使用tf.Session()时未对模型进行训练

import tensorflow as tf

# 定义模型
x = tf.placeholder(tf.float32, shape=[None, input_dim])
y_true = tf.placeholder(tf.float32, shape=[None, output_dim])
weights = tf.Variable(tf.zeros([input_dim, output_dim]))
biases = tf.Variable(tf.zeros([output_dim]))
y_pred = tf.matmul(x, weights) + biases

# 定义损失函数和优化算法
loss = tf.reduce_mean(tf.square(y_true - y_pred))
optimizer = tf.train.GradientDescentOptimizer(learning_rate)
train_op = optimizer.minimize(loss)

# 创建会话并初始化变量
with tf.Session() as sess:
    sess.run(tf.global_variables_initializer())

    # 执行训练步骤
    for i in range(num_epochs):
        sess.run(train_op, feed_dict={x: input_data, y_true: target_data})

    # 在训练完成后，可以使用训练好的模型进行预测或评估
    y_pred_val = sess.run(y_pred, feed_dict={x: test_data})