HuggingFaceTB

This is the home for small LLMs (SmolLM) and high quality pre-training datasets, such as Cosmopedia and Smollm-Corpus.

We released:

• Cosmopedia: the largest open synthetic dataset, with 25B tokens and more than 30M samples. It contains synthetic textbooks, blog posts, stories, posts, and WikiHow articles generated by Mixtral-8x7B-Instruct-v0.1.
• Cosmo-1B a 1B model trained on Cosmopedia.
• FineWeb-Edu: a filtered version of FineWeb dataset for educational content
• Smollm-Corpus: the pre-training corpus of SmolLM models including Cosmopedia v0.2, FineWeb-Edu and Python-Edu.
• SmolLM models and SmolLM2: a series of strong small models in three sizes: 135M, 360M and 1.7B

For more details check our blog posts: https://huggingface.co/blog/cosmopedia and https://huggingface.co/blog/smollm