https://huggingface.co/MaziyarPanahi

There's so many! Some of my all-time favourite open-source contributors on Hugging Face, in no particular order, and very biased for Information Retrieval and related:

• https://huggingface.co/orionweller - Lots of IR (and adjacent) papers with checkpoints, always as open in terms of models/data as possible
• https://huggingface.co/antoinelouis - Tons of French IR models from before it was cool: dense, sparse, cross-encoder, and late-interaction, he's done it all
• https://huggingface.co/stefan-it - Perhaps #1 worldwide on most interested in NER, very very up to date on the latest developments
• https://huggingface.co/rasyosef - Lots of Amharic models, including base decoders, encoders, as well as dense and sparse IR models
• https://huggingface.co/prithivida - Has trained and released each type of IR model, often with fully custom training, quite impressive

https://huggingface.co/robinhad - literally carried the burden of promoting Ukrainian to a mid-resource language by creating tons of datasets, models and benchmarks with some notable work in Crimean Tatar language promotion which is greatly underrepresented lately! Also tons of work not only in text-to-text but speech-to-text/text-to-speech 🤗

• Lots of models and experiments, for the Italian community and not only: https://huggingface.co/DeepMount00
• An icon on this platform, with more than 500 models shared: https://huggingface.co/mrm8488