Language inclusivity in AI takes a significant leap forward with the development of multimodal large language models (LLMs) like GPT-4 and LLaMA, tailored for low-resource languages such as Amharic. Despite the challenge posed by limited training data, researchers have innovatively augmented datasets using translation models, paving the way for a more inclusive digital landscape.

• A deep dive into Amharic LLaMA and LLaVA, LLMs designed for low-resource languages.
• Utilization of open source translation models for extensive data augmentation.
• Integration of an image encoder to create a multimodal LLM that comprehends both text and images.
• Contribution of an Amharic benchmarking dataset to the global AI community.

This initiative not only showcases the potential of AI in embracing linguistic diversity but also sets a precedent for developing robust models that cater to a global audience. It encourages further exploration into bridging language barriers and enhancing accessibility using AI.