DeepLM: Large-scale Nonlinear Least Squares on Deep Learning Frameworks using Stochastic Domain Decomposition (CVPR 2021)
Please install the following
Then, run the example via
sh example.sh
A full set of test data can be downloaded from the BAL page (Software & Data -> Available Dataset).
Please check the examples folder for more application and features.
after run
python3 bundle_adjuster.py --balFile filename --device cpu
there is an error:
Traceback (most recent call last): File "bundle_adjuster.py", line 5, in <module> import BACore ImportError: /home/wlh/DeepLM/build/BACore.cpython-38-x86_64-linux-gnu.so: undefined symbol: _Z16THPVariable_WrapN2at6TensorE
ubuntu1804 without cuda
Hi, the open source code seems not include stochastic domain decomposition in bundle adjustment. I have tested deeplm on 1dsfm datasets, I noticed that current version can only handle less than 50w projections optimization on rtx 2080 gpu. And current version can not fix first image poses when optimize all image poses like ceres solver.
Hi, I have implemented a new loss function in bundle adjustment to solve the shared camera intrinsic optimization. But I found when all images share the same camera, the optimization speed is about 20 times slower than the case that each image has an camera to be optimized.
cmake .. -DCMAKE_BUILD_TYPE=Release -DWITH_CUDA=ON
-- The CXX compiler identification is GNU 8.3.0
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Found OpenMP_CXX: -fopenmp (found version "4.5")
-- Found OpenMP: TRUE (found version "4.5")
-- The CUDA compiler identification is unknown
-- Check for working CUDA compiler: /opt/tiger/cuda/bin/nvcc
-- Check for working CUDA compiler: /opt/tiger/cuda/bin/nvcc -- broken
CMake Error at /usr/share/cmake-3.16/Modules/CMakeTestCUDACompiler.cmake:46 (message):
The CUDA compiler
"/opt/tiger/cuda/bin/nvcc"
is not able to compile a simple test program.
It fails with the following output: Change Dir: /opt/tiger/code/DeepLM/build/CMakeFiles/CMakeTmp
Hello Jingwei @hjwdzh, this is Jiaming Sun from SenseTime & ZJU.
I came across this repo recently and found the paper title very interesting. I've been looking for a pytorch-based second-order optimizer suited for NLLS problems for a long time😁! Glad you have worked on it and got the paper accepted in CVPR, I can expect the paper to be very solid and inspiring. I'm also very interested since it can be very useful in a project I'm currently working on.
I wonder if you could give some rough timeline on the release of the paper and the code? It would be really helpful for me to plan my current project accordingly.
Thanks very much!
Best, Jiaming Sun
DeepMLS: Deep Implicit Moving Least-Squares Functions for 3D Reconstruction This repository contains the implementation of the paper: Deep Implicit Mo
T-LOAM: Truncated Least Squares Lidar-only Odometry and Mapping in Real-Time The first Lidar-only odometry framework with high performance based on tr
High-Performance MLS-MPM Solver with Cutting and Coupling (CPIC) (MIT License) A Moving Least Squares Material Point Method with Displacement Disconti
arXiv, porject page, paper Blind Image Decomposition (BID) Blind Image Decomposition is a novel task. The task requires separating a superimposed imag
YOLTv5 rapidly detects objects in arbitrarily large aerial or satellite images that far exceed the ~600×600 pixel size typically ingested by deep learning object detection frameworks.
RGBT Crowd Counting Lingbo Liu, Jiaqi Chen, Hefeng Wu, Guanbin Li, Chenglong Li, Liang Lin. "Cross-Modal Collaborative Representation Learning and a L
PointNetVLAD: Deep Point Cloud Based Retrieval for Large-Scale Place Recognition PointNetVLAD: Deep Point Cloud Based Retrieval for Large-Scale Place
Skeleton Aware Multi-modal Sign Language Recognition By Songyao Jiang, Bin Sun, Lichen Wang, Yue Bai, Kunpeng Li and Yun Fu. Smile Lab @ Northeastern
Portrait Photo Retouching with PPR10K Paper | Supplementary Material PPR10K: A Large-Scale Portrait Photo Retouching Dataset with Human-Region Mask an
HSC4D: Human-centered 4D Scene Capture in Large-scale Indoor-outdoor Space Using Wearable IMUs and LiDAR. CVPR 2022 [Project page | Video] Getting sta
DeepONet: Learning nonlinear operators The source code for the paper Learning nonlinear operators via DeepONet based on the universal approximation th
PiSL A python implementation of Physics-informed Spline Learning for nonlinear dynamics discovery. Sun, F., Liu, Y. and Sun, H., 2021. Physics-informe
Knover Knover is a toolkit for knowledge grounded dialogue generation based on PaddlePaddle. Knover allows researchers and developers to carry out eff
NLopt is a library for nonlinear local and global optimization, for functions with and without gradient information. It is designed as a simple, unifi
GraspNet Baseline Baseline model for "GraspNet-1Billion: A Large-Scale Benchmark for General Object Grasping" (CVPR 2020). [paper] [dataset] [API] [do
Large-Scale Long-Tailed Recognition in an Open World [Project] [Paper] [Blog] Overview Open Long-Tailed Recognition (OLTR) is the author's re-implemen
DRNet for Video Indvidual Counting (CVPR 2022) Introduction This is the official PyTorch implementation of paper: DR.VIC: Decomposition and Reasoning
MMdnn MMdnn is a comprehensive and cross-framework tool to convert, visualize and diagnose deep learning (DL) models. The "MM" stands for model manage
Bulk2Space Spatially resolved single-cell deconvolution of bulk transcriptomes using Bulk2Space Bulk2Space is a spatial deconvolution method based on
103 Dec 22, 2022
183 Dec 1, 2022
2.2k Dec 31, 2022
64 Dec 20, 2022
145 Jan 1, 2023
37 Dec 8, 2022
294 Dec 12, 2022
128 Dec 8, 2022
184 Dec 11, 2022
51 Nov 29, 2022
239 Jan 2, 2023
8 Jul 13, 2022
607 Dec 31, 2022
1.4k Dec 25, 2022
209 Dec 29, 2022
761 Dec 26, 2022
35 Nov 22, 2022
5.7k Jan 9, 2023
60 Dec 27, 2022