SRFT: A Single-Stage Method with Supervised and Reinforcement Fine-Tuning for Reasoning | Xiaol.x | Podwise