Skip to content

Utility for extracting layer weights and biases from Keras models

Notifications You must be signed in to change notification settings

kristpapadopoulos/keras-extract-weights

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

20 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Keras extract weights from model to python dictionary

The file extract_parameters.py

  • takes Keras h5py model parameter file and outputs a python dictionary with option to output as pickle file
  • extract_parameters.py --h5_path='file_path' --output_path='file_path'

Tested with:

h5py  3.1.0

Each Keras layer has a group that has members: 1 for the biases and 1 for the parameters of the layer.


Example

Output for 3 layer neural network - see dataset folder for input file.

Releases

No releases published

Packages

No packages published

Languages