s452639/psi
1
1
Fork 0
Code Issues Pull Requests Releases Wiki Activity

neural_networks2

Browse Source
This commit is contained in:
s452693 2021-06-20 17:59:01 +02:00
parent 59945fb8de
commit 9ca08c7afb
1 changed files with 25 additions and 35 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

56
src/neural_network.py
View File

@ -15,20 +15,10 @@ dig_train_images, dig_train_labels = extract_training_samples('digits')
dig_test_images, dig_test_labels = extract_test_samples('digits') dig_test_images, dig_test_labels = extract_test_samples('digits')
let_train_images, let_train_labels = extract_training_samples('letters') let_train_images, let_train_labels = extract_training_samples('letters')
let_test_images, let_test_labels = extract_test_samples('letters') let_test_images, let_test_labels = extract_test_samples('letters')
#print(dig_train_images.shape)
#def plotdigit(image):
# img = np.reshape(image, (-1, 28))
# imshow(img, cmap='Greys', vmin=0, vmax=255)
print(dig_train_images.shape)
"""
dig_train_images = dig_train_images / 255
dig_test_images = dig_test_images / 255
let_train_images = let_train_images / 255
let_test_images = let_test_images / 255
dig_train_images = [torch.tensor(image, dtype=torch.float32) for image in dig_train_images]
"""
#print(dig_train_images[0]) #print(dig_train_images[0])
dig_train_images = dig_train_images.reshape(len(dig_train_images),28*28) dig_train_images = dig_train_images.reshape(len(dig_train_images),28*28)
d_train = dig_train_images[:1000] d_train = dig_train_images[:1000]
@ -45,13 +35,16 @@ print(d_labelstest)
class NeuralNetwork: class NeuralNetwork:
def __init__(self, inputNodes, hiddenNodes, outputNodes, learningGrade): def __init__(self, inputNodes, hiddenNodes, outputNodes, learningGrade, fileWeight, fileHidden):
self.inodes = inputNodes self.inodes = inputNodes
self.hnodes = hiddenNodes self.hnodes = hiddenNodes
self.onodes = outputNodes self.onodes = outputNodes
self.weights = (np.random.rand(self.hnodes, self.inodes) - 0.5) """te pierwsze dwa użyj przy nauce, potem zostaw cały czas te 2"""
self.hidden = (np.random.rand(self.onodes, self.hnodes) - 0.5) #self.weights = (np.random.rand(self.hnodes, self.inodes) - 0.5)
#self.hidden = (np.random.rand(self.onodes, self.hnodes) - 0.5)
self.weights = np.load(fileWeight)
self.hidden = np.load(fileHidden)
#print( 'Matrix1 \n', self.weights) #print( 'Matrix1 \n', self.weights)
#print( 'Matrix2 \n', self.hidden) #print( 'Matrix2 \n', self.hidden)
@ -89,6 +82,10 @@ class NeuralNetwork:
pass pass
def saveTraining(self, fileWeight, fileHidden):
np.save(fileWeight, self.weights)
np.save(fileHidden, self.hidden)
def query(self, inputsList): def query(self, inputsList):
inputs = np.array(inputsList, ndmin=2).T inputs = np.array(inputsList, ndmin=2).T
@ -104,27 +101,18 @@ class NeuralNetwork:
""" """ dodaj tablicę literek"""
def getAccurancy(predictons,Y):
print(predictons,Y)
return np.sum(predictons=Y)/Y.size
def getPredictions(A2):
return np.argmax(A2,0)
"""
#n = NeuralNetwork(inputNodes=3, hiddenNodes=5, outputNodes=2, learningGrade=0.2) #n = NeuralNetwork(inputNodes=3, hiddenNodes=5, outputNodes=2, learningGrade=0.2)
n = NeuralNetwork(inputNodes=784, hiddenNodes=200, outputNodes=10, learningGrade=0.1) digitNetwork = NeuralNetwork(inputNodes=784, hiddenNodes=200, outputNodes=10, learningGrade=0.1, fileWeight="Dweights.npy", fileHidden="Dhidden.npy")
def trainNetwork(n): def trainNetwork(n, fWeight, fHidden, trainingSamples):
epochs = 10 epochs = 10
outputNodes = 10 outputNodes = 10
for e in range(epochs): for e in range(epochs):
m=0 m=0
print('Epoch', e+1) print('Epoch', e+1)
for record in d_train: for record in trainingSamples:
inputs = (np.asfarray(record[0:])/255 * 0.99) + 0.01 inputs = (np.asfarray(record[0:])/255 * 0.99) + 0.01
#print(inputs.shape) #print(inputs.shape)
@ -136,15 +124,17 @@ def trainNetwork(n):
m+=1 m+=1
pass pass
pass pass
n.saveTraining(fileWeight=fWeight, fileHidden=fHidden)
trainNetwork(n)
record = d_test[0] ##################################### ODPALANIE TRAINING
#trainNetwork(digitNetwork, "Dweights.npy", "Dhidden.npy", d_train)
#record = d_test[0]
#print('Label', d_labelstest[0]) #print('Label', d_labelstest[0])
inputs = np.asfarray(record[0:])/ 255 * 0.99 + 0.01 #inputs = np.asfarray(record[0:])/ 255 * 0.99 + 0.01
#print(n.query(inputs)) #print(n.query(inputs))
#testing