保存与加载权重数据

深度学习笔记/前馈神经网络/feedforward_neural_network.py

@@ -17,7 +17,7 @@ def relu(z):
 L = 0
 
 ##### neural network model
-def neural_network(X, Y, learning_rate=0.01, num_iterations=8000, lambd=0):
+def neural_network(X, Y, learning_rate=0.01, num_iterations=2000, lambd=0):
     n0, m = X.shape
     n1 = 20
     n2 = 7
@@ -124,21 +124,33 @@ def vectorize_sequences(sequences, dimension=10000):
         results[i, sequence] = 1.  # 索引results矩阵中的位置，赋值为1，全部都是从第0行0列开始的
     return results
 
+def load(path="./weigths.npz"):
+    weigths = np.load(path, allow_pickle=True) 
+    return weigths
+
+def save(weigths, path="./weigths.npz"):
+    np.savez(path, param_w=weigths["param_w"], param_b=weigths["param_b"])
+    ### np.savetxt("./weigths.out", weigths["param_w"], fmt='%s')
+
 (train_data, train_labels), (test_data, test_labels) = imdb.load_data(path="imdb/imdb.npz",num_words=10000)
 
 # Our vectorized training data
-x_train = vectorize_sequences(train_data[:1000])
+x_train = vectorize_sequences(train_data[:100])
 # Our vectorized test data
 x_test = vectorize_sequences(test_data[:5])
 
-y_train = np.asarray(train_labels[:1000]).astype('float32')
+y_train = np.asarray(train_labels[:100]).astype('float32')
 y_test = np.asarray(test_labels[:5]).astype('float32')
 
-parameters = neural_network(x_train.T, y_train)
-y_p = predict(x_test.T, parameters)
+parameters = neural_network(x_train.T, y_train, num_iterations=1000)
+save(parameters)
+weigths = load()
+
+y_p = predict(x_test.T, weigths)
 print(y_p)
 print(y_test)
-y_p2 = predict(x_train[:5].T, parameters)
+
+y_p2 = predict(x_train[:5].T, weigths)
 print(y_p2)
 print(y_train[:5])