Datasets and Model Checkpoints for Paper "From SFT to RL: Demystifying the Post-Training Pipeline for LLM-based Vulnerability Detection"