Embracing Single Stride 3D Object Detector with Sparse Transformer

Related tags

Deep LearningSST
Overview

SST: Single-stride Sparse Transformer

This is the official implementation of paper:

Embracing Single Stride 3D Object Detector with Sparse Transformer

Authors: Lue Fan, Ziqi Pang, Tianyuan Zhang, Yu-Xiong Wang, Hang Zhao, Feng Wang, Naiyan Wang, Zhaoxiang Zhang

Paper Link (Check again on Monday)

Introduction and Highlights

  • SST is a single-stride network, which maintains original feature resolution from the beginning to the end of the network. Due to the characterisric of single stride, SST achieves exciting performances on small object detection (Pedestrian, Cyclist).
  • For simplicity, except for backbone, SST is almost the same with the basic PointPillars in MMDetection3D. With such a basic setting, SST achieves state-of-the-art performance in Pedestrian and Cyclist and outperforms PointPillars more than 10 AP only at a cost of 1.5x latency.
  • SST consists of 6 Regional Sparse Attention (SRA) blocks, which deal with the sparse voxel set. It's similar to Submanifold Sparse Convolution (SSC), but much more powerful than SSC. It's locality and sparsity guarantee the efficiency in the single stride setting.
  • The SRA can also be used in many other task to process sparse point clouds. Our implementation of SRA only relies on the pure Python APIs in PyTorch without engineering efforts as taken in the CUDA implementation of sparse convolution.
  • Large room for further improvements. For example, second stage, anchor-free head, IoU scores and advanced techniques from ViT, etc.

Usage

PyTorch >= 1.9 is highly recommended for a better support of the checkpoint technique.

Our immplementation is based on MMDetection3D, so just follow their getting_started and simply run the script: run.sh. Then you will get a basic results of SST after 5~7 hours (depends on your devices).

We only provide the single-stage model here, as for our two-stage models, please follow LiDAR-RCNN. It's also a good choice to apply other powerful second stage detectors to our single-stage SST.

Main results

Single-stage Model (based on PointPillars) on Waymo validation split

#Sweeps Veh_L1 Ped_L1 Cyc_L1
SST_1f 1 73.57 80.01 70.72
SST_3f 3 75.16 83.24 75.96

Note that we train the 3 classes together, so the performance above is a little bit lower than that reported in our paper.

TODO

  • Build SRA block with similar API as Sparse Convolution for more convenient usage.

Acknowlegement

This project is based on the following codebases.

Owner
TuSimple
The Future of Trucking
TuSimple
QKeras: a quantization deep learning library for Tensorflow Keras

QKeras github.com/google/qkeras QKeras 0.8 highlights: Automatic quantization using QKeras; Stochastic behavior (including stochastic rouding) is disa

Google 437 Jan 03, 2023
Offical implementation for "Trash or Treasure? An Interactive Dual-Stream Strategy for Single Image Reflection Separation".

Trash or Treasure? An Interactive Dual-Stream Strategy for Single Image Reflection Separation (NeurIPS 2021) by Qiming Hu, Xiaojie Guo. Dependencies P

Qiming Hu 31 Dec 20, 2022
Repository for the electrical and ICT benchmark model developed in the ERIGrid 2.0 project.

Benchmark Model Electrical and ICT System This repository contains the documentation, code, and models for the electrical and ICT benchmark model deve

ERIGrid 2.0 1 Nov 29, 2021
IDA file loader for UF2, created for the DEFCON 29 hardware badge

UF2 Loader for IDA The DEFCON 29 badge uses the UF2 bootloader, which conveniently allows you to dump and flash the firmware over USB as a mass storag

Kevin Colley 6 Feb 08, 2022
Space-event-trace - Tracing service for spaceteam events

space-event-trace Tracing service for TU Wien Spaceteam events. This service is

TU Wien Space Team 2 Jan 04, 2022
Detecting drunk people through thermal images using Deep Learning (CNN)

Drunk Detection CNN Detecting drunk people through thermal images using Deep Learning (CNN) Dataset We used thermal images provided by Electronics Lab

Giacomo Ferretti 3 Oct 27, 2022
Official implementation of NeurIPS'2021 paper TransformerFusion

TransformerFusion: Monocular RGB Scene Reconstruction using Transformers Project Page | Paper | Video TransformerFusion: Monocular RGB Scene Reconstru

Aljaz Bozic 118 Dec 25, 2022
A Pythonic library for Nvidia Codec.

A Pythonic library for Nvidia Codec. The project is still in active development; expect breaking changes. Why another Python library for Nvidia Codec?

Zesen Qian 12 Dec 27, 2022
SMIS - Semantically Multi-modal Image Synthesis(CVPR 2020)

Semantically Multi-modal Image Synthesis Project page / Paper / Demo Semantically Multi-modal Image Synthesis(CVPR2020). Zhen Zhu, Zhiliang Xu, Anshen

316 Dec 01, 2022
PyTorch implementation of DreamerV2 model-based RL algorithm

PyDreamer Reimplementation of DreamerV2 model-based RL algorithm in PyTorch. The official DreamerV2 implementation can be found here. Features ... Run

118 Dec 15, 2022
PAthological QUpath Obsession - QuPath and Python conversations

PAQUO: PAthological QUpath Obsession Welcome to paquo 👋 , a library for interacting with QuPath from Python. paquo's goal is to provide a pythonic in

Bayer AG 60 Dec 31, 2022
Code and Data for NeurIPS2021 Paper "A Dataset for Answering Time-Sensitive Questions"

Time-Sensitive-QA The repo contains the dataset and code for NeurIPS2021 (dataset track) paper Time-Sensitive Question Answering dataset. The dataset

wenhu chen 35 Nov 14, 2022
A PyTorch implementation for our paper "Dual Contrastive Learning: Text Classification via Label-Aware Data Augmentation".

Dual-Contrastive-Learning A PyTorch implementation for our paper "Dual Contrastive Learning: Text Classification via Label-Aware Data Augmentation". Y

hoshi-hiyouga 85 Dec 26, 2022
[CVPR 2022] "The Principle of Diversity: Training Stronger Vision Transformers Calls for Reducing All Levels of Redundancy" by Tianlong Chen, Zhenyu Zhang, Yu Cheng, Ahmed Awadallah, Zhangyang Wang

The Principle of Diversity: Training Stronger Vision Transformers Calls for Reducing All Levels of Redundancy Codes for this paper: [CVPR 2022] The Pr

VITA 16 Nov 26, 2022
Integrated Semantic and Phonetic Post-correction for Chinese Speech Recognition

Integrated Semantic and Phonetic Post-correction for Chinese Speech Recognition | paper | dataset | pretrained detection model | Authors: Yi-Chang Che

Yi-Chang Chen 1 Aug 23, 2022
(CVPR 2022) Pytorch implementation of "Self-supervised transformers for unsupervised object discovery using normalized cut"

(CVPR 2022) TokenCut Pytorch implementation of Tokencut: Self-supervised Transformers for Unsupervised Object Discovery using Normalized Cut Yangtao W

YANGTAO WANG 200 Jan 02, 2023
A flag generation AI created using DeepAIs API

Vex AI or Vexiology AI is an Artifical Intelligence created to generate custom made flag design texts. It uses DeepAIs API. Please be aware that you must include your own DeepAI API key. See instruct

Bernie 10 Apr 06, 2022
This is the official source code for SLATE. We provide the code for the model, the training code, and a dataset loader for the 3D Shapes dataset. This code is implemented in Pytorch.

SLATE This is the official source code for SLATE. We provide the code for the model, the training code and a dataset loader for the 3D Shapes dataset.

Gautam Singh 66 Dec 26, 2022
Learning to Predict Gradients for Semi-Supervised Continual Learning

Learning to Predict Gradients for Semi-Supervised Continual Learning Code for project: "Learning to Predict Gradients for Semi-Supervised Continual Le

Yan Luo 2 Mar 05, 2022