Question

我必须提交我的预测（每个预测都是2289个班级之一！）。但是，我不知道该怎么做。救救我！

模型看起来像这样：

model = Sequential()
model.add(Conv2D(32, kernel_size = (3, 3), input_shape=(32,64,1), activation = 'relu'))
model.add(Conv2D(64, (3,3), activation='relu'))
model.add(MaxPooling2D(pool_size=2))
model.add(Dropout(0.25))
model.add(Flatten())
model.add(Dense(2000, activation = 'relu'))
model.add(Dropout(0.6))
model.add(Dense(2289, activation='softmax'))
model.summary()
model.compile(loss = 'sparse_categorical_crossentropy', optimizer = 'adam', metrics = ['accuracy'])

我使用了拟合函数

history = model.fit(trainimage,trainlabel, validation_data=(validimage,validlabel), epochs=10, batch_size=200)

我想知道的是如何提取给定测试图像的预测

# Indexing of predictions
argpreds = np.argmax(history)

# Load label dictionary
with open('./drive/My Drive/Colab Notebooks/CoE202_KakaoArena/y_vocab.pickle', 'rb') as f:
    y_dict = pickle.load(f)

# Inverse label dictionary
inv_y_dict = dict((y,x) for x,y in y_dict.items())
submissions = [inv_y_dict[argpred] for argpred in argpreds]

# Write the results to 'submissions.csv'
f = open('./drive/My Drive/Colab Notebooks/CoE202_KakaoArena/submissions.csv', 'w')
for i, j in zip(pids, submissions):
    line = '{},{}\n'.format(i,j)
    f.write(line)
f.close()

这是我助教代码的修改。但是我收到一个错误“ TypeError：'numpy.int64'对象不可迭代”

感谢您的帮助

Answer 1

通常，当您使用列表等变量时，错误对象是不可迭代的。因此，它的唯一一维数据类型（例如整型列表）。

如何提取模型的输出？

1 个答案: