我为此奋斗了很多。面临的挑战是，我可能会在一种样式上绘制几个不同的图层，然后调用 balance_dataframe = pd.read_csv("balance.csv", sep=",")
balance_dataframe = balance_dataframe.reindex(np.random.permutation(balance_dataframe.index))
balance_dataframe = balance_dataframe[balance_dataframe["Class"] != "B"]
data = preprocessFeatures(balance_dataframe)
target = preprocessClasses(balance_dataframe)
train_features = data.head(500)
train_classes = target.head(500)
test_features = data.tail(200)
test_classes = target.tail(200)
_ = train_model(
learning_rate=0.0001,
steps=100,
batch_size=70,
hidden_units=[10, 10],
training_examples=train_features,
training_targets=train_classes,
validation_examples=test_features,
validation_targets=test_classes)
（例如，进入卫星视图），这会擦除我的所有图层。
答案 0 :(得分：3)
我通过复制我关心的图层和源并重新应用来解决它
setStyle