My implementation of transformers related papers for computer vision in pytorch

Last update: Nov 10, 2021

Overview

vision_transformers

This is my personnal repo to implement new transofrmers based and other computer vision DL models

I am currenlty working without a lot of GPU ressources therefore I mainly trained models on CIFAR 10. But my implementation are build to be fast and effective at scale.

Current paper implemented:

An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale, from Dosovitskiy et al (2020)
Patch Are All You Need ? anonymous

Baseline:

Deep Residual Learning for Image Recognition, from He et al (2015)

Models are implemented in pure pytorch and trained via pytorchlightning. Dependencies are managed by poetry. It is included an Dockerfile to create a cuda ready container with jupyter lab inside. On the development part, I use jupytext in order to avoid commit every metadata change on the notebook. Fully tested with pytest and formatted with black and isort.

If you want to create a project with similar config, just use my boilerplat.

How to use it ?

first install the dependecies:

poetry install

Then, only for development:

add the precommit hook

poetry run pre-commit install

sync the notebook (only once)

poetry shell
make notebook-sync

launch a jupyter lab session

poetry run jupyter lab

Use tensorboard

poetry shell
make tensorboard

Format the code without the precommit hook

poetry shell
make formatting

Tests:

to run the tests:

poetry shell
make tests

You might also like...

Build fully-functioning computer vision models with PyTorch

Detecto is a Python package that allows you to build fully-functioning computer vision and object detection models with just 5 lines of code. Inferenc

576 Dec 29, 2022

A PyTorch-Based Framework for Deep Learning in Computer Vision

TorchCV: A PyTorch-Based Framework for Deep Learning in Computer Vision @misc{you2019torchcv, author = {Ansheng You and Xiangtai Li and Zhen Zhu a

2.2k Jan 9, 2023

Open Source Differentiable Computer Vision Library for PyTorch

Kornia is a differentiable computer vision library for PyTorch. It consists of a set of routines and differentiable modules to solve generic computer

7.6k Jan 4, 2023

An Agnostic Computer Vision Framework - Pluggable to any Training Library: Fastai, Pytorch-Lightning with more to come

IceVision is the first agnostic computer vision framework to offer a curated collection with hundreds of high-quality pre-trained models from torchvision, MMLabs, and soon Pytorch Image Models. It orchestrates the end-to-end deep learning workflow allowing to train networks with easy-to-use robust high-performance libraries such as Pytorch-Lightning and Fastai

789 Dec 29, 2022

My implementation of transformers related papers for computer vision in pytorch

Related tags

Overview

vision_transformers

How to use it ?

launch a jupyter lab session

Use tensorboard

Format the code without the precommit hook

Tests:

You might also like...

Build fully-functioning computer vision models with PyTorch

A PyTorch-Based Framework for Deep Learning in Computer Vision

Open Source Differentiable Computer Vision Library for PyTorch

An Agnostic Computer Vision Framework - Pluggable to any Training Library: Fastai, Pytorch-Lightning with more to come

Spiking Neural Network for Computer Vision using SpikingJelly framework and Pytorch-Lightning

Implementation of self-attention mechanisms for general purpose. Focused on computer vision modules. Ongoing repository.

The Incredible PyTorch: a curated list of tutorials, papers, projects, communities and more relating to PyTorch.

Explainability for Vision Transformers (in PyTorch)

PyTorch code for Vision Transformers training with the Self-Supervised learning method DINO

Releases(0.1.0)

0.1.0(Nov 10, 2021)

Owner

samsja

Robust Lane Detection via Expanded Self Attention (WACV 2022)

Official repo for our 3DV 2021 paper "Monocular 3D Reconstruction of Interacting Hands via Collision-Aware Factorized Refinements".

Model Zoo of BDD100K Dataset

KITTI-360 Annotation Tool is a framework that developed based on python(cherrypy + jinja2 + sqlite3) as the server end and javascript + WebGL as the front end.

Official Code Release for "CLIP-Adapter: Better Vision-Language Models with Feature Adapters"

Model of an AI powered sign language interpreter.

Denoising Normalizing Flow

RefineGNN - Iterative refinement graph neural network for antibody sequence-structure co-design (RefineGNN)

A PyTorch implementation of unsupervised SimCSE

Frequency Spectrum Augmentation Consistency for Domain Adaptive Object Detection

Experiments and code to generate the GINC small-scale in-context learning dataset from "An Explanation for In-context Learning as Implicit Bayesian Inference"

Bridging Vision and Language Model

Medical Image Segmentation using Squeeze-and-Expansion Transformers

Improving Deep Network Debuggability via Sparse Decision Layers

A curated list of the latest breakthroughs in AI (in 2021) by release date with a clear video explanation, link to a more in-depth article, and code.

Jupyter notebooks for the code samples of the book "Deep Learning with Python"

Normalizing Flows with a resampled base distribution

basic tutorial on pytorch

Plenoxels: Radiance Fields without Neural Networks

Implicit Model Specialization through DAG-based Decentralized Federated Learning