RUN.py

@@ -6,12 +6,11 @@ training_data, validation_data, test_data = mnist_loader.load_data_wrapper()
 import network
 import dataset_loader
 
-
+net = network.Network([262144, 30, 10])                                         #Testé : 94,56%
-net = network.Network([262144,50, 20, 30, 10])                                         #Testé : 94,56%
 net.SGD(dataset_loader.loadTrainingSet("training"), 30, 10, 3.0, test_data=dataset_loader.loadTestSet("test"))
 
 # net = network.Network([784, 100, 10])                                          #Marche mieux apparemment
-#net.SGD(dataset_loader.loadTrainingSet("setcomplete"), 30, 10, 3.0, test_data=dataset_loader.loadTestSet("setcomplete"))
+# net.SGD(training_data, 30, 10, 3.0, test_data=test_data)
 
 # net = network.Network([784, 100, 10])                                        #Marche pas bien apparemment
 # net.SGD(training_data, 30, 10, 0.001, test_data=test_data)

dataset_loader.py

@@ -29,7 +29,7 @@ def loadSet(path):
 
 	for name in filelist:
 
-		if i >= 500:
+		if i >= 100:
 
 			break
 		
@@ -40,7 +40,7 @@ def loadSet(path):
 				pix = im.load()
 				temparray = []
 
-				result.append(int(name.split("/")[-1][0]))
+				result.append(name.split("/")[-1][0])
 				
 				for x in range(im.size[0]):
 
@@ -49,7 +49,8 @@ def loadSet(path):
 						temparray.append(pix[x, y] / 255)
 
 				pixels.append(temparray)
-				print(str("%.2f" % round(i / (len(filelist) if len(filelist) < 500 else 500) * 100, 2)) + "% Done, ram usage: " + str("%.2f" % round(resource.getrusage(resource.RUSAGE_SELF).ru_maxrss / (1024*1024), 2)) + "Go", end = '\r')
+				print(temparray)
+				print(str("%.2f" % round(i / (len(filelist) if len(filelist) < 100 else 100) * 100, 2)) + "% Done, ram usage: " + str("%.2f" % round(resource.getrusage(resource.RUSAGE_SELF).ru_maxrss / (1024*1024), 2)) + "Go", end = '\r')
 				i += 1
 
 	print("max ram usage: " + str(resource.getrusage(resource.RUSAGE_SELF).ru_maxrss / (1024*1024)) + "Go")
@@ -63,7 +64,7 @@ def loadTrainingSet(path):
 
 	set = loadSet(path)
 
-	training_inputs = [np.reshape(x, (784, 1)) for x in set[0]]
+	training_inputs = [np.reshape(x, (262144, 1)) for x in set[0]]
 	training_results = [vectorized_result(int(y)) for y in set[1]]
 	training_data = zip(training_inputs, training_results)
 
@@ -75,7 +76,7 @@ def loadTestSet(path):
 
 	set = loadSet(path)
 
-	test_inputs = [np.reshape(x, (784, 1)) for x in set[0]]
+	test_inputs = [np.reshape(x, (262144, 1)) for x in set[0]]
 	test_data = zip(test_inputs, set[1])
 
 	return test_data

mnist_loader.py

@@ -66,8 +66,6 @@ def load_data_wrapper():
     validation_data = zip(validation_inputs, va_d[1])
     test_inputs = [np.reshape(x, (784, 1)) for x in te_d[0]]
     test_data = zip(test_inputs, te_d[1])
-    print(te_d[0])
-    print("1:", te_d[1])
     return (training_data, validation_data, test_data)
 
 def vectorized_result(j):

network.py

@@ -132,8 +132,6 @@ class Network(object):
         neuron in the final layer has the highest activation."""
         test_results = [(np.argmax(self.feedforward(x)), y)
                         for (x, y) in test_data]
-
-        print(test_data[0], test_data[1])
         return sum(int(x == y) for (x, y) in test_results)
 
     def cost_derivative(self, output_activations, y):