Diffckt Official Dataset

This Dataset is the official dataset accompanying the paper "DiffCkt: A Diffusion Model-Based Hybrid Neural Network Framework for Automatic Transistor-Level Generation of Analog Circuits". Designed to advance automated analog circuit design, this dataset provides high-quality training and research resources for the field. Constructed based on the TSMC 65nm CMOS process, it contains over 400,000 pairs of amplifier structures and performance metrics .