resnet18

最新推荐文章于 2024-07-14 13:36:04 发布

菜鸟鱼

最新推荐文章于 2024-07-14 13:36:04 发布

阅读量829

点赞数 1

分类专栏： tensorflow

本文链接：https://blog.csdn.net/y786256773/article/details/109641517

版权

本文详细介绍了ResNet18模型的架构和工作原理，重点探讨了残差块在解决深度网络退化问题中的作用。通过使用Python3.6和TensorFlow1.4实现ResNet18，展示了如何在深度学习项目中应用该模型。

摘要由CSDN通过智能技术生成

python3.6 tensorflow1.4

import tensorflow as tf
import numpy as np
class resnet:
    def __init__(self, input, calsses,TRAINING):
        self.X = input
        self.classes = calsses
        self.TRAINING = TRAINING
        self.create()

    def create(self):

        assert (self.X.shape == (self.X.shape[0], 224, 224, 3))

        # stage 1
        self.x = tf.layers.conv2d(self.X, filters=64, kernel_size=(7, 7), strides=(2, 2), padding='same',name='conv1')
        self.x = tf.layers.batch_normalization(self.x, axis=3, name='bn_conv1', training=self.TRAINING)
        self.x = tf.nn.relu(self.x)
        self.x = tf.layers.max_pooling2d(self.x, pool_size=(3, 3), strides=(2, 2),padding='same')
        print(self.x)