Official implementation of the paper Label-Efficient Semantic Segmentation with Diffusion Models

Overview

Label-Efficient Semantic Segmentation with Diffusion Models

Official implementation of the paper Label-Efficient Semantic Segmentation with Diffusion Models

This code is based on datasetGAN and guided-diffusion.

Note: use --recurse-submodules when clone.

 

Overview

The paper investigates the representations learned by the state-of-the-art DDPMs and shows that they capture high-level semantic information valuable for downstream vision tasks. We design a simple segmentation approach that exploits these representations and outperforms the alternatives in the few-shot operating point in the context of semantic segmentation.

DDPM-based Segmentation

 

Dependencies

  • Python >= 3.7
  • Packages: see requirements.txt

 

Datasets

The evaluation is performed on 6 collected datasets with a few annotated images in the training set: Bedroom-18, FFHQ-34, Cat-15, Horse-21, CelebA-19 and ADE-Bedroom-30. The number corresponds to the number of semantic classes.

datasets.tar.gz (~47Mb)

 

DDPM

Pretrained DDPMs

The models trained on LSUN are adopted from guided-diffusion. FFHQ-256 is trained by ourselves using the same model parameters as for the LSUN models.

LSUN-Bedroom: lsun_bedroom.pt
FFHQ-256: ffhq.pt
LSUN-Cat: lsun_cat.pt
LSUN-Horse: lsun_horse.pt

Run

  1. Download the datasets:
      bash datasets/download_datasets.sh
  2. Download the DDPM checkpoint:
       bash checkpoints/ddpm/download_checkpoint.sh
  3. Check paths in experiments/ /ddpm.json
  4. Run: bash scripts/ddpm/train_interpreter.sh

Available checkpoint names: lsun_bedroom, ffhq, lsun_cat, lsun_horse
Available dataset names: bedroom_28, ffhq_34, cat_15, horse_21, celeba_19, ade_bedroom_30

How to improve the performance

  1. Set input_activations=true in experiments/ /ddpm.json .
       In this case, the feature dimension is 18432.
  2. Tune for a particular task what diffusion steps and UNet blocks to use.

 

DatasetDDPM

Synthetic datasets

To download DDPM-produced synthetic datasets (50000 samples, ~7Gb):
bash synthetic-datasets/gan/download_synthetic_dataset.sh

Run | Option #1

  1. Download the synthetic dataset:
       bash synthetic-datasets/ddpm/download_synthetic_dataset.sh
  2. Check paths in experiments/ /datasetDDPM.json
  3. Run: bash scripts/datasetDDPM/train_deeplab.sh

Run | Option #2

  1. Download the datasets:
       bash datasets/download_datasets.sh

  2. Download the DDPM checkpoint:
       bash checkpoints/ddpm/download_checkpoint.sh

  3. Check paths in experiments/ /datasetDDPM.json

  4. Train an interpreter on a few DDPM-produced annotated samples:
       bash scripts/datasetDDPM/train_interpreter.sh

  5. Generate a synthetic dataset:
       bash scripts/datasetDDPM/generate_dataset.sh
        Please specify the hyperparameters in this script for the available resources.
        On 8xA100 80Gb, it takes about 12 hours to generate 10000 samples.

  6. Run: bash scripts/datasetDDPM/train_deeplab.sh
       One needs to specify the path to the generated data. See comments in the script.

Available checkpoint names: lsun_bedroom, ffhq, lsun_cat, lsun_horse
Available dataset names: bedroom_28, ffhq_34, cat_15, horse_21

 

SwAV

Pretrained SwAVs

We pretrain SwAV models using the official implementation on the LSUN and FFHQ-256 datasets:

LSUN-Bedroom: lsun_bedroom.pth
FFHQ-256: ffhq.pth
LSUN-Cat: lsun_cat.pth
LSUN-Horse: lsun_horse.pth

Training setup:

Dataset epochs batch-size multi-crop num-prototypes
LSUN 200 1792 2x256 + 6x108 1000
FFHQ-256 400 2048 2x224 + 6x96 200

Run

  1. Download the datasets:
       bash datasets/download_datasets.sh
  2. Download the SwAV checkpoint:
       bash checkpoints/swav/download_checkpoint.sh
  3. Check paths in experiments/ /swav.json
  4. Run: bash scripts/swav/train_interpreter.sh

Available checkpoint names: lsun_bedroom, ffhq, lsun_cat, lsun_horse
Available dataset names: bedroom_28, ffhq_34, cat_15, horse_21, celeba_19, ade_bedroom_30

 

DatasetGAN

Opposed to the official implementation, more recent StyleGAN2(-ADA) models are used.

Synthetic datasets

To download GAN-produced synthetic datasets (50000 samples):

bash synthetic-datasets/gan/download_synthetic_dataset.sh

Run

Since we almost fully adopt the official implementation, we don't provide our reimplementation here. However, one can still reproduce our results:

  1. Download the synthetic dataset:
      bash synthetic-datasets/gan/download_synthetic_dataset.sh
  2. Change paths in experiments/ /datasetDDPM.json
  3. Change paths and run: bash scripts/datasetDDPM/train_deeplab.sh

Available dataset names: bedroom_28, ffhq_34, cat_15, horse_21

 

Results

  • Performance in terms of mean IoU:
Method Bedroom-28 FFHQ-34 Cat-15 Horse-21 CelebA-19 ADE-Bedroom-30
ALAE 20.0 ± 1.0 48.1 ± 1.3 -- -- 49.7 ± 0.7 15.0 ± 0.5
VDVAE -- 57.3 ± 1.1 -- -- 54.1 ± 1.0 --
GAN Inversion 13.9 ± 0.6 51.7 ± 0.8 21.4 ± 1.7 17.7 ± 0.4 51.5 ± 2.3 11.1 ± 0.2
GAN Encoder 22.4 ± 1.6 53.9 ± 1.3 32.0 ± 1.8 26.7 ± 0.7 53.9 ± 0.8 15.7 ± 0.3
SwAV 41.0 ± 2.3 54.7 ± 1.4 44.1 ± 2.1 51.7 ± 0.5 53.2 ± 1.0 30.3 ± 1.5
DatasetGAN 31.3 ± 2.7 57.0 ± 1.0 36.5 ± 2.3 45.4 ± 1.4 -- --
DatasetDDPM 46.9 ± 2.8 56.0 ± 0.9 45.4 ± 2.8 60.4 ± 1.2 -- --
DDPM 46.1 ± 1.9 57.0 ± 1.4 52.3 ± 3.0 63.1 ± 0.9 57.0 ± 1.0 32.3 ± 1.5

 

  • Examples of segmentation masks predicted by the DDPM-based method:
DDPM-based Segmentation

 

Cite

@misc{baranchuk2021labelefficient,
      title={Label-Efficient Semantic Segmentation with Diffusion Models}, 
      author={Dmitry Baranchuk and Ivan Rubachev and Andrey Voynov and Valentin Khrulkov and Artem Babenko},
      year={2021},
      eprint={2112.03126},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}
Owner
Yandex Research
Yandex Research
OMLT: Optimization and Machine Learning Toolkit

OMLT is a Python package for representing machine learning models (neural networks and gradient-boosted trees) within the Pyomo optimization environment.

C⚙G - Imperial College London 179 Jan 02, 2023
PyTorch Code for NeurIPS 2021 paper Anti-Backdoor Learning: Training Clean Models on Poisoned Data.

Anti-Backdoor Learning PyTorch Code for NeurIPS 2021 paper Anti-Backdoor Learning: Training Clean Models on Poisoned Data. Check the unlearning effect

Yige-Li 51 Dec 07, 2022
Pure python implementation reverse-mode automatic differentiation

MiniGrad A minimal implementation of reverse-mode automatic differentiation (a.k.a. autograd / backpropagation) in pure Python. Inspired by Andrej Kar

Kenny Song 76 Sep 12, 2022
Home for cuQuantum Python & NVIDIA cuQuantum SDK C++ samples

Welcome to the cuQuantum repository! This public repository contains two sets of files related to the NVIDIA cuQuantum SDK: samples: All C/C++ sample

NVIDIA Corporation 147 Dec 27, 2022
Official code for our CVPR '22 paper "Dataset Distillation by Matching Training Trajectories"

Dataset Distillation by Matching Training Trajectories Project Page | Paper This repo contains code for training expert trajectories and distilling sy

George Cazenavette 256 Jan 05, 2023
Pytorch Implementation of Residual Vision Transformers(ResViT)

ResViT Official Pytorch Implementation of Residual Vision Transformers(ResViT) which is described in the following paper: Onat Dalmaz and Mahmut Yurt

ICON Lab 41 Dec 08, 2022
Code for layerwise detection of linguistic anomaly paper (ACL 2021)

Layerwise Anomaly This repository contains the source code and data for our ACL 2021 paper: "How is BERT surprised? Layerwise detection of linguistic

6 Dec 07, 2022
The undersampled DWI image using Slice-Interleaved Diffusion Encoding (SIDE) method can be reconstructed by the UNet network.

UNet-SIDE The undersampled DWI image using Slice-Interleaved Diffusion Encoding (SIDE) method can be reconstructed by the UNet network. For Super Reso

TIANTIAN XU 1 Jan 13, 2022
A GUI to automatically create a TOPAS-readable MLC simulation file

Python script to create a TOPAS-readable simulation file descriring a Multi-Leaf-Collimator. Builds the MLC using the data from a 3D .stl file.

Sebastian Schäfer 0 Jun 19, 2022
DetCo: Unsupervised Contrastive Learning for Object Detection

DetCo: Unsupervised Contrastive Learning for Object Detection arxiv link News Sparse RCNN+DetCo improves from 45.0 AP to 46.5 AP(+1.5) with 3x+ms trai

Enze Xie 234 Dec 18, 2022
⚾🤖⚾ Automatic baseball pitching overlay in realtime

⚾ Automatically overlaying pitch motion and trajectory with machine learning! This project takes your baseball pitching clips and automatically genera

Tony Chou 240 Dec 05, 2022
Key information extraction from invoice document with Graph Convolution Network

Key Information Extraction from Scanned Invoices Key information extraction from invoice document with Graph Convolution Network Related blog post fro

Phan Hoang 39 Dec 16, 2022
A PyTorch implementation of the architecture of Mask RCNN

EDIT (AS OF 4th NOVEMBER 2019): This implementation has multiple errors and as of the date 4th, November 2019 is insufficient to be utilized as a reso

Sai Himal Allu 975 Dec 30, 2022
Convert scikit-learn models to PyTorch modules

sk2torch sk2torch converts scikit-learn models into PyTorch modules that can be tuned with backpropagation and even compiled as TorchScript. Problems

Alex Nichol 101 Dec 16, 2022
Tightness-aware Evaluation Protocol for Scene Text Detection

TIoU-metric Release on 27/03/2019. This repository is built on the ICDAR 2015 evaluation code. If you propose a better metric and require further eval

Yuliang Liu 206 Nov 18, 2022
Minimalistic PyTorch training loop

Backbone for PyTorch training loop Will try to keep it minimalistic. pip install back from back import Bone Features Progress bar Checkpoints saving/l

Kashin 4 Jan 16, 2020
Dynamic View Synthesis from Dynamic Monocular Video

Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfer This repository contains code to compute depth from a

Intelligent Systems Lab Org 2.3k Jan 01, 2023
This is the workbook I created while I was studying for the Qiskit Associate Developer exam. I hope this becomes useful to others as it was for me :)

A Workbook for the Qiskit Developer Certification Exam Hello everyone! This is Bartu, a fellow Qiskitter. I have recently taken the Certification exam

Bartu Bisgin 66 Dec 10, 2022
Code for 'Self-Guided and Cross-Guided Learning for Few-shot segmentation. (CVPR' 2021)'

SCL Introduction Code for 'Self-Guided and Cross-Guided Learning for Few-shot segmentation. (CVPR' 2021)' We evaluated our approach using two baseline

34 Oct 08, 2022
Sample Code for "Pessimism Meets Invariance: Provably Efficient Offline Mean-Field Multi-Agent RL"

Sample Code for "Pessimism Meets Invariance: Provably Efficient Offline Mean-Field Multi-Agent RL" This is the official codebase for Pessimism Meets I

3 Sep 19, 2022