OCRia-1: A new SOTA 4B document parser (llamaindex.ai)

would be cool to see energy usage associated with training

Interesting, got an issue in my project for OCR-ing images. I'll definitely try this out since ollama is supported anyways