http://www-cvr.ai.uiuc.edu/ponce_grp/data/ 我用birds那個 #以下是code import os import glob # 查找文件 from keras.models import Sequential from keras.layers.core import Flatten, Dense, Dropout from keras.layers.convolutional import Convolution2D, MaxPooling2D, ZeroPadding2D from keras.optimizers import SGD import numpy as np from pandas import Series, DataFrame from keras.preprocessing.image import ImageDataGenerator from keras.models import Sequential from keras.layers.core import Dense, Dropout, Activation, Flatten from keras.layers.advanced_activations import PReLU from keras.layers.convolutional import Convolution2D, MaxPooling2D from keras.optimizers import SGD, Adadelta, Adagrad from keras.utils import np_utils, generic_utils from six.moves import range def load_data(): sed = 1000 data = np.empty((540,150,150,3),dtype="float32") label = np.empty((540,)) imgs = os.listdir(r"C:\Users\user\Desktop\birds\train") num = len(imgs) times = 0 time = 0 for i in range(num): if "egr" in imgs[i]: img = Image.open("C:/Users/user/Desktop/birds/train/" + imgs[i]) arr = np.asarray(img, dtype="float32") arr.resize((150,150,3)) data[i, :, :, :] = arr label[i] = 0 #times +=1 elif "man" in imgs[i]: img = Image.open("C:/Users/user/Desktop/birds/train/" + imgs[i]) arr = np.asarray(img, dtype="float32") arr.resize((150,150,3)) data[i, :, :, :] = arr label[i] = 1 #times +=1 elif "owl" in imgs[i]: img = Image.open("C:/Users/user/Desktop/birds/train/" + imgs[i]) arr = np.asarray(img, dtype="float32") arr.resize((150,150,3)) data[i, :, :, :] = arr label[i] = 2 #times +=1 elif "puf" in imgs[i]: img = Image.open("C:/Users/user/Desktop/birds/train/" + imgs[i]) arr = np.asarray(img, dtype="float32") arr.resize((150,150,3)) data[i, :, :, :] = arr label[i] = 3 #times +=1 elif "tou" in imgs[i]: img = Image.open("C:/Users/user/Desktop/birds/train/" + imgs[i]) arr = np.asarray(img, dtype="float32") arr.resize((150,150,3)) data[i, :, :, :] = arr label[i] = 4 #times +=1 elif "wod" in imgs[i]: img = Image.open("C:/Users/user/Desktop/birds/train/" + imgs[i]) arr = np.asarray(img, dtype="float32") arr.resize((150,150,3)) data[i, :, :, :] = arr label[i] = 5 #times +=1 else: break return data,label train_data,train_labels = load_data() def load_data1(): sed = 1000 data = np.empty((60,150,150,3),dtype="float32") label = np.empty((60,)) imgs = os.listdir(r"C:\Users\user\Desktop\birds\test") num = len(imgs) times = 0 time = 0 for i in range(num): if "egr" in imgs[i]: img = Image.open("C:/Users/user/Desktop/birds/test/" + imgs[i]) arr = np.asarray(img, dtype="float32") arr.resize((150,150,3)) data[i, :, :, :] = arr label[i] = 0 #times +=1 elif "man" in imgs[i]: img = Image.open("C:/Users/user/Desktop/birds/test/" + imgs[i]) arr = np.asarray(img, dtype="float32") arr.resize((150,150,3)) data[i, :, :, :] = arr label[i] = 1 #times +=1 elif "owl" in imgs[i]: img = Image.open("C:/Users/user/Desktop/birds/test/" + imgs[i]) arr = np.asarray(img, dtype="float32") arr.resize((150,150,3)) data[i, :, :, :] = arr label[i] = 2 #times +=1 elif "puf" in imgs[i]: img = Image.open("C:/Users/user/Desktop/birds/test/" + imgs[i]) arr = np.asarray(img, dtype="float32") arr.resize((150,150,3)) data[i, :, :, :] = arr label[i] = 3 #times +=1 elif "tou" in imgs[i]: img = Image.open("C:/Users/user/Desktop/birds/test/" + imgs[i]) arr = np.asarray(img, dtype="float32") arr.resize((150,150,3)) data[i, :, :, :] = arr label[i] = 4 #times +=1 elif "wod" in imgs[i]: img = Image.open("C:/Users/user/Desktop/birds/test/" + imgs[i]) arr = np.asarray(img, dtype="float32") arr.resize((150,150,3)) data[i, :, :, :] = arr label[i] = 5 #times +=1 else: break return data,label datatest,labeltest = load_data1() y_train=np_utils.to_categorical(train_labels,num_classes=6) #np.utils.to_categorical : 若 y＝2 －> [[ 0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]] y_test=np_utils.to_categorical(labeltest,num_classes=6) model = Sequential() model.add(Convolution2D(32, (3, 3),input_shape=(150, 150,3))) model.add(Activation('relu')) model.add(MaxPooling2D(pool_size=(2, 2))) model.add(Convolution2D(64, (3, 3))) model.add(Activation('relu')) model.add(MaxPooling2D(pool_size=(2, 2))) model.add(Convolution2D(64, (3, 3))) model.add(Activation('relu')) model.add(MaxPooling2D(pool_size=(2, 2))) model.add(Flatten()) model.add(Dense(64)) model.add(Activation('relu')) model.add(Dropout(0.5)) model.add(Dense(6)) model.add(Activation('softmax')) model.compile(loss='categorical_crossentropy', optimizer='rmsprop', metrics=['accuracy']) model.fit(train_data, y_train, nb_epoch=50, batch_size=20, validation_data=(datatest, y_test)) 我跑出的結果長這樣 我覺得很怪 Epoch 1/50 540/540 [==============================] - 21s 40ms/step - loss: 2.0500 - acc: 0.1630 - val_loss: 1.9806 - val_acc: 0.1667 Epoch 2/50 540/540 [==============================] - 21s 39ms/step - loss: 1.9121 - acc: 0.1574 - val_loss: 1.8918 - val_acc: 0.1667 Epoch 3/50 540/540 [==============================] - 21s 39ms/step - loss: 1.8998 - acc: 0.1704 - val_loss: 1.8457 - val_acc: 0.1667 Epoch 4/50 540/540 [==============================] - 21s 39ms/step - loss: 1.8353 - acc: 0.1704 - val_loss: 1.8256 - val_acc: 0.1667 Epoch 5/50 540/540 [==============================] - 21s 38ms/step - loss: 1.8449 - acc: 0.1574 - val_loss: 1.8175 - val_acc: 0.1667 Epoch 6/50 540/540 [==============================] - 21s 38ms/step - loss: 1.8065 - acc: 0.1796 - val_loss: 1.8098 - val_acc: 0.1667 Epoch 7/50 540/540 [==============================] - 21s 39ms/step - loss: 1.8136 - acc: 0.1778 - val_loss: 1.8035 - val_acc: 0.1667 Epoch 8/50 540/540 [==============================] - 22s 41ms/step - loss: 1.8419 - acc: 0.1852 - val_loss: 1.7981 - val_acc: 0.1667 Epoch 9/50 540/540 [==============================] - 21s 39ms/step - loss: 1.8219 - acc: 0.1685 - val_loss: 1.7956 - val_acc: 0.1667 Epoch 10/50 540/540 [==============================] - 22s 40ms/step - loss: 1.7954 - acc: 0.1889 - val_loss: 1.7936 - val_acc: 0.1667 Epoch 11/50 540/540 [==============================] - 22s 40ms/step - loss: 1.8192 - acc: 0.1741 - val_loss: 1.7926 - val_acc: 0.1667 Epoch 12/50 540/540 [==============================] - 21s 40ms/step - loss: 1.8244 - acc: 0.1741 - val_loss: 1.7921 - val_acc: 0.1667 Epoch 13/50 540/540 [==============================] - 23s 42ms/step - loss: 1.8233 - acc: 0.1815 - val_loss: 1.7919 - val_acc: 0.1667 Epoch 14/50 540/540 [==============================] - 22s 40ms/step - loss: 1.7947 - acc: 0.1889 - val_loss: 1.7919 - val_acc: 0.1667 Epoch 15/50 540/540 [==============================] - 21s 39ms/step - loss: 1.7991 - acc: 0.2019 - val_loss: 1.7919 - val_acc: 0.1667 Epoch 16/50 540/540 [==============================] - 21s 39ms/step - loss: 1.8107 - acc: 0.2037 - val_loss: 1.7923 - val_acc: 0.1667 Epoch 17/50 540/540 [==============================] - 21s 40ms/step - loss: 1.8087 - acc: 0.1815 - val_loss: 1.7945 - val_acc: 0.1500 Epoch 18/50 540/540 [==============================] - 21s 39ms/step - loss: 1.8248 - acc: 0.2019 - val_loss: 1.7977 - val_acc: 0.1667 Epoch 19/50 540/540 [==============================] - 21s 39ms/step - loss: 1.8012 - acc: 0.1611 - val_loss: 1.7968 - val_acc: 0.1667 Epoch 20/50 540/540 [==============================] - 21s 39ms/step - loss: 1.8415 - acc: 0.1796 - val_loss: 1.7933 - val_acc: 0.1667 Epoch 21/50 540/540 [==============================] - 21s 39ms/step - loss: 1.8390 - acc: 0.1778 - val_loss: 1.8040 - val_acc: 0.1667 Epoch 22/50 540/540 [==============================] - 21s 39ms/step - loss: 1.8363 - acc: 0.1778 - val_loss: 1.8007 - val_acc: 0.1667 Epoch 23/50 540/540 [==============================] - 21s 39ms/step - loss: 1.8053 - acc: 0.1630 - val_loss: 1.7787 - val_acc: 0.1500 Epoch 24/50 540/540 [==============================] - 21s 39ms/step - loss: 1.8064 - acc: 0.1704 - val_loss: 1.7940 - val_acc: 0.1667 Epoch 25/50 540/540 [==============================] - 21s 39ms/step - loss: 1.8006 - acc: 0.1481 - val_loss: 1.7979 - val_acc: 0.1667 Epoch 26/50 540/540 [==============================] - 21s 39ms/step - loss: 1.8004 - acc: 0.1630 - val_loss: 1.7972 - val_acc: 0.1667 Epoch 27/50 540/540 [==============================] - 21s 39ms/step - loss: 1.8358 - acc: 0.1870 - val_loss: 1.7968 - val_acc: 0.1667 Epoch 28/50 540/540 [==============================] - 21s 39ms/step - loss: 1.7930 - acc: 0.1759 - val_loss: 1.7969 - val_acc: 0.1667 Epoch 29/50 540/540 [==============================] - 23s 42ms/step - loss: 1.8084 - acc: 0.1648 - val_loss: 1.7971 - val_acc: 0.1667 主要是這邊都維持一樣val_acc: 0.1667 我覺得好像不太對, 請問有高人遇過這樣問題嗎? 謝謝回答 --

※ 發信站: 批踢踢實業坊(ptt.cc), 來自: 140.116.234.34 ※ 文章網址: https://webptt.com/m.aspx?n=bbs/Python/M.1532189728.A.D8D.html