simple softmax

.gitignore

@@ -139,3 +139,5 @@ dmypy.json
 cython_debug/
 
 sgd_hw/
+
+mnist.pickle

layer.py

@@ -188,8 +188,10 @@ class SoftmaxWithNegativeLogLikelihood(OpTree):
     def __init__(self, i, y):
         super().__init__()
         self.i = i
-        self.s = softmaxHelp(i)
-        self.v = -y*self.s
+        self.s = softmaxHelp(i.numpy())
+        self.y = y
+        self.v = -y*np.log(self.s)
+        self.v = np.sum(self.v,axis=self.v.ndim-1)
     
     def __str__(self):
         return f"SoftmaxWithNegativeLogLikelihoodOp"
@@ -205,7 +207,7 @@ class SoftmaxWithNegativeLogLikelihood(OpTree):
         return self.s
 
     def backprop(self,seed):
-        self.i.backprop(seed * (self.s-y))
+        self.i.backprop(seed * (self.s-self.y))
 
 class Variable(OpTree):
     def __init__(self,x):

p4.py

@@ -1,10 +1,68 @@
 from sklearn import datasets
 import numpy as np
-X, y = datasets.fetch_openml('mnist_784', version=1, return_X_y=True, cache=True, as_frame= False)
-print(X,y)
+from layer import *
+import os
+import pickle
+import matplotlib
+import matplotlib.pyplot as plt
+import random
+
+matplotlib.use("TkAgg")
+
+PICKLE_DATA_FILENAME = "mnist.pickle"
+if not os.path.exists(PICKLE_DATA_FILENAME):
+    X, y = datasets.fetch_openml('mnist_784', return_X_y=True, cache=True, as_frame= False)
+    with open(PICKLE_DATA_FILENAME,"wb") as file:
+        pickle.dump(X,file)
+        pickle.dump(y,file)
+else:
+    with open(PICKLE_DATA_FILENAME,"rb") as file:
+        X = pickle.load(file)
+        y = pickle.load(file)
+
+i = random.randint(0,len(X) - 1)
+#plt.imshow(X[0].reshape(28,28),cmap='gray',interpolation='none')
+#plt.show()
+
+#simple normalize
+X = X / 255
+
+y = np.array([int(i) for i in y])
+Y = np.eye(10)[y]
 
 gen:np.random.Generator = np.random.default_rng()
+eta = 0.01
 
-input_var = Variable(X)
-weight = Variable(gen.normal(100,784))
-bias = Variable(np.array([1]))
+MiniBatchN = 100
+
+weight1 = Variable(gen.normal(0,1,size=(784,10)))
+bias1 = Variable(gen.normal(0,1,size=(10)))
+
+accuracy_list = []
+
+for iteration in range(0,100):
+    choiced_index = gen.choice(range(0,60000),MiniBatchN)
+    input_var = Variable(X[choiced_index])
+    U1 = (input_var @ weight1 + bias1)
+    J = SoftmaxWithNegativeLogLikelihood(U1,Y[choiced_index])
+
+    J.backprop(np.ones(()))
+    #update variable
+    weight1 = Variable(weight1.numpy() - (weight1.grad) * eta)
+    bias1 = Variable(bias1.numpy() - (bias1.grad) * eta)
+    if iteration % 5 == 0:
+        print(iteration,'iteration : avg(J) == ',np.average(J.numpy()))
+        s = J.softmax_numpy()
+        #print(Y[0:1000].shape)
+        s = np.round(s)
+        confusion = (np.transpose(Y[choiced_index])@s)
+        accuracy = np.trace(confusion).sum() / MiniBatchN
+        print('accuracy : ',accuracy * 100,'%')
+        accuracy_list.append(accuracy)
+
+plt.title("accuracy")
+plt.plot(np.linspace(0,len(accuracy_list),len(accuracy_list)),accuracy_list)
+plt.show()
+plt.title("confusion matrix")
+plt.imshow(confusion,cmap='gray')
+plt.show()