** pyhessian
This package `pyhessian' contains the following components:
pyhessian.py
A python3 module to estimate Hessian matrices using TensorFlow
(https://www.tensorflow.org) and NumPy (http://www.numpy.org)
pyhessian_example.py
An example showing how to apply HessianEstimator on
a feed-forward neural network TensorFlow model
pyhessian is written in Python3 and depends on the TensorFlow and
NumPy libraries.
pyhessian
~~~~~~~~~
pyhessian is a module used to estimate Hessian matrices of TensorFlow
models. See the accompagnying paper "Efficient Computation of Hessian Matrices
in TensorFlow" found at: http://arxiv.org/abs/1905.05559
For models where the full Hessian matrix is too large to fit in memory,
an alternative can be to use the approximate eigendecomposition of the Hessian matrix. See the module pydeepdelta
(https://github.com/gknilsen/pydeepdelta) described in the paper "On the Delta Method for Uncertainty Approximation in Deep Learning"
found at https://arxiv.org/abs/1912.00832
Where to get
~~~~~~~~~~~~
By GIT (development):
https://github.com/gknilsen/pyhessian
Author
~~~~~~
Author: Geir K. Nilsen <geir.kjetil.nilsen@gmail.com>
My blog: http://octovoid.com
Other people have contributed to hessian_estimator theory development, see the
THANKS file.
License
~~~~~~~
Copyright (c) 2019 by Geir K. Nilsen (geir.kjetil.nilsen@gmail.com) and
the University of Bergen.
pyhessian is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with hessian_estimator. If not, see <http://www.gnu.org/licenses/>.