Difference between revisions of "Team:Florida Atlantic/Model"

Line 25: Line 25:
 
layer of the network was a fully-connected layer with 150 nodes and
 
layer of the network was a fully-connected layer with 150 nodes and
 
hyperbolic tangent activation function. The output layer contained two
 
hyperbolic tangent activation function. The output layer contained two
nodes with a softmax activation function./<p>
+
nodes with a softmax activation function.<p>
 
<img src="https://static.igem.org/mediawiki/2017/e/ee/T--Florida_Atlantic--LSTMGraph.png">
 
<img src="https://static.igem.org/mediawiki/2017/e/ee/T--Florida_Atlantic--LSTMGraph.png">
  

Revision as of 14:42, 1 November 2017

Florida_Atlantic

Model

LSTM Machine Learning Model

The LSTM model contained two lstm layers with 300 nodes each. The first lstm layer used dropout with a dropout probability of 0.5 to avoid overfitting the training data. The third layer of the network was a fully-connected layer with 150 nodes and hyperbolic tangent activation function. The output layer contained two nodes with a softmax activation function.