Open Source Projects in Africa

**Ethiopian Logos Open-Source Project** The Chapa-Et/ethiopianlogos repository is an open-source initiative providing free, high-quality vector logos of Ethiopian companies. This project aims to collect and share logos from various Ethiopian businesses, making them easily accessible for developers and designers. The project utilizes Node.js and is hosted on GitHub.

Andegna's Calender is a Python package that enables conversion between the Ethiopian calendar and other calendar systems, including the Gregorian calendar. It also supports Amharic date formatting. This tool is useful for applications requiring date conversions in Ethiopian or Geez date formats.

**COVID-19 Data Repository for South Africa** The dsfsi/covid19za repository provides a comprehensive dataset and interactive dashboard for tracking COVID-19 cases in South Africa. This open-source project aggregates coronavirus data for analysis and visualization, serving as a valuable resource for researchers, policymakers, and the public.

HerdNet is an open-source project developed by Alexandre-Delplanque that uses deep learning for precise detection and counting of African mammals in aerial imagery. This technology has the potential to aid in wildlife conservation and livestock management in Africa. The project is based on a paper titled "From Crowd to Herd Counting".

**ZikoJS** ZikoJS is a JavaScript library developed in Morocco, aiming to simplify coding with a focus on effortless development. It provides a framework for building user interfaces and offers mathematical capabilities.

**Factbook API** Location: GitHub Description: A free, open public domain dataset of country profiles from the World Factbook, available in JSON format without requiring an API key. This data covers various regions, including Africa, and can be useful for research and development projects.

**Instadeepai Catx** Location: Global (Open-source) Category: Artificial Intelligence, Machine Learning Summary: Catx is a contextual bandits library for continuous action trees with smoothing, built using JAX and Python, enabling efficient decision-making in complex environments. This open-source library is designed for developers and researchers working on deep learning applications.

InstaNovo is a machine learning-based tool for de novo peptide sequencing, enabling accurate and database-free peptide identification in large-scale proteomics experiments. Developed by InstaDeepAI, this open-source solution is suitable for researchers and scientists working with proteomics data. It can be accessed through GitHub at https://github.com/instadeepai/InstaNovo.

Amharic-Tokenizer is a fast and accurate language processing tool for the Amharic language, developed by sefineh-ai. It uses Syllable-aware BPE (Byte Pair Encoding) and is trainable for improved performance. This open-source tool supports language processing and deep learning applications for the Amharic language.

**Instadeepai/mlip** Instadeepai/mlip is an open-source library for efficient training and application of Machine Learning Interatomic Potentials (MLIP) in Africa. This library enables researchers to develop accurate and scalable MLIP models for various materials science applications. It is available on GitHub for developers and researchers to access and contribute to.

**Jumanji: Reinforcement Learning Environments** Jumanji is a collection of scalable reinforcement learning environments developed in JAX, a high-performance Python library. This suite provides a diverse set of environments for training and testing AI models. It is maintained by Instadeepai and is suitable for research purposes.

Mava is a research-friendly codebase for fast experimentation in multi-agent reinforcement learning using JAX, developed by instadeepai. It supports the development and testing of multi-agent systems and reinforcement learning models. This open-source project is suitable for researchers and developers working in the field of multi-agent reinforcement learning.