Skip to content

This is an upgrade for DeepSense using Tensorflow2

Notifications You must be signed in to change notification settings

TrajM/DeepSenseTF2

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 

Repository files navigation

DeepSenseTF2

This is an upgrade for DeepSense using Tensorflow2.

Based on the work of https://github.com/yscacaca/DeepSense and the Dataset upgrade of https://github.com/zhezh/DeepSense, I re-wrote the HHAR task of DeepSense using TensorFlow 2. The code is designed in a modular way to add as much sensors as needed and configure the CNNs kernels and strides depending on the training and design needs.

Training and evaluation Dataset can be downloaded from: https://drive.google.com/drive/folders/1k7PRcJk4Bd0cVAMTHfO01Gd55-uGvlhA?usp=sharing

This work is intended for embedded systems implementation. The model is trained and converted to Tensorflow Lite with "DeepSense_Training.ipynb"

The file "DeepSense_TFLite_Decompose.ipynb" is an extra step to decompose DeepSense into 3 sub-networks: the CNN part, the RNN part and the Fully-Connected Dense part for further performance measurements and network partitioning for an embedded system implementation.

About

This is an upgrade for DeepSense using Tensorflow2

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published