Unofficial Pytorch Lightning implementation of Contrastive Syn-to-Real Generalization (ICLR, 2021)

Last update: Sep 23, 2021

Related tags

Deep Learning CSG

Overview

CSG-lightning

Unofficial Pytorch Lightning implementation of Contrastive Syn-to-Real Generalization (ICLR 2021).

Based on:

Environment Setup

Tested in a Python 3.8 environment in Linux and Windows with:

Pytorch: 1.8.1
Pytorch Lightning: 1.3.1
Lightning bolts: 0.3.3
Torchmetrics: 0.3.2
1x RTX 3070

Installing the dependencies:

pip install pytorch-lightning lightning-bolts torchmetrics

Classification (VisDA17)

Dataset Setup

Download VisDA17 dataset from official website or, use the provided script for your convenience.

# The script downloads and extracts VisDA17 dataset.
# Note: It takes a very long time to download full dataset.
python datasets/prepare_visda17.py

If you downloaded the dataset manually, extract and place them as below.

📂 datasets
 ┣ 📂 visda17
 ┃ ┣ 📂 train
 ┃ ┃ 📂 validation
 ┗ ┗ 📂 test

How to run

Training

Simply run:

python run.py

or with options,

usage: run.py [-h] [-o OUTPUT] [-r ROOT] [-e EPOCHS] [-lr LEARNING_RATE] [-bs BATCH_SIZE] [-wd WEIGHT_DECAY] [--task {classification,segmentation}] [--encoder {resnet101,deeplab50,deeplab101}] [--momentum MOMENTUM] [--num-classes NUM_CLASSES] [--eval-only] [--gpus GPUS]
              [--resume RESUME] [--dev-run] [--exp-name EXP_NAME] [--augmentation AUGMENTATION] [--seed SEED] [--fc-dim FC_DIM] [--no-apool] [--single-network] [--stages STAGES [STAGES ...]] [--emb-dim EMB_DIM] [--emb-depth EMB_DEPTH] [--num-patches NUM_PATCHES]
              [--moco-weight MOCO_WEIGHT] [--moco-queue-size MOCO_QUEUE_SIZE] [--moco-momentum MOCO_MOMENTUM] [--moco-temperature MOCO_TEMPERATURE]

Evaluation

python run.py --eval-only --resume https://github.com/ryanking13/CSG/releases/download/v0.2/csg_resnet101.ckpt

Results

Model	Accuracy
CSG (from paper)	64.1
CSG (reimpl)	67.1

Semantic Segmentation

Dataset Setup (GTA5 ==> Cityscapes)

Download GTA5 and Cityscapes datasets.

Place them as below.

📂 datasets
 ┣ 📂 GTA5
 ┃ ┣ 📂 images 
 ┃ ┃ ┣ 📜 00001.png
 ┃ ┃ ┣ ...
 ┃ ┃ ┗ 📜 24966.png
 ┃ ┃ ┣ 📂 labels
 ┃ ┃ ┣ 📜 00001.png
 ┃ ┃ ┣ ...
 ┃ ┃ ┗ 📜 24966.png
 ┣ 📂 cityscapes
 ┃ ┣ 📂 leftImg8bit
 ┃ ┃ ┣ 📂 train
 ┃ ┃ ┃ 📂 val
 ┗ ┗ ┗ 📂 test
 ┃ ┣ 📂 gtFine 
 ┃ ┃ ┣ 📂 train
 ┃ ┃ ┃ 📂 val
 ┗ ┗ ┗ 📂 test

How to run

Training

Simply run:

./run_seg.sh

Evaluation

./run_seg --eval-only --resume https://github.com/ryanking13/CSG/releases/download/v0.2/csg_deeplab50.ckpt

Results

Model	IoU
CSG (from paper)	35.27
CSG (reimpl)	34.71

Differences from official implementation

Warmup LR scheduler
No layerwise LR modification
RandAugment augmentation types

Known Issues

I got error Distributed package doesn't have NCCL built in

On windows, nccl is not supported, try:

set PL_TORCH_DISTRIBUTED_BACKEND=gloo

You might also like...

Unofficial PyTorch implementation of "RTM3D: Real-time Monocular 3D Detection from Object Keypoints for Autonomous Driving" (ECCV 2020)

RTM3D-PyTorch The PyTorch Implementation of the paper: RTM3D: Real-time Monocular 3D Detection from Object Keypoints for Autonomous Driving (ECCV 2020

271 Nov 29, 2022

RGBD-Net - This repository contains a pytorch lightning implementation for the 3DV 2021 RGBD-Net paper.

[3DV 2021] We propose a new cascaded architecture for novel view synthesis, called RGBD-Net, which consists of two core components: a hierarchical depth regression network and a depth-aware generator network.

4 May 26, 2022

An essential implementation of BYOL in PyTorch + PyTorch Lightning

Essential BYOL A simple and complete implementation of Bootstrap your own latent: A new approach to self-supervised Learning in PyTorch + PyTorch Ligh

48 Sep 27, 2022

Unofficial implementation of "TTNet: Real-time temporal and spatial video analysis of table tennis" (CVPR 2020)

TTNet-Pytorch The implementation for the paper "TTNet: Real-time temporal and spatial video analysis of table tennis" An introduction of the project c

438 Dec 29, 2022

Implementation of Geometric Vector Perceptron, a simple circuit for 3d rotation equivariance for learning over large biomolecules, in Pytorch. Idea proposed and accepted at ICLR 2021

Geometric Vector Perceptron Implementation of Geometric Vector Perceptron, a simple circuit with 3d rotation equivariance for learning over large biom

59 Nov 24, 2022

Pytorch implementation of BRECQ, ICLR 2021

BRECQ Pytorch implementation of BRECQ, ICLR 2021 @inproceedings{ li&gong2021brecq, title={BRECQ: Pushing the Limit of Post-Training Quantization by Bl

148 Dec 28, 2022

Implementation of 'lightweight' GAN, proposed in ICLR 2021, in Pytorch. High resolution image generations that can be trained within a day or two

512x512 flowers after 12 hours of training, 1 gpu 256x256 flowers after 12 hours of training, 1 gpu Pizza 'Lightweight' GAN Implementation of 'lightwe

1.5k Jan 2, 2023

This repo contains the pytorch implementation for Dynamic Concept Learner (accepted by ICLR 2021).

DCL-PyTorch Pytorch implementation for the Dynamic Concept Learner (DCL). More details can be found at the project page. Framework Grounding Physical

31 Jan 6, 2023

An implementation demo of the ICLR 2021 paper Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural Networks in PyTorch.

Neural Attention Distillation This is an implementation demo of the ICLR 2021 paper Neural Attention Distillation: Erasing Backdoor Triggers from Deep

84 Jan 4, 2023

Unofficial Pytorch Lightning implementation of Contrastive Syn-to-Real Generalization (ICLR, 2021)

Related tags

Overview

CSG-lightning

Environment Setup

Classification (VisDA17)

Dataset Setup

How to run

Results

Semantic Segmentation

Dataset Setup (GTA5 ==> Cityscapes)

How to run

Results

Differences from official implementation

Known Issues

You might also like...

Unofficial PyTorch implementation of "RTM3D: Real-time Monocular 3D Detection from Object Keypoints for Autonomous Driving" (ECCV 2020)

RGBD-Net - This repository contains a pytorch lightning implementation for the 3DV 2021 RGBD-Net paper.

An essential implementation of BYOL in PyTorch + PyTorch Lightning

Unofficial implementation of "TTNet: Real-time temporal and spatial video analysis of table tennis" (CVPR 2020)

Implementation of Geometric Vector Perceptron, a simple circuit for 3d rotation equivariance for learning over large biomolecules, in Pytorch. Idea proposed and accepted at ICLR 2021

Pytorch implementation of BRECQ, ICLR 2021

Implementation of 'lightweight' GAN, proposed in ICLR 2021, in Pytorch. High resolution image generations that can be trained within a day or two

This repo contains the pytorch implementation for Dynamic Concept Learner (accepted by ICLR 2021).

An implementation demo of the ICLR 2021 paper Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural Networks in PyTorch.

Releases(v0.2)

v0.2(Aug 11, 2021)

v0.1(May 12, 2021)

Owner

Gyeongjae Choi

[ICCV 2021] Deep Hough Voting for Robust Global Registration

Reporting and Visualization for Hazardous Events

⚾🤖⚾ Automatic baseball pitching overlay in realtime

Implementation of the pix2pix model on satellite images

TVNet: Temporal Voting Network for Action Localization

Defending against Model Stealing via Verifying Embedded External Features

A simple API wrapper for Discord interactions.

An Industrial Grade Federated Learning Framework

Code for 'Single Image 3D Shape Retrieval via Cross-Modal Instance and Category Contrastive Learning', ICCV 2021

Hierarchical-Bayesian-Defense - Towards Adversarial Robustness of Bayesian Neural Network through Hierarchical Variational Inference (Openreview)

Retrieval.pytorch - The code we used in [2020 DIGIX]

Official code for our CVPR '22 paper "Dataset Distillation by Matching Training Trajectories"

Official PyTorch implementation of PICCOLO: Point-Cloud Centric Omnidirectional Localization (ICCV 2021)

Code for "Multi-View Multi-Person 3D Pose Estimation with Plane Sweep Stereo"

Fast mesh denoising with data driven normal filtering using deep variational autoencoders

Python Classes: Medical Insurance Project using Object Oriented Programming Concepts

Self-supervised learning on Graph Representation Learning (node-level task)

List of all dependencies affected by node-ipc malicious commit

Dense Prediction Transformers

Sign Language Transformers (CVPR'20)