Blogs
Technical deep-dives into verification tools, ONNX infrastructure, and neural network research.
PropDAG: A Template Framework for Bound Propagation on Neural Network DAGs
A reusable template framework for bound propagation on directed acyclic graphs, separating algorithm logic from graph traversal infrastructure.
ShapeONNX: Solving ONNX's Dynamic Shape Problem
A dual-track shape inference tool that resolves ONNX's dynamic shapes to concrete static values for neural network verification workflows.
SlimONNX: A Story of Optimizing Neural Networks for Verification
A pure Python toolkit for optimizing ONNX models specifically for verification workflows, achieving 100% optimization success across VNN-COMP 2024 benchmarks.
TorchONNX: A Compiler for ONNX-to-PyTorch Conversion
A pure Python compiler that converts ONNX models to native PyTorch code through a 6-stage pipeline, achieving 100% success on VNN-COMP 2024 benchmarks.
TorchVNNLIB: A Story of Fast VNN-LIB Property Loading
A tool for converting VNN-LIB verification properties to binary formats, achieving 10-100x faster loading with a two-tier processing architecture.
Wraact: Polytope Approximation for Sound Neural Network Verification
A unified framework for tight convex hull approximation of neural network activation functions, supporting 7 core activation types with sub-millisecond performance.