Researchers have developed SAMpLE, an open-source framework that integrates machine learning models into SystemC-AMS virtual prototypes. This framework allows ML models to function as first-class Timed Dataflow components, simplifying their incorporation into simulations. SAMpLE offers two execution backends: one for online training of lightweight models in C++ and another for executing externally developed models without manual integration. It utilizes ONNX for model exchange, enabling the evaluation of various ML-based solutions within a unified and reproducible simulation environment. AI
IMPACT This framework could streamline the integration of AI models into hardware design simulations, potentially accelerating development cycles.
RANK_REASON The cluster describes a new research paper detailing a framework for integrating machine learning into virtual prototyping. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →