chyelang / hw2_image_classification
利用TensorFlow(TF),本作业实现了一个基于full convolution stack、inception v2 module等模块的图片分类网络,纵向共包含10层包含参数的层,采用自己搭建的卷积层。对于dset1与dset2采用同样的网络结构,分别进行训练。最终在dset1验证集(约含900张图,下同)上的Top1分类准确率约为0.52,在dset2验证集上的Top1准确率约为0.63。在K80显卡的单核上,该模型的训练速度约为205张图每秒,最终所得模型的checkpoint约为45MB。
☆12Updated 6 years ago
Alternatives and similar repositories for hw2_image_classification:
Users that are interested in hw2_image_classification are comparing it to the libraries listed below
- some small codes about deep learning☆51Updated 6 years ago
- 一个面向初学者的,友好的Keras入门教程☆123Updated 5 years ago
- keras融合inception,vgg,residual_net做一个超快的迁移学习模型☆11Updated 6 years ago
- The code on deep learning.☆74Updated 6 years ago
- Models built with TensorFlow☆34Updated 7 years ago
- 深度学习入门的一些简单例子☆355Updated 6 years ago
- ☆47Updated 7 years ago
- some real example of machine learn algorithm☆65Updated 3 years ago
- mnist classification with tensorflow ( nn, cnn, lstm, nlstm, bi-lstm, cnn-rnn)☆106Updated 7 years ago
- use TensorFlow and Two stream CNN train and test the CK+ dataset☆50Updated 7 years ago
- TensorFlow学习☆122Updated 7 years ago
- data augmentation on python☆35Updated 7 years ago
- TensorFlow的练习代码☆44Updated 6 years ago
- TensorFlow实现Kaggle猫狗大战☆57Updated 7 years ago
- a baseline for baidu dog classification competition.☆113Updated 6 years ago
- U-Net图像分割练习题两则☆126Updated 6 years ago
- 唐源黄文坚所著书籍《tensorflow实战》的源码实现☆27Updated 7 years ago
- tensorflow_alexnet_classify (details http://www.cnblogs.com/vipyoumay/p/7686230.html)☆106Updated 6 years ago