This project creates a streamlit app to allow users to quickly find and view detailed information about ML tasks and how various models have performed on popular benchmarks. The app is designed to be user-friendly and easy to navigate. One theoretical use case is for journalists to quickly find a layman description of a given ML task and, if necessary, report on how specific capabilities related to that task (or its subtasks) have improved over time.
- Search for SOTA models by task, dataset, or model name in a "nested" directory structure
- View detailed information about a task, including:
- A layman's description of the task
- A list of subtasks (if applicable)
- A list of datasets used to benchmark the task
- A chart showing the performance of various models on selected task over time
- Papers with Code for all ML tasks, datasets, and model evaluation metrics
Clone the repository and navigate to the project directory:
git clone https://github.com/zachpinto/SOTA-explorer.git
cd SOTA-ExplorerInstall the required packages:
pip install -r requirements.txtTo run the application, execute the following command:
streamlit run HOME.pyPull requests are welcome. I am not a professional web developer, so major changes to other front-end technologies are welcome.
MIT