keras
shelleyHLX
这个作者很懒,什么都没留下…
展开
-
keras中inception_v3的权重和模型架构
inception_v3https://github.com/fchollet/deep-learning-models/releases/download/v0.5/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5import keras.layers as KLimport keras.models as KMimport kerasprint(keras.__version__)from keras.application原创 2020-12-24 13:20:52 · 507 阅读 · 0 评论 -
keras中MobileNet的模型架构
import keras.layers as KLimport keras.models as KMimport kerasfrom keras.applications.mobilenet import MobileNetimport os# os.environ['CUDA_DEVICES_VISIBLE'] = '1'print(keras.__version__)input_image = KL.Input(shape=[512, 512, 3], name="input_imag原创 2020-12-02 15:02:43 · 691 阅读 · 0 评论 -
keras中resnet50的模型架构
import keras.layers as KLimport keras.models as KMimport kerasimport osos.environ['CUDA_DEVICES_VISIBLE'] = '1'print(keras.__version__)from keras.applications.resnet50 import ResNet50input_image = KL.Input(shape=[512, 512, 3], name="input_image"原创 2020-12-02 14:27:58 · 424 阅读 · 0 评论 -
keras中resnet101的模型架构
import keras.layers as KLimport keras.models as KMimport kerasimport osos.environ['CUDA_DEVICES_VISIBLE'] = '1'print(keras.__version__)from keras.applications.resnet import ResNet101input_image = KL.Input(shape=[512, 512, 3], name="input_image..原创 2020-12-02 14:02:45 · 672 阅读 · 0 评论 -
keras中densenet201的权重和模型架构
https://github.com/keras-team/keras-applications/releases/download/densenet/densenet201_weights_tf_dim_ordering_tf_kernels.h5https://github.com/keras-team/keras-applications/releases/download/densenet/densenet201_weights_tf_dim_ordering_tf_kernels_notop.h原创 2020-12-02 13:35:39 · 1550 阅读 · 2 评论 -
keras中densenet169模型的权重和架构
import keras.layers as KLimport keras.models as KMimport kerasfrom keras.applications.densenet import DenseNet169import os# os.environ['CUDA_DEVICES_VISIBLE'] = '1'print(keras.__version__)input_image = KL.Input(shape=[512, 512, 3], name="input_im.原创 2020-12-01 16:24:36 · 903 阅读 · 0 评论 -
keras中densenet121模型的权重和架构
densenet121https://github.com/keras-team/keras-applications/releases/download/densenet/densenet121_weights_tf_dim_ordering_tf_kernels_notop.h5from keras.applications.densenet import DenseNet121input_image = KL.Input(shape=[512, 512, 3], name="input_ima原创 2020-12-01 16:20:16 · 1711 阅读 · 2 评论 -
keras中resnet50的权重和模型架构
resnet50https://github.com/fchollet/deep-learning-models/releases/download/v0.2/resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5from keras.applications.resnet50 import ResNet50input_image = KL.Input(shape=[512, 512, 3], name="input_image")resnet5原创 2020-12-01 16:05:56 · 1840 阅读 · 1 评论 -
keras中resnet101的权重和模型架构
resnet101https://github.com/keras-team/keras-applications/releases/download/resnet/resnet101_weights_tf_dim_ordering_tf_kernels.h5import tensorflow as tfimport numpy as np# import kerasimport keras.layers as KLimport keras.models as KMimport keras原创 2020-12-01 16:03:19 · 1180 阅读 · 0 评论 -
keras中resnet152的权重和模型架构
文章目录resnet152resnet152https://github.com/keras-team/keras-applications/releases/download/resnet/resnet152_weights_tf_dim_ordering_tf_kernels.h5import keras.layers as KLimport keras.models as KMimport kerasimport os# os.environ['CUDA_DEVICES_VISIBL原创 2020-12-01 16:08:39 · 1045 阅读 · 0 评论 -
【tf.keras】官方教程一 Keras overview
目录Sequential Model:(the simplest type of model)Getting started with the Keras Sequential modelSpecifying the input shapeCompilationTraining使用tf.keras建立一个简易的模型使用Numpy数组进行训练:使用tf.data 数据集进行训练验证和预测使用tf.keras构建复杂模型The functional API模型子.转载 2020-11-25 14:08:32 · 555 阅读 · 0 评论 -
深度学习题目
L1正则化和L2正则化可以看做是损失函数的惩罚项。所谓『惩罚』是指对损失函数中的某些参数做一些限制。L1正则化的模型建叫做Lasso回归,使用L2正则化的模型叫做Ridge回归(岭回归。但是使用正则化来防止过拟合的原理是什么?L1和L2正则化有什么区别呢?1 L1正则化与L2正则化L1正则化的表达如下,其中为L1正则化项,L1正则化是指权值向量w 中各个元素的绝对值之和。L2正则化项表达式如下,其中为L2正则化项,L2正则化是指权值向量w 中各个元素的平方和然后再求平方根。L1和L2...转载 2020-11-25 14:00:08 · 944 阅读 · 0 评论 -
keras中vgg16,vgg19的下载和基本模型的架构
文章目录vgg19inception_v3vgg16resnet50densenet121用迅雷下载,速度比较快一些。文本检测的ctpn需要把图片下采样16倍,将所有编码网络进行总结。权重位置:/home/hlx/.keras/models/keras的版本:vgg19https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg19_weights_tf_dim_ordering_tf_kernels_n原创 2020-11-17 13:12:08 · 1779 阅读 · 0 评论