing function name in arkitscenes_dataparser.py.
2023-02-26 16:29:57 -08:00
Jingchen Ye 871eb0f7d3 Rename friends dataset to sitcoms3d (#1503)
* Rename friends dataset with official name sitcoms3d

* Rename downloaded folder

* Fix pylint
2023-02-25 14:38:23 -08:00
Matthew Tancik 69fdc7db9c Datamanager eval fix (#1482)
* Fix datamanager

* Datamanager tweak

* Hopefully fixed now

* readd comment
2023-02-22 16:53:23 -08:00
Matthew Tancik 9e68a46abe Fix datamanager (#1481) 2023-02-22 16:08:53 -08:00
Ethan Weber 3066819241 Fix eval dataloader to use eval ray generator (#1473)
fix eval ray generator
2023-02-22 09:53:42 -08:00
Alexander Kristoffersen f8e8d8266e Patch Sampling (#1458)
* patch sampling.

* changed str to literal.

* clean up sample selector.
2023-02-19 22:14:27 -08:00
Pablo Vela a37c73fb07 sdfstudio dataparser/dataset (#1381)
* add downloads for sdfstudio datasets

* add sdfstudio parser to work with sdfstudio data

* fix licensing and doc strings

* fix linter errors

* move depths/normals to metadata, create sdfdataset

* remove extra scenebox parameters

* more linting errors fix

* fix wrong annotation

* add in missing depth/normal files

* minor fixes

---------
2023-02-12 10:31:31 -08:00
LSong 9a4f5e8e45 DyCheck dataset parser (#1270)
* DyCheck dataset parser

* fix lint
2023-01-23 13:59:28 -08:00
matsurenandMatthew Tancik 408ab709ec Fix typo (#1231)
* Fix typo COLAMP->COLMAP

* Fix more typo

Co-authored-by: Matthew Tancik <tancik@berkeley.edu>
2023-01-16 12:12:55 -08:00
Matthew Tancik 335c5c3661 Fix dynamic batch logic during inference (#1160) 2022-12-21 22:50:05 -07:00
Nikhil Mohan 29df594334 Minimal disk dataparser (#1128)
* wip

* merge upstream

* clkeanup

* cleanup

* cleanup

* cleanup

* more cleanup

* cleanup

* cleanup docs

* lint

* style fix
2022-12-16 20:16:20 +00:00
Matthew Tancik a0a3def162 Fix datamanager dataparser_outputs (#1123)
* Fix datamanager dataparser_outputs

* Cast to float
2022-12-13 13:53:09 -08:00
Matthew Tancik b41ce8060a Save out dataparser transforms (#1105)
* Save out dataparser transforms

* Make missing folder
2022-12-11 13:02:50 -08:00
Kevin Chen 87c3ead680 Add Support for Equirectangular Images (#966) 2022-11-28 21:29:02 -08:00
Ethan Weber b23d76cc47 Cameras TensorDataclass and Phototourism (#961)
This branch gives us the flexibility to use Cameras as a tensordataclass, giving arbitrary batch dimensions, slicing, etc all for free. It tests this with the phototourism dataset.

Tensordataclass changes:
Originally, the tensordataclass would look at tensors in an object and take everything in front of the last dimension to be the batch dimensions, but in the case of the camera_to_world matrices, everything except for the last TWO are the important dimensions, so we needed to have a way for the tensordataclass to acknowledge these specific fields where there are multiple data dimensions behind the batch dimensions. This is important for things like reshaping where we need to know which dimensions get reshaped and which dimensions get left alone.

Cameras changes:
Now subclass of tensordataclass, which required a change to the tensordataclass itself.
The possibility of extra batch dimensions is the hairy part that requires the majority of these changes. The extra batch dimensions aren't a hassle in every function except in ray generation. The ray generation is where this operates different relative to the old cameras. The reason for this is before, for each ray we wanted to generate, we could specify a single index into our batch of cameras, but if you have multiple camera dimensions, for each ray, you need to now specify one index per camera batch dimension, meaning our camera_indices we specify need to have one extra dimension at the end, a dimension of size Cameras.ndim (number of batch dimensions in the camera object).
I also added some neat logic in generate_rays where if you want your output raybundle to be batched, if you specify batched camera_indices, you will receive the same batch size for the output raybundle, meaning you can easily generate raybundles corresponding to things like image patches and whatnot.

Phototourism changes:
The difficulty here is that images in the phototourism dataset are of different sizes, which is difficult because we can no longer batch the image tensors. We also need to better utilize caching in the CacheDataloader since the vanilladatamanager defaults don't really use it as well as they could, and without properly utilizing it, the rays / sec are maybe 25% of the standard nerfstudio dataset speeds (which are 10% of the size of the phototourism datasets). To do this, we specify a collate function in the existing datamanager configs, and we make a subclass of the vanilladatamanagerconfig named VariableResDataManagerConfig which will have a custom collate function storing the images in a list, and using better defaults for caching in the dataloader, caching ~40 images at a time. The slowest part of using a list is pixel sampling, where we added a custom if case for if our pixel sampler receives a batch of images that are in a list, and the lack of ability to vectorize operations over a list means that caching in the dataloader is even more important.
2022-11-25 22:16:11 -08:00
Nikhil Mohan a82d19f549 Support downsample in dataset (#973)
* init

* tests passing

* pass through arguments

* verbose

* remove debug prints

* switch to bilinear

* update viewer

* make parser outputs private

* change name
2022-11-19 17:06:34 +00:00
Matthew Tancik 6baf31dc9a Fix python 3.7 support (#959) 2022-11-15 09:51:23 -08:00
Nikhil Mohan b5c34f0154 Add 'inference' mode to pipeline (#956)
* cleanup test mode

* cleanup eval loading

* change default

* cleanup

* cleanup

* comment
2022-11-14 16:56:36 +00:00
Zak Murezandzmurez 6ea7dbf934 NuScenes dataparser (#939)
* nuscenes dataparser

* cleanup

* style

* cleanup

* cleanup

* cleanup

Co-authored-by: zmurez <zmurez@eng.ucsd.edu>
2022-11-11 18:40:36 -08:00
Brent YiandEthan Weber ad8599aad0 Add missing __init__.py files (#935)
* Add missing __init__.py files

* fixing lint

Co-authored-by: Ethan Weber <eweb0124@gmail.com>
2022-11-10 16:56:13 -08:00
Nikhil Mohan 1a24f3e58c Refactor Data Pipeline (#896)
* add none check

* black formatting

* cleanup

* cleanup

* fix tests

* remove new line

* init

* toy example working

* semantic nerf working

* simplify pixel samplers

* cleanup + fix tests

* comments

* fix tests

* fix tests
2022-11-06 07:40:43 -08:00