final submission XOR 99.3 mnist 96.8

Assignment2/main.py

@@ -13,16 +13,16 @@ def taskXor():
 	# TASK 3a (Marks 7) - YOUR CODE HERE
 	# raise NotImplementedError
 	###############################################
-	YTrain , YVal, YTest = np.array([int(i[1]==1) for i in YTrain]).reshape((-1,1)),  np.array([int(i[1]==1) for i in YVal]).reshape((-1,1)),  np.array([int(i[1]==1) for i in YTest]).reshape((-1,1))
-	lr,batchSize,epochs = 0.01,50,10
+	# YTrain , YVal, YTest = np.array([int(i[1]==1) for i in YTrain]).reshape((-1,1)),  np.array([int(i[1]==1) for i in YVal]).reshape((-1,1)),  np.array([int(i[1]==1) for i in YTest]).reshape((-1,1))
+	lr,batchSize,epochs = -0.8,25,3500
 	nn1 = nn.NeuralNetwork(lr, batchSize, epochs)
 	# Add layers to neural network corresponding to inputs and outputs of given data
 	input_layer = XTrain.shape[1]
-	hidden_layer = 2
-	output_layer = 1
+	hidden_layer = 20
+	output_layer = 2
 # 	activation_fn = 'softmax' # 'relu'
 	nn1.addLayer(nn.FullyConnectedLayer(input_layer,hidden_layer,'relu'))
-	nn1.addLayer(nn.FullyConnectedLayer(hidden_layer,output_layer,'relu'))
+	nn1.addLayer(nn.FullyConnectedLayer(hidden_layer,output_layer,'softmax'))
 
 	nn1.train(XTrain, YTrain, XVal, YVal)
 	pred, acc = nn1.validate(XTest, YTest)
@@ -48,7 +48,7 @@ def preprocessMnist(X):
 def taskMnist():
 	XTrain, YTrain, XVal, YVal, XTest, _ = loadMnist()
 	# Create a NeuralNetwork object 'nn1' as follows with optimal parameters. For parameter definition, refer to py file.
-	lr,batchSize,epochs = 0.005,256,10000
+	lr,batchSize,epochs = -0.005,256,10000
 	nn1 = nn.NeuralNetwork(lr, batchSize, epochs)
 	# Add layers to neural network corresponding to inputs and outputs of given data
 	input_layer = XTrain.shape[1]

Assignment2/nn.py

@@ -260,8 +260,8 @@ class FullyConnectedLayer:
 		# This function should actually update the weights using the gradients computed in the backwardpass
 		###############################################
 		# TASK 1h (Marks 2) - YOUR CODE HERE
-		self.weights += lr*(self.weightsGrad)
-		self.biases += lr*(self.biasesGrad)
+		self.weights -= lr*(self.weightsGrad)
+		self.biases -= lr*(self.biasesGrad)
     
 # 		raise NotImplementedError
 		###############################################