From e192c7d9572c14075358536cba1b14c7f828f032 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?K=C3=BChle=2C=20Laura=20Christine=20=28lakue103=29?=
<laura.kuehle@uni-duesseldorf.de>
Date: Tue, 6 Apr 2021 19:44:55 +0200
Subject: [PATCH] Replaced all 'ThreeLayerNet' classes with 'ThreeLayerReLu'.
---
ANN_Model.py | 87 +++++++++++++++---------------------
Artificial_Neural_Network.py | 7 +--
Troubled_Cell_Detector.py | 7 ++-
3 files changed, 46 insertions(+), 55 deletions(-)
diff --git a/ANN_Model.py b/ANN_Model.py
index d806738..59821a3 100644
--- a/ANN_Model.py
+++ b/ANN_Model.py
@@ -1,59 +1,46 @@
# -*- coding: utf-8 -*-
"""
@author: Laura C. Kühle, Soraya Terrab (sorayaterrab)
+
+TODO: Combine all ThreeLayerNet classes in one class -> Done
+
+INFO: /home/laura/anaconda3/lib/python3.7/site-packages/torch/nn/modules
+
"""
import torch
# Define Neural Network
-# Model with Linear -> 2 ReLU layers -> Linear -> SoftMax
-class ThreeLayerNet(torch.nn.Module):
- def __init__(self, d_in, h, d_out):
- super(ThreeLayerNet, self).__init__()
- self.input_linear = torch.nn.Linear(d_in, h)
- self.middle_linear = torch.nn.Linear(h, h)
- self.output_linear = torch.nn.Linear(h, d_out)
- self.output_softmax = torch.nn.Softmax(dim=1)
-
- def forward(self, x_in):
- h_relu1 = self.input_linear(x_in).clamp(min=0)
- h_relu2 = self.middle_linear(h_relu1).clamp(min=0)
- y_pred = self.output_linear(h_relu2)
- y_pred = self.output_softmax(y_pred)
- return y_pred
-
-
-# Model with Linear -> 2 ReLU layers -> Linear -> SoftMax
-class ThreeLayerNetDifferentNeuronsSoftMax(torch.nn.Module):
- def __init__(self, d_in, h1, h2, d_out):
- super(ThreeLayerNetDifferentNeuronsSoftMax, self).__init__()
- self.input_linear = torch.nn.Linear(d_in, h1)
- self.middle_linear = torch.nn.Linear(h1, h2)
- self.output_linear = torch.nn.Linear(h2, d_out)
- self.output_softmax = torch.nn.Softmax(dim=1)
-
- def forward(self, x_in):
- h_relu1 = self.input_linear(x_in).clamp(min=0)
- h_relu2 = self.middle_linear(h_relu1).clamp(min=0)
- y_pred = self.output_linear(h_relu2)
- y_pred = self.output_softmax(y_pred)
- return y_pred
-
-
-# Model with Linear -> 2 ReLU layers -> Linear -> SoftMax
-class ThreeLayerNetDifferentNeuronsSigmoid(torch.nn.Module):
- def __init__(self, d_in, h1, h2, d_out):
- super(ThreeLayerNetDifferentNeuronsSigmoid, self).__init__()
- self.input_linear = torch.nn.Linear(d_in, h1)
- self.middle_linear = torch.nn.Linear(h1, h2)
- self.output_linear = torch.nn.Linear(h2, d_out)
- self.output_softmax = torch.nn.Sigmoid()
-
- def forward(self, x_in):
- h_relu1 = self.input_linear(x_in).clamp(min=0)
- h_relu2 = self.middle_linear(h_relu1).clamp(min=0)
- y_pred = self.output_linear(h_relu2)
- y_pred = self.output_softmax(y_pred)
- return y_pred
-
+# Model with Linear -> ReLu -> Linear -> ReLu -> Linear -> any activation function
+class ThreeLayerReLu(torch.nn.Module):
+ def __init__(self, config):
+ super().__init__()
+
+ input_size = config.pop('input_size', 5)
+ first_hidden_size = config.pop('first_hidden_size', 8)
+ second_hidden_size = config.pop('second_hidden_size', 4)
+ output_size = config.pop('output_size', 2)
+ activation_function = config.pop('activation_function', 'Sigmoid')
+ activation_config = config.pop('activation_config', {})
+
+ if not hasattr(torch.nn.modules.activation, activation_function):
+ raise ValueError('Invalid activation function: "%s"' % activation_function)
+
+ self._name = self.__class__.__name__ + '_' + str(first_hidden_size) + '_' + str(second_hidden_size) + '_'\
+ + activation_function
+
+ self._input_layer = torch.nn.Linear(input_size, first_hidden_size)
+ self._first_hidden_layer = torch.nn.Linear(first_hidden_size, second_hidden_size)
+ self._second_hidden_layer = torch.nn.Linear(second_hidden_size, output_size)
+ self._output_layer = getattr(torch.nn.modules.activation, activation_function)(**activation_config)
+
+ def forward(self, input_data):
+ prediction = self._input_layer(input_data).clamp(min=0)
+ prediction = self._first_hidden_layer(prediction).clamp(min=0)
+ prediction = self._second_hidden_layer(prediction)
+ prediction = self._output_layer(prediction)
+ return prediction
+
+ def get_name(self):
+ return self._name
diff --git a/Artificial_Neural_Network.py b/Artificial_Neural_Network.py
index 932df6e..02a8cdd 100644
--- a/Artificial_Neural_Network.py
+++ b/Artificial_Neural_Network.py
@@ -30,7 +30,7 @@ class ModelTrainer(object):
self.batch_size = config.pop('batch_size', 500)
self._num_epochs = config.pop('num_epochs', 1000)
self._threshold = config.pop('threshold', 1e-5)
- self._model = config.pop('model', 'ThreeLayerNet')
+ self._model = config.pop('model', 'ThreeLayerReLu')
self._model_config = config.pop('model_config', {})
self._loss_function = config.pop('loss_function', 'BCELoss')
self._loss_config = config.pop('loss_config', {})
@@ -169,8 +169,9 @@ class ModelTrainer(object):
pass
-# d_in_1, h1_1, h2_1, d_out_1 = 5, 8, 4, 2
-# model = ANN_Model.ThreeLayerNetDifferentNeuronsSoftMax(d_in_1, h1_1, h2_1, d_out_1)
+# model_config = {'input_size': 5, 'first_hidden_size': 8, 'second_hidden_size': 4, 'output_size': 2,
+# 'activation_function': 'Softmax', 'activation_config': {'dim': 1}}
+# model = ANN_Model.ThreeLayerReLu(model_config)
# learning_rate = 1e-10
# config = {'lr': learning_rate}
# # optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)
diff --git a/Troubled_Cell_Detector.py b/Troubled_Cell_Detector.py
index b83b735..2d42b76 100644
--- a/Troubled_Cell_Detector.py
+++ b/Troubled_Cell_Detector.py
@@ -198,8 +198,11 @@ class ArtificialNeuralNetwork(TroubledCellDetector):
super()._reset(config)
self._stencil_len = config.pop('stencil_len', 3)
- self._model = config.pop('model', 'ThreeLayerNetDifferentNeuronsSoftMax')
- self._model_config = config.pop('model_config', {'d_in': self._stencil_len+2, 'h1': 8, 'h2': 4, 'd_out': 2})
+ self._model = config.pop('model', 'ThreeLayerReLu')
+ self._model_config = config.pop('model_config', {'input_size': self._stencil_len+2, 'first_hidden_size': 8,
+ 'second_hidden_size': 4, 'output_size': 2,
+ 'activation_function': 'Softmax',
+ 'activation_config': {'dim': 1}})
self._model_state = config.pop('model_state', 'Train24k24k_Valid8k8k_Norm12ReLU8+4nodesSM1Adamlr1e-2MSE.pt')
if not hasattr(ANN_Model, self._model):
--
GitLab