説明なし

Ravi e9a845a293 README created online with Bitbucket 6 年 前
.ipynb_checkpoints 25d68667cb commit with assisted DE 6 年 前
__pycache__ 5c2211c4cd ready for global opt 6 年 前
datasets 9af47cfec1 conv layer nets added and working 6 年 前
materials 1d7b8580a1 again 6 年 前
models 25d68667cb commit with assisted DE 6 年 前
results 25d68667cb commit with assisted DE 6 年 前
README e9a845a293 README created online with Bitbucket 6 年 前
barplotting.ipynb 5c2211c4cd ready for global opt 6 年 前
de_conv.ipynb 25d68667cb commit with assisted DE 6 年 前
hypoptim.ipynb 25d68667cb commit with assisted DE 6 年 前
hyptune.py 1e493c3bd5 further develp in conv scatter nets 6 年 前
make_dataset.py 1d7b8580a1 again 6 年 前
mxmodel.ipynb 25d68667cb commit with assisted DE 6 年 前
pygmoimp.ipynb 25d68667cb commit with assisted DE 6 年 前
pymietest.ipynb 25d68667cb commit with assisted DE 6 年 前
scatternet.ipynb 25d68667cb commit with assisted DE 6 年 前
scipydiffe.ipynb 25d68667cb commit with assisted DE 6 年 前
scnets.py 85e716f303 Almost ready for global optimization 6 年 前
snlay.py 85e716f303 Almost ready for global optimization 6 年 前

README

https://bitbucket.org/rshegde/deepmie-deep-learning-electromagnetic-scattering

Abstract
===============================================================
We propose the use of Deep Convolutional Neural Networks (DCNN)
in concert with Differential Evolution for solving inverse problems in
electromagnetics. We show that networks built using innovations like
convolutional layers, batch normalization, parametric REctified Linear Unit
(ReLU) and residual blocks achieve comparable performance to previously
reported dense fully-connected networks but achieve drastic reduction in
parameter weights and training epoch lengths. The proposed lightweight CNNs
are used in concert with Differential Evolution global optimization to achieve
8x speedup in convergence time. Systematic hyperparameter grid searches are
reported to permit faster model discovery for application to other problems.

See the following arxiv paper for details


Requirements
================================================================
1) CUDA capable GPU
2) Python 3.5
3) Numpy, Matplotlib, Pandas, Scipy, Scikitlearn
4) Cudatooklit 9.0, CuDNN 7.1
5) Mxnet-GPU, Mxnet-cumkl
6) Keras, Keras-mxnet
7) Scattnlay -- needed if you want to generate other datasets
8) Jupyter notebook
9) ipyparallel for calling multiple runs at the same time

Getting started
==================================
Most of the code is placed in Jupyter notebooks for interactivity.
PyMIEtest.ipynb -- familiarizes you with generating datasets
Scatternets.ipynb -- allows building and visualizing models
hypoptim.ipynb -- hyperparameter grid search using scikit CV grid search
scipydiffe -- differential evolution using compiled models.