Skip to content

Project overview

Ana Robles edited this page Apr 14, 2023 · 2 revisions

About the project

In this project we will develop a system that allows the user to generate high-quality images in response to text prompts. To achive this, we will integrate two technologies: the HuggingFace library and the ChatGPT Large Language Model. The user will provide a textual prompt that describes the image they want to create. The ChatGPT Large Language Model will be used to enhance this prompt by adding a more thorough description of the desired image. This enriched prompt will be used as input to the pre-trained models for image generation, which are accessed through HuggingFace. Also, for demo purposes we will include the use of Graddio to create an interfaces.

Goal

This project aims to investigate the benefits of using GitHub Copilot and ChatGTP-4 on developer productivity. GitHub Copilot is a code generation tool developed by GitHub in partnership with OpenAI. It uses machine learning to generate code suggestions based on natural language inputs from the developer. The project will focus on comparing the productivity of developers using GitHub Copilot to those not using it. To achieve this, the project will be broken up into two sections: the first days without the tools, and the second days utilizing them.

SPACE Framework

To have a measurement and have an holistic view based on our goal of comparing the benefits of using GitHub Copilot on developer productivity, we will use the SPACE productivity framework, which describe several aspects, as satisfaction and well-being, performance, activity, communication and collaboration and efficency and flow. Allowing us to identify which areas of developer productivity are most affected by the use of GitHub Copilot, and to make recommendations for how Encora can use the tool most effectively.

Tools and others

Choseen Language: Python. It's a popular language because of it's simplicity, it has a large community and support, and a large number of available libraries and frameworks.

Choseen Framework: LangChain. Currently one of the hottest new python's framework. This framework is used for developing end-to-end applications powered by large language models.

Choseen IDE: Visual Studio Code Visual Studio Code is a lightweight and customizable IDE that supports a wide range of programming languages. It provides support for debugging, syntax highlighting, and auto-completion, as well as a range of extensions for data science and machine learning development. This IDE will be connected to use copilot.

Clone this wiki locally