https://github.com/pierrekieffer/neuralnetworkgenerator

Test and generate the best Neural Net model based on a series of parameters
https://github.com/pierrekieffer/neuralnetworkgenerator

deep-learning keras neural-networks

Last synced: about 2 months ago
JSON representation

Test and generate the best Neural Net model based on a series of parameters

• Host: GitHub
• URL: https://github.com/pierrekieffer/neuralnetworkgenerator
• Owner: PierreKieffer
• Created: 2018-12-02T21:04:14.000Z (about 6 years ago)
• Default Branch: master
• Last Pushed: 2019-06-21T09:17:34.000Z (over 5 years ago)
• Last Synced: 2023-03-07T01:31:41.779Z (almost 2 years ago)
• Topics: deep-learning, keras, neural-networks
• Language: Python
• Homepage:
• Size: 6.84 KB
• Stars: 0
• Watchers: 0
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# NeuralNetworkGenerator

This library provides full process to Test and generate the best Neural Net model based on a series of parameters 

for a generic input dataframe 

### load_processed_data

- load data  

### generate_nn_list

- Return a list of dictionnaries. Each dict has different param for a neural net. The input is a dict of parameters : 

	param_choices = {

	'nb_neurons' : [32,64,128],

	'nb_layers' : [3,4,5,6],

	'activation' : ['relu','elu','sigmoid'],

	'optimizer' : ['adam','rmsprop','adamax','sgd'],

	}

### compile_model

- Build a neural net from generate_nn_list output

### train_networks

- Classic Train and test models

### cross_validation

- Train and Test models on n differents train_test_split of the data, and return the score mean for each architecture of neural net