Tf2.0中没有keras.fit或estimator的Tensorboard显示标量

# 导入相关库
import tensorflow as tf
from tensorflow.keras import layers

# 创建一个FileWriter对象，指定记录日志的目录
log_dir = 'logs'
file_writer = tf.summary.create_file_writer(log_dir)

# 定义模型和优化器
model = tf.keras.Sequential([
    layers.Dense(64, activation='relu'),
    layers.Dense(10, activation='softmax')
])
optimizer = tf.keras.optimizers.Adam()

# 定义损失函数和评价指标
loss_fn = tf.keras.losses.SparseCategoricalCrossentropy()
accuracy_metric = tf.keras.metrics.SparseCategoricalAccuracy()

# 开始训练循环
for epoch in range(num_epochs):
    for batch, (x, y) in enumerate(train_dataset):
        with tf.GradientTape() as tape:
            logits = model(x)
            loss = loss_fn(y, logits)
        
        # 计算梯度并更新模型参数
        grads = tape.gradient(loss, model.trainable_variables)
        optimizer.apply_gradients(zip(grads, model.trainable_variables))
        
        # 更新评价指标
        accuracy_metric.update_state(y, logits)
        
        # 手动记录训练损失和准确率
        with file_writer.as_default():
            tf.summary.scalar('train_loss', loss, step=epoch*num_batches+batch)
            tf.summary.scalar('train_accuracy', accuracy_metric.result(), step=epoch*num_batches+batch)
        
        # 清除评价指标的状态
        accuracy_metric.reset_states()

# 导入相关库
import tensorflow as tf
from tensorflow.keras import layers

# 创建一个FileWriter对象，指定记录日志的目录
log_dir = 'logs'
file_writer = tf.summary.create_file_writer(log_dir)

# 定义模型和优化器
model = tf.keras.Sequential([
    layers.Dense(64, activation='relu'),
    layers.Dense(10, activation='softmax')
])
optimizer = tf.keras.optimizers.Adam()

# 定义损失函数和评价指标
loss_fn = tf.keras.losses.SparseCategoricalCrossentropy()
accuracy_metric = tf.keras.metrics.SparseCategoricalAccuracy()

# 开始profile记录
tf.profiler.experimental.start(log_dir)

# 开始训练循环
for epoch in range(num_epochs):
    for batch, (x, y) in enumerate(train_dataset):
        with tf.GradientTape() as tape:
            logits = model(x)
            loss = loss_fn(y, logits)
        
        # 计算梯度并更新模型参数
        grads = tape.gradient(loss, model.trainable_variables)
        optimizer.apply_gradients(zip(grads, model.trainable_variables))
        
        # 更新评价指标
        accuracy_metric.update_state(y, logits)
        
        # 清除评价指标的状态
        accuracy_metric.reset_states()

# 停止profile记录
tf.profiler.experimental.stop()