Skip to content

LorenzoValente3/Autoencoder-for-FPGA

Repository files navigation

Autoencoder-for-FPGA

In this repo it is presented an implementation of a Deep Autoencoder architecture trained on the MNIST database in FPGA, focusing on machine vision tasks for the data reconstruction and classification in the latent dimension. To implement machine learning (ML) models in FPGAs, a companion compiler based on High-Level Synthesis (HLS) called hls4ml is used. Furthermore, an optimization using both compression and quantization of Neural Networks is performed to obtain sensible reduction in model size, latency and energy consumption.

About

Autoencoder model for FPGA implementation using hls4ml. Repository for Applied Electronics Project.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published