无法挤压dim[1]，预期尺寸为1，输入形状为[?,4] 的'metrics/accuracy/Squeeze'(操作:'Squeeze')为4 [英] Can not squeeze dim[1], expected a dimension of 1, got 4 for 'metrics/accuracy/Squeeze' (op: 'Squeeze') with input shapes: [?,4]

问题描述

张量流 - 2.1.0

Tensorflow - 2.1.0

Python - 3.6

Python - 3.6

我已经在 stackoverflow 上搜索过这个问题，但找不到解决方案.

I have searched this issue on stackoverflow but could not find solution.

我正在尝试使用 tensorflow 创建一个聊天机器人.这是错误:

I am trying to create a chatbot using tensorflow. This is error:

不能挤压dim[1]，期望维度为1，得到4个'metrics/accuracy/Squeeze'(操作:'Squeeze')，输入形状:[?,4].

Can not squeeze dim[1], expected a dimension of 1, got 4 for 'metrics/accuracy/Squeeze' (op: 'Squeeze') with input shapes: [?,4].

这是代码:

words = []
    classes = []
    documents = []
    ignore_words = ['?', '!']
    data_file = open('fil.json').read()
    intents = json.loads(data_file)
    for intent in intents['intents']:
        for pattern in intent['question']:
            w = nltk.word_tokenize(pattern)
            words.extend(w)
            documents.append((w, intent['tag']))
            if intent['tag'] not in classes:
                classes.append(intent['tag'])
    words = [lemmatizer.lemmatize(w.lower()) for w in words if w not in ignore_words]
    words = sorted(list(set(words)))
    classes = sorted(list(set(classes)))
    pickle.dump(words, open('words.pkl', 'wb'))
    pickle.dump(classes, open('classes.pkl', 'wb'))
    training = []
    output_empty = [0] * len(classes)
    for doc in documents:
        bag = []
        pattern_words = doc[0]
        pattern_words = [lemmatizer.lemmatize(word.lower()) for word in pattern_words]
        for w in words:
            bag.append(1) if w in pattern_words else bag.append(0)
        output_row = list(output_empty)
        output_row[classes.index(doc[1])] = 1

        training.append([bag, output_row])
    random.shuffle(training)
    training = np.array(training)
    train_x = list(training[:, 0])
    train_y = list(training[:, 1])
    print("Training data created")
    model = tf.keras.models.Sequential([
        tf.keras.layers.Dense(128, input_shape=(len(train_x[0]),), activation='relu'),
        tf.keras.layers.Dropout(0.2),
        tf.keras.layers.Dense(len(train_y[0]), activation='softmax')
    ])
    model.compile(optimizer='adam',
                  loss='sparse_categorical_crossentropy',
                  metrics=['accuracy'])
    hist = model.fit(train_x, train_y, epochs=5)
    model.save('chatbot_model.h5', hist)

    print("model created")

推荐答案

确实，由于您的模型使用 softmax 密集层作为输出，您应该将其损失设置为 categorical_crossentropy.之后您得到的错误是已知问题.有几个选项可以帮助您解决此问题:

Indeed, since your model is using softmax dense layer as output, you should set its loss to categorical_crossentropy. The error you get after that is a known issue. There's a few options that could help you to fix it:

• 检查您的 Python 环境.确保您没有与 TensorFlow 一起安装 Keras，并且您的 Keras-Applications 和 Keras-Preprocessing 包是最新的.
• 在 model.fit() 中调用集合 workers=0.
• 尝试将您的 TensorFlow 软件包降级到 2.0.1.

• Check your Python environment. Ensure that you don't have Keras installed along with TensorFlow and that your Keras-Applications and Keras-Preprocessing packages are up to date.
• In model.fit() call set workers=0.
• Try to downgrade your TensorFlow package to 2.0.1.

这篇关于无法挤压dim[1]，预期尺寸为1，输入形状为[?,4] 的'metrics/accuracy/Squeeze'(操作:'Squeeze')为4的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！