Submanifold sparse convolutional networks

Overview

Submanifold Sparse Convolutional Networks

This is the PyTorch library for training Submanifold Sparse Convolutional Networks.

Spatial sparsity

This library brings Spatially-sparse convolutional networks to PyTorch. Moreover, it introduces Submanifold Sparse Convolutions, that can be used to build computationally efficient sparse VGG/ResNet/DenseNet-style networks.

With regular 3x3 convolutions, the set of active (non-zero) sites grows rapidly:
submanifold
With Submanifold Sparse Convolutions, the set of active sites is unchanged. Active sites look at their active neighbors (green); non-active sites (red) have no computational overhead:
submanifold
Stacking Submanifold Sparse Convolutions to build VGG and ResNet type ConvNets, information can flow along lines or surfaces of active points.

Disconnected components don't communicate at first, although they will merge due to the effect of strided operations, either pooling or convolutions. Additionally, adding ConvolutionWithStride2-SubmanifoldConvolution-DeconvolutionWithStride2 paths to the network allows disjoint active sites to communicate; see the 'VGG+' networks in the paper.
Strided Convolution, convolution, deconvolution
Strided Convolution, convolution, deconvolution
From left: (i) an active point is highlighted; a convolution with stride 2 sees the green active sites (ii) and produces output (iii), 'children' of hightlighted active point from (i) are highlighted; a submanifold sparse convolution sees the green active sites (iv) and produces output (v); a deconvolution operation sees the green active sites (vi) and produces output (vii).

Dimensionality and 'submanifolds'

SparseConvNet supports input with different numbers of spatial/temporal dimensions. Higher dimensional input is more likely to be sparse because of the 'curse of dimensionality'.

Dimension Name in 'torch.nn' Use cases
1 Conv1d Text, audio
2 Conv2d Lines in 2D space, e.g. handwriting
3 Conv3d Lines and surfaces in 3D space or (2+1)D space-time
4 - Lines, etc, in (3+1)D space-time

We use the term 'submanifold' to refer to input data that is sparse because it has a lower effective dimension than the space in which it lives, for example a one-dimensional curve in 2+ dimensional space, or a two-dimensional surface in 3+ dimensional space.

In theory, the library supports up to 10 dimensions. In practice, ConvNets with size-3 SVC convolutions in dimension 5+ may be impractical as the number of parameters per convolution is growing exponentially. Possible solutions include factorizing the convolutions (e.g. 3x1x1x..., 1x3x1x..., etc), or switching to a hyper-tetrahedral lattice (see Sparse 3D convolutional neural networks).

Hello World

SparseConvNets can be built either by defining a function that inherits from torch.nn.Module or by stacking modules in a sparseconvnet.Sequential:

import torch
import sparseconvnet as scn

# Use the GPU if there is one, otherwise CPU
device = 'cuda:0' if torch.cuda.is_available() else 'cpu'

model = scn.Sequential().add(
    scn.SparseVggNet(2, 1,
                     [['C', 8], ['C', 8], ['MP', 3, 2],
                      ['C', 16], ['C', 16], ['MP', 3, 2],
                      ['C', 24], ['C', 24], ['MP', 3, 2]])
).add(
    scn.SubmanifoldConvolution(2, 24, 32, 3, False)
).add(
    scn.BatchNormReLU(32)
).add(
    scn.SparseToDense(2, 32)
).to(device)

# output will be 10x10
inputSpatialSize = model.input_spatial_size(torch.LongTensor([10, 10]))
input_layer = scn.InputLayer(2, inputSpatialSize)

msgs = [[" X   X  XXX  X    X    XX     X       X   XX   XXX   X    XXX   ",
         " X   X  X    X    X   X  X    X       X  X  X  X  X  X    X  X  ",
         " XXXXX  XX   X    X   X  X    X   X   X  X  X  XXX   X    X   X ",
         " X   X  X    X    X   X  X     X X X X   X  X  X  X  X    X  X  ",
         " X   X  XXX  XXX  XXX  XX       X   X     XX   X  X  XXX  XXX   "],

        [" XXX              XXXXX      x   x     x  xxxxx  xxx ",
         " X  X  X   XXX       X       x   x x   x  x     x  x ",
         " XXX                X        x   xxxx  x  xxxx   xxx ",
         " X     X   XXX       X       x     x   x      x    x ",
         " X     X          XXXX   x   x     x   x  xxxx     x ",]]


# Create Nx3 and Nx1 vectors to encode the messages above:
locations = []
features = []
for batchIdx, msg in enumerate(msgs):
    for y, line in enumerate(msg):
        for x, c in enumerate(line):
            if c == 'X':
                locations.append([y, x, batchIdx])
                features.append([1])
locations = torch.LongTensor(locations)
features = torch.FloatTensor(features).to(device)

input = input_layer([locations,features])
print('Input SparseConvNetTensor:', input)
output = model(input)

# Output is 2x32x10x10: our minibatch has 2 samples, the network has 32 output
# feature planes, and 10x10 is the spatial size of the output.
print('Output SparseConvNetTensor:', output)

Examples

Examples in the examples folder include

For example:

cd examples/Assamese_handwriting
python VGGplus.py

Setup

Tested with PyTorch 1.3, CUDA 10.0, and Python 3.3 with Conda.

conda install pytorch torchvision cudatoolkit=10.0 -c pytorch # See https://pytorch.org/get-started/locally/
git clone [email protected]:facebookresearch/SparseConvNet.git
cd SparseConvNet/
bash develop.sh

To run the examples you may also need to install unrar:

apt-get install unrar

License

SparseConvNet is BSD licensed, as found in the LICENSE file. Terms of use. Privacy

Links

  1. ICDAR 2013 Chinese Handwriting Recognition Competition 2013 First place in task 3, with test error of 2.61%. Human performance on the test set was 4.81%. Report
  2. Spatially-sparse convolutional neural networks, 2014 SparseConvNets for Chinese handwriting recognition
  3. Fractional max-pooling, 2014 A SparseConvNet with fractional max-pooling achieves an error rate of 3.47% for CIFAR-10.
  4. Sparse 3D convolutional neural networks, BMVC 2015 SparseConvNets for 3D object recognition and (2+1)D video action recognition.
  5. Kaggle plankton recognition competition, 2015 Third place. The competition solution is being adapted for research purposes in EcoTaxa.
  6. Kaggle Diabetic Retinopathy Detection, 2015 First place in the Kaggle Diabetic Retinopathy Detection competition.
  7. Submanifold Sparse Convolutional Networks, 2017 Introduces deep 'submanifold' SparseConvNets.
  8. Workshop on Learning to See from 3D Data, 2017 First place in the semantic segmentation competition. Report
  9. 3D Semantic Segmentation with Submanifold Sparse Convolutional Networks, 2017 Semantic segmentation for the ShapeNet Core55 and NYU-DepthV2 datasets, CVPR 2018
  10. Unsupervised learning with sparse space-and-time autoencoders (3+1)D space-time autoencoders
  11. ScanNet 3D semantic label benchmark 2018 0.726 average IOU.
  12. MinkowskiEngine is an alternative implementation of SparseConvNet; 0.736 average IOU for ScanNet.
  13. SpConv: PyTorch Spatially Sparse Convolution Library is an alternative implementation of SparseConvNet.
  14. Live Semantic 3D Perception for Immersive Augmented Reality describes a way to optimize memory access for SparseConvNet.
  15. OccuSeg real-time object detection using SparseConvNets.
  16. TorchSparse implements 3D submanifold convolutions.
  17. TensorFlow 3D implements submanifold convolutions.

Citations

If you find this code useful in your research then please cite:

3D Semantic Segmentation with Submanifold Sparse Convolutional Networks, CVPR 2018
Benjamin Graham,
Martin Engelcke,
Laurens van der Maaten,

@article{3DSemanticSegmentationWithSubmanifoldSparseConvNet,
  title={3D Semantic Segmentation with Submanifold Sparse Convolutional Networks},
  author={Graham, Benjamin and Engelcke, Martin and van der Maaten, Laurens},
  journal={CVPR},
  year={2018}
}

and/or

Submanifold Sparse Convolutional Networks, https://arxiv.org/abs/1706.01307
Benjamin Graham,
Laurens van der Maaten,

@article{SubmanifoldSparseConvNet,
  title={Submanifold Sparse Convolutional Networks},
  author={Graham, Benjamin and van der Maaten, Laurens},
  journal={arXiv preprint arXiv:1706.01307},
  year={2017}
}
The code repository for EMNLP 2021 paper "Vision Guided Generative Pre-trained Language Models for Multimodal Abstractive Summarization".

Vision Guided Generative Pre-trained Language Models for Multimodal Abstractive Summarization [Paper] accepted at the EMNLP 2021: Vision Guided Genera

CAiRE 42 Jan 07, 2023
Pytorch implementation of "Forward Thinking: Building and Training Neural Networks One Layer at a Time"

forward-thinking-pytorch Pytorch implementation of Forward Thinking: Building and Training Neural Networks One Layer at a Time Requirements Python 2.7

Kim Heecheol 65 Oct 06, 2022
sssegmentation is a general framework for our research on strongly supervised semantic segmentation.

sssegmentation is a general framework for our research on strongly supervised semantic segmentation.

445 Jan 02, 2023
Official implementation of paper Gradient Matching for Domain Generalization

Gradient Matching for Domain Generalisation This is the official PyTorch implementation of Gradient Matching for Domain Generalisation. In our paper,

94 Dec 23, 2022
:hot_pepper: R²SQL: "Dynamic Hybrid Relation Network for Cross-Domain Context-Dependent Semantic Parsing." (AAAI 2021)

R²SQL The PyTorch implementation of paper Dynamic Hybrid Relation Network for Cross-Domain Context-Dependent Semantic Parsing. (AAAI 2021) Requirement

huybery 60 Dec 31, 2022
SiT: Self-supervised vIsion Transformer

This repository contains the official PyTorch self-supervised pretraining, finetuning, and evaluation codes for SiT (Self-supervised image Transformer).

Sara Ahmed 275 Dec 28, 2022
根据midi文件演奏“风物之诗琴”的脚本 "Windsong Lyre" auto play

Genshin-lyre-auto-play 简体中文 | English 简介 根据midi文件演奏“风物之诗琴”的脚本。由Python驱动,在此承诺, ⚠️ 项目内绝不含任何能够引起安全问题的代码。 前排提示:所有键盘在动但是原神没反应的都是因为没有管理员权限,双击run.bat或者以管理员模式

御坂17032号 386 Jan 01, 2023
SpinalNet: Deep Neural Network with Gradual Input

SpinalNet: Deep Neural Network with Gradual Input This repository contains scripts for training different variations of the SpinalNet and its counterp

H M Dipu Kabir 142 Dec 30, 2022
Doubly Robust Off-Policy Evaluation for Ranking Policies under the Cascade Behavior Model

Doubly Robust Off-Policy Evaluation for Ranking Policies under the Cascade Behavior Model About This repository contains the code to replicate the syn

Haruka Kiyohara 12 Dec 07, 2022
Image inpainting using Gaussian Mixture Models

dmfa_inpainting Source code for: MisConv: Convolutional Neural Networks for Missing Data (to be published at WACV 2022) Estimating conditional density

Marcin Przewięźlikowski 8 Oct 09, 2022
Pretraining on Dynamic Graph Neural Networks

Pretraining on Dynamic Graph Neural Networks Our article is PT-DGNN and the code is modified based on GPT-GNN Requirements python 3.6 Ubuntu 18.04.5 L

7 Dec 17, 2022
Code for Greedy Gradient Ensemble for Visual Question Answering (ICCV 2021, Oral)

Greedy Gradient Ensemble for De-biased VQA Code release for "Greedy Gradient Ensemble for Robust Visual Question Answering" (ICCV 2021, Oral). GGE can

21 Jun 29, 2022
PyTorch and GPyTorch implementation of the paper "Conditioning Sparse Variational Gaussian Processes for Online Decision-making."

Conditioning Sparse Variational Gaussian Processes for Online Decision-making This repository contains a PyTorch and GPyTorch implementation of the pa

Wesley Maddox 16 Dec 08, 2022
nnFormer: Interleaved Transformer for Volumetric Segmentation

nnFormer: Interleaved Transformer for Volumetric Segmentation Code for paper "nnFormer: Interleaved Transformer for Volumetric Segmentation ". Please

jsguo 610 Dec 28, 2022
Official implementation for ICDAR 2021 paper "Handwritten Mathematical Expression Recognition with Bidirectionally Trained Transformer"

Handwritten Mathematical Expression Recognition with Bidirectionally Trained Transformer Description Convert offline handwritten mathematical expressi

Wenqi Zhao 87 Dec 27, 2022
Code for 1st place solution in Sleep AI Challenge SNU Hospital

Sleep AI Challenge SNU Hospital 2021 Code for 1st place solution for Sleep AI Challenge (Note that the code is not fully organized) Refer to the notio

Saewon Yang 13 Jan 03, 2022
An implementation of Equivariant e2 convolutional kernals into a convolutional self attention network, applied to radio astronomy data.

EquivariantSelfAttention An implementation of Equivariant e2 convolutional kernals into a convolutional self attention network, applied to radio astro

2 Nov 09, 2021
The code for our paper Semi-Supervised Learning with Multi-Head Co-Training

Semi-Supervised Learning with Multi-Head Co-Training (PyTorch) Abstract Co-training, extended from self-training, is one of the frameworks for semi-su

cmc 6 Dec 04, 2022
Code repository of the paper Neural circuit policies enabling auditable autonomy published in Nature Machine Intelligence

Neural Circuit Policies Enabling Auditable Autonomy Online access via SharedIt Neural Circuit Policies (NCPs) are designed sparse recurrent neural net

8 Jan 07, 2023
AutoVideo: An Automated Video Action Recognition System

AutoVideo is a system for automated video analysis. It is developed based on D3M infrastructure, which describes machine learning with generic pipeline languages. Currently, it focuses on video actio

Data Analytics Lab at Texas A&M University 267 Dec 17, 2022