dataastron的博客

1_绘图效果较为简洁image_input = Input(shape=(224, 224, 3))vision_model1 = ResNet50(include_top=False, weights='imagenet',input_tensor=Input(shape=(224, 224, 3)))encoded_image1 = vision_model1(image_input)x = GlobalAveragePooling2D(name='avg_pool')(encoded_imag

先查版本对症下药import tensorflow as tfprint(tf.__version__)tf1.x版本from keras.backend.tensorflow_backend import set_sessionconfig = tf.ConfigProto()config.gpu_options.per_process_gpu_memory_fraction = ...

实验介绍本文参考keras中文官方文档操作。传送门Keras框架搭建Keras中mnist数据集测试3行代码这么方便>>> git clone https://github.com/fchollet/keras.git>>> cd keras/examples/>>> python mnist_mlp.py执行日志root@astron:/asdata# git clone https://gi

自定义学习率代码可以加快训练的速度，下面是代码的内容和学习率的截图。import matplotlib.pyplot as pltdef lrfn(epoch): LR_START = 0.00001 LR_MAX = 0.00005 *8#* strategy.num_replicas_in_sync LR_MIN = 0.00001 LR_RAMPUP_EP...

import efficientnet.tfkeras