如何解决tensorflow中的功能列不推荐警告

我的tensorflow版本是2.4.1，当我想将数据传输到特征列并创建估计器时，终端说我的特征列API不推荐使用。我感到很困惑。有人能帮我吗？谢谢

这是我的代码：

csv_file = 'train.csv'
csv_data = pd.read_csv(csv_file, low_memory = False)
csv_df = pd.DataFrame(csv_data) 

CONTI_FEATURES = ['Age', 'Smoking']
CATE_FEATURES = ['Gender', 'ICD9Code']

def print_transformation(feature, continuous, size):
   
    feature_names = [feature]

    d = dict(zip(feature_names, [csv_data[feature]]))

    if continuous == True:
        c = tf.feature_column.numeric_column(feature)
        feature_columns = [c]
    else:
        c = tf.feature_column.categorical_column_with_hash_bucket(feature, hash_bucket_size=size)
        c_indicator = tf.feature_column.indicator_column(c)
        feature_columns = [c_indicator]

## Use input_layer to print the value
    input_layer = tf.feature_column.input_layer(
        features=d,
        feature_columns=feature_columns
        )
    
    return input_layer

Age_feature = print_transformation(feature = "Age", continuous = True, size = 2)
Smoking_feature = print_transformation(feature = "Smoking", continuous = True, size = 2)
Gender_feature = print_transformation(feature = "Gender", continuous = False, size = 2)
Diagnosis_feature = print_transformation(feature = "ICD9Code", continuous = False, size = 1000)

continuous_features = [Age_feature, Smoking_feature]
categorical_features = [Gender_feature, Diagnosis_feature]

model = tf.estimator.LinearClassifier(
  n_classes = 2,
  model_dir = "ongoing/train",
  feature_columns = categorical_features + continuous_features
)

---

需要使用弃用警告消息来通知希望使用TensorFlow实际版本的开发人员将来会发生什么，并给他们时间进行调整

不用担心弃用警告，它不会影响代码性能。 您可以使用以下任一代码段关闭警告

import logging, os
logging.disable(logging.WARNING)
os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"
import tensorflow as tf

另一种选择是使用下面的代码

import tensorflow as tf
tf.compat.v1.logging.set_verbosity(tf.compat.v1.logging.ERROR)