Feature Map Transform Coding for Energy-Efficient CNN Inference

This code implements the papper "Feature Map Transform Coding for Energy-Efficient CNN Inference"

Arxiv link - TODO

Flow Diagram

HW Implementation

Full explanation of the Hardware implementation, including VHDL code can be found in: https://github.com/CompressTeam/TransformCodingInference/tree/master/FPGA

Datasets

To run this code you need the training (1 batch for calibration) and validation set of ILSVRC2012 data

To get the ILSVRC2012 data, you should register on their site for access

Running instructions

python main.py --data --model <Model name (resnet18 / resnet50 / resnet101 / inception_v3 / mobilenet_v2)> --actBitwidth --weightBitwidth <4/8> --transform --transformType (eye/pca/pcaQ/pcaT)

Links to models with quantized weights can be download from: https://www.mediafire.com/file/ajna79opjt53c12/qmodels.tar/file

unzip the file and put in folder ./qmodels/

Results

Acknowledgments

TODO

Citation

TODO