我將許多長(zhǎng)度為 100 和 3 個(gè)特征的時(shí)間序列輸入到 1D Convnet 中。我有太多這些無(wú)法使用 numpy 數(shù)組，因此我需要使用 Dataset.from_generator()。問(wèn)題是當(dāng)我在數(shù)據(jù)集上訓(xùn)練模型時(shí)，它給出了錯(cuò)誤：expected conv1d_input to have 3 dimensions, but got array with shape (100, 3)下面的代碼演示了這個(gè)問(wèn)題。生成器將每個(gè)元素生成為預(yù)期的 (100,3) 數(shù)組。為什么模型無(wú)法將生成器輸出識(shí)別為有效？非常感謝您的幫助。朱利安import numpy as npimport tensorflow as tfdef create_timeseries_element(): # returns a random time series of 100 intervals, each with 3 features, # and a random one-hot array of 5 entries data = np.random.rand(100,3) label = np.eye(5, dtype='int')[np.random.choice(5)] return data, labeldef data_generator(): d, l = create_timeseries_element() yield (d, l)model = tf.keras.models.Sequential([ tf.keras.layers.Conv1D(128, 9, activation='relu', input_shape=(100, 3)), tf.keras.layers.Conv1D(128, 9, activation='relu'), tf.keras.layers.MaxPooling1D(2), tf.keras.layers.Conv1D(256, 5, activation='relu'), tf.keras.layers.Conv1D(256, 5, activation='relu'), tf.keras.layers.GlobalAveragePooling1D(), tf.keras.layers.Dropout(0.5), tf.keras.layers.Dense(5, activation='softmax')])model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])x_train = []y_train = []for _ in range(1000): d, l = create_timeseries_element() x_train.append(d) y_train.append(l)x_train = np.array(x_train)y_train = np.array(y_train)# train model with numpy arrays - this worksmodel.fit(x=x_train, y=y_train)ds = tf.data.Dataset.from_generator(data_generator, output_types=(tf.float32, tf.int32), output_shapes=(tf.TensorShape([100, 3]), tf.TensorShape([5])))# train model with dataset - this failsmodel.fit(ds)
1 回答
慕哥9229398
TA貢獻(xiàn)1877條經(jīng)驗(yàn) 獲得超6個(gè)贊
模型需要一批/樣本列表。您可以通過(guò)在創(chuàng)建數(shù)據(jù)集時(shí)簡(jiǎn)單地設(shè)置批處理屬性來(lái)做到這一點(diǎn)，如下所示：
ds = tf.data.Dataset.from_generator(data_generator, output_types=(tf.float32, tf.int32),
output_shapes=(tf.TensorShape([100, 3]), tf.TensorShape([5])))
ds = ds.batch(16)
您也可以在準(zhǔn)備樣品時(shí)采用另一種方式。這樣，您需要擴(kuò)展樣本維度，以便樣本充當(dāng)批次（您也可以傳遞樣本列表）并且您必須在output_shapes數(shù)據(jù)集和create_timeseries_element函數(shù)中進(jìn)行以下修改
def create_timeseries_element():
# returns a random time series of 100 intervals, each with 3 features,
# and a random one-hot array of 5 entries
# Expand dimensions to create a batch of single sample
data = np.expand_dims(np.random.rand(100, 3), axis=0)
label = np.expand_dims(np.eye(5, dtype='int')[np.random.choice(5)], axis=0)
return data, label
ds = tf.data.Dataset.from_generator(data_generator, output_types=(tf.float32, tf.int32), output_shapes=(tf.TensorShape([None, 100, 3]), tf.TensorShape([None, 5])))
上述更改將為數(shù)據(jù)集的每個(gè)時(shí)期僅提供一個(gè)批次（第一個(gè)解決方案的樣本）。您可以通過(guò)在定義數(shù)據(jù)集時(shí)將參數(shù)傳遞給data_generator函數(shù)來(lái)生成所需的批次（第一個(gè)解決方案的樣本）（例如 25 個(gè)），如下所示：
def data_generator(count=1):
for _ in range(count):
d, l = create_timeseries_element()
yield (d, l)
ds = tf.data.Dataset.from_generator(data_generator, args=[25], output_types=(tf.float32, tf.int32), output_shapes=(tf.TensorShape([None, 100, 3]), tf.TensorShape([None, 5])))
添加回答
舉報(bào)
0/150
提交
取消