Results

Arbitrary Precision Printed Ternary Neural Networks with Approximate Linear Threshold Gates

Each folder contains a Pareto-optimal TNN resulted by our multi-objective optimization. Area and power values refer to the classifier only, i.e., without including interfacing costs.