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

🏖 Keras Implementation of Painting outside the box

An executor that loads ONNX models and embeds documents using the ONNX runtime.

Official codebase for ICLR oral paper Unsupervised Vision-Language Grammar Induction with Shared Structure Modeling

CondLaneNet: a Top-to-down Lane Detection Framework Based on Conditional Convolution

Fast Learning of MNL Model From General Partial Rankings with Application to Network Formation Modeling

Code for "PV-RAFT: Point-Voxel Correlation Fields for Scene Flow Estimation of Point Clouds", CVPR 2021

Datasets, Transforms and Models specific to Computer Vision

Fully Automatic Page Turning on Real Scores

A Topic Modeling toolbox

Combining Automatic Labelers and Expert Annotations for Accurate Radiology Report Labeling Using BERT

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

ADOP: Approximate Differentiable One-Pixel Point Rendering

We evaluate our method on different datasets (including ShapeNet, CUB-200-2011, and Pascal3D+) and achieve state-of-the-art results, outperforming all the other supervised and unsupervised methods and 3D representations, all in terms of performance, accuracy, and training time.

Simple torch.nn.module implementation of Alias-Free-GAN style filter and resample

Generate images from texts. In Russian

The Video-based Accident Detection System built in Python

A CROSS-MODAL FUSION NETWORK BASED ON SELF-ATTENTION AND RESIDUAL STRUCTURE FOR MULTIMODAL EMOTION RECOGNITION

Generative Adversarial Text-to-Image Synthesis

Аналитика доходности инвестиционного портфеля в Тинькофф брокере

Testing and Estimation of structural breaks in Stata