ReasonFlux Series - Open-source innovative LLM post-training algorithms focusing on data selection, reinforcement learning, and inference scaling - View it on GitHub