def main():
training_dataset = produce_dataset(
os.path.join(os.path.dirname(os.path.realpath(__file__)),
'training_data', 'pickle', 'training_playlists',
'training_4.txt'))
test_dataset = produce_dataset(
os.path.join(os.path.dirname(os.path.realpath(__file__)),
'training_data', 'pickle', 'test_playlists',
'test_4.txt'))
settings = {
# Required settings
"n_inputs":
8, # Number of input signals
"layers": [
(8, sigmoid_function
), # First Hidden Layer (number of nodes, activation function)
(1, sigmoid_function) # Output layer
],
# Optional settings
"initial_bias_value":
0.0,
"weights_low":
-0.1,
"weights_high":
0.1
}
network = NeuralNet.load_network_from_file("%s.pkl" % "training54point1")
training_set = training_dataset
test_set = test_dataset
cost_function = cross_entropy_cost
RMSprop(network,
training_set,
test_set,
cost_function,
ERROR_LIMIT=0.1,
max_iterations=100000,
batch_size=400)
network.save_network_to_file("%s.pkl" % "training2")
from nimblenet.data_structures import Instance
from nimblenet.tools import print_test
import random
import smbus
import math
import time
import os
import RPi.GPIO as GPIO
import requests
import json
import datetime
import numpy as np
##-------------------BELLE--------------------
## Loads Belle's beautiful brain
network = NeuralNet.load_network_from_file("Belle_5.pkl")
# Print a network test
#print_test( network, training_data, cost_function )
##------------------LISA--------------------
# Power management registers
power_mgmt_1 = 0x6b
power_mgmt_2 = 0x6c
gyro_scale = 131.0
accel_scale = 16384.0
address = 0x68 # This is the address value read via the i2cdetect command
import sys
from Playlist import Playlist
from nimblenet.data_structures import Instance
from nimblenet.neuralnet import NeuralNet
from playlist_recommender_utilities import *
username = sys.argv[1]
playlist = sys.argv[2]
base_playlist = Playlist.Playlist(username, playlist)
base_playlist.get_playlist()
base_playlist.generate_playlist_vector()
base_playlist.generate_normalized_aggregate_vector()
# Funtion to return the 100 closet songs to playlist normalized_aggregate_vectors
hundred_song_set = gather_hundred_songs(base_playlist)
prediction_set = create_prediction_data(hundred_song_set)
network = NeuralNet.load_network_from_file("%s.pkl" % "training54point1")
recommended_values = network.predict(prediction_set)
returned_playlist = generate_top_twentyfive(recommended_values,
hundred_song_set)
for song in returned_playlist:
print song.attributes['title']
def load_network():
return NeuralNet.load_network_from_file(network_file)
"n_inputs": 14, # Number of network input signals
"layers": [(5, sigmoid_function), (4, sigmoid_function)],
# [ (number_of_neurons, activation_function) ]
# The last pair in the list dictate the number of output signals
# Optional settings
"weights_low": -0.1, # Lower bound on the initial weight value
"weights_high": 0.4, # Upper bound on the initial weight value
}
# initialize the neural network
network = NeuralNet(settings)
network.check_gradient(training_data, cost_function)
## load a stored network configuration
network = NeuralNet.load_network_from_file("redsk_00009.pkl")
ctx = Context()
ctx.init()
# Create the user generator
user = UserGenerator()
user.create(ctx)
#Obtener imagen
depth = DepthGenerator()
depth.create(ctx)
depth.set_resolution_preset(DefResolution.RES_VGA)
depth.fps = 30
ctx.start_generating_all()