taskswithcode/grit

GRiT: A Generative Region-to-text Transformer for Object Understanding (https://arxiv.org/abs/2212.00280)

We spotlight Authors With Code, make it easier to find solutions for Tasks (With Code), and enable learners to grasp Concepts With Code. These initiatives are work in progress

What is the AuthorsWithCode initiative?

Machine learning researchers and practitioners contributing to open source are the true catalysts AI progress. Their contributions have paved the way for innovations across industries, with many corporations benefiting significantly from this open collaboration.

It could be a strategic advantage for businesses leveraging machine learning to directly invest in these researchers and practitioners. Especially those who share code and models with licenses that encourage widespread use. Presently, while use of open source with permissible licenses primarily requires just attribution from companies, businesses can demonstrate their commitment to innovation and benefit even more by actively investing in the community that propels their advancements.

GitHub sponsorship is one direct way to accomplish this. However, the majority of them (~98%) have not enabled sponsorship yet. At this point, all we can do is a symbolic gesture of support by starring their repos.

AuthorsWithCode pursues two primary objectives:

1. To inspire both individuals and businesses to sponsor researchers and practitioners.
2. To create a direct channel for both rewarding and investing in them by encouraging them to activate GitHub sponsorship.

An auxiliary objective is to fill the current gap in finding both paper and code contributions by an author in one place - an author-centric view. This gap is not covered by Google Scholar, which focuses primarily on citations. Additionally, Google Scholar overlooks an important segment that contributes to open-source implementations of papers: practitioners, some of whom have never published a single paper. Paperswithcode, in contrast, provides a paper-centric view with links to accompanying code, encompassing both researchers and practitioners

We address these objectives with an app that facilitates the discovery of these contributors and simplifies the process of investing in them.

AuthorsWithCode relies on openly published data from paperswithcode, to create an author-centric view by combining it with GitHub data available through APIs.

Other contributions to Open Research & Education

• - Watch. Walkthrough. Interact. ConceptsWithCode ML concepts explained, accompanied by a notebook and a code walkthrough

• - TasksWithCode Discover and compare SOTA solutions for tasks with a few keystrokes/clicks

• - Free APIs for tasks using SOTA models (with permissive licenses). Please use for evaluation only to avoid overwhelming servers

• - demo apps hosted on TWC servers and/or on Hugging Face

• - notebooks in Google colab

• - Github repos with experiments that are work in progress

• - Aperiodic posts on Substack, Ghost, and Medium

• - Twitter, Sigmoid Social, Threads, Youtube, TikTok

Support for open research & education

• Platinum level supporter of Sigmoid Social - a Mastodon instance for people researching, working on, or just interested in AI
• Supporting NLP News on Substack
• Supporting researchers/practitioners here on Github (see "Sponsoring" block/tab on this page ).