Activation Density based Mixed-Precision Quantization for Energy Efficient Neural Networks

no code implementations • 12 Jan 2021 • Karina Vasquez, Yeshwanth Venkatesha, Abhiroop Bhattacharjee, Abhishek Moitra, Priyadarshini Panda

Additionally, we find that integrating AD based quantization with AD based pruning (both conducted during training) yields up to ~198x and ~44x energy reductions for VGG19 and ResNet18 architectures respectively on PIM platform compared to baseline 16-bit precision, unpruned models.

Model Compression Quantization