TF2 可视化Loss函数，并且导出模型图

pip安装依赖pydot和graphviz并且安装软件sudo apt install graphviz，有个坑，windows安装软件之后安装的依赖是pydot-ng

注意：模型的第一层需要把形状传进去

import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
import datetime

import tensorflow as tf
import numpy as np
from tensorflow import keras
from tensorflow.keras import datasets, layers, optimizers, Sequential, metrics
from tensorflow.keras.utils import plot_model

conv_layers = [
    layers.Conv2D(64, kernel_size=[3, 3], padding='same', activation=tf.nn.relu, input_shape=[32, 32, 3]),
    layers.Conv2D(64, kernel_size=[3, 3], padding='same', activation=tf.nn.relu),
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same'),

    layers.Conv2D(128, kernel_size=[3, 3], padding='same', activation=tf.nn.relu),
    layers.Conv2D(128, kernel_size=[3, 3], padding='same', activation=tf.nn.relu),
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same'),

    layers.Conv2D(256, kernel_size=[3, 3], padding='same', activation=tf.nn.relu),
    layers.Conv2D(256, kernel_size=[3, 3], padding='same', activation=tf.nn.relu),
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same'),

    layers.Conv2D(512, kernel_size=[3, 3], padding='same', activation=tf.nn.relu),
    layers.Conv2D(512, kernel_size=[3, 3], padding='same', activation=tf.nn.relu),
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same'),

    layers.Conv2D(512, kernel_size=[3, 3], padding='same', activation=tf.nn.relu),
    layers.Conv2D(512, kernel_size=[3, 3], padding='same', activation=tf.nn.relu),
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same'),

    layers.Flatten(),

    layers.Dense(256, activation=tf.nn.relu),
    layers.Dense(128, activation=tf.nn.relu),
    layers.Dense(100, activation=None),
]

def preprocess(x, y):
    x = tf.cast(x, dtype=tf.float32) /255.
    y = tf.cast(y, dtype=tf.int32)
    return x, y

(x, y), (x_test, y_test) = datasets.cifar100.load_data()
# (50000, 32, 32, 3) (50000, 1) (10000, 32, 32, 3) (10000, 1)
y = tf.squeeze(y, axis=1)
y_test = tf.squeeze(y_test, axis=1)

train_db = tf.data.Dataset.from_tensor_slices((x, y))
train_db = train_db.shuffle(1000).map(preprocess).batch(64)

test_db = tf.data.Dataset.from_tensor_slices((x_test, y_test))
test_db = test_db.map(preprocess).batch(64)


logdir = "logs/" + datetime.datetime.now().strftime("%Y%m%d-%H%M%S")
writer = tf.summary.create_file_writer(logdir=logdir)

def main():
    model = Sequential(conv_layers)
    model.summary()
    plot_model(model=model, to_file="model.png", show_shapes=True, dpi=300)


    variables = model.trainable_variables
    optimizer = optimizers.Adam(lr=1e-4)

    for epoch in range(10):
        for step, (x, y) in enumerate(train_db):
            with tf.GradientTape() as tape:
                logits = model(x)
                y_onehot = tf.one_hot(y, depth=100)
                loss = tf.losses.categorical_crossentropy(y_onehot, logits, from_logits=True)
                loss = tf.reduce_mean(loss)

            grads = tape.gradient(loss, variables)
            optimizer.apply_gradients(zip(grads, variables))

            if step % 100 == 0:
                print(epoch, step, 'loss', float(loss))

        with writer.as_default():
            tf.summary.scalar("train_loss", loss, epoch)


        totol_num = 0
        totol_correct = 0
        for x, y in test_db:
            logits = model(x)
            prob = tf.nn.softmax(logits, axis=1)
            pred = tf.argmax(prob, axis=1)
            pred = tf.cast(pred, dtype=tf.int32)

            correct = tf.cast(tf.equal(pred, y), dtype=tf.int32)
            correct = tf.reduce_sum(correct)
            totol_num += x.shape[0]
            totol_correct += int(correct)

        acc = totol_correct / totol_num
        print(epoch, 'acc:', acc)

        with writer.as_default():
            tf.summary.scalar("val_acc", acc, epoch)


if __name__ == '__main__':
    main()