- 1、本文档共45页,可阅读全部内容。
- 2、有哪些信誉好的足球投注网站(book118)网站文档一经付费(服务费),不意味着购买了该文档的版权,仅供个人/单位学习、研究之用,不得用于商业用途,未经授权,严禁复制、发行、汇编、翻译或者网络传播等,侵权必究。
- 3、本站所有内容均由合作方或网友上传,本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺!文档内容仅供研究参考,付费前请自行鉴别。如您付费,意味着您自己接受本站规则且自行承担风险,本站不退款、不进行额外附加服务;查看《如何避免下载的几个坑》。如果您已付费下载过本站文档,您可以点击 这里二次下载。
- 4、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等,请点击“版权申诉”(推荐),也可以打举报电话:400-050-0827(电话支持时间:9:00-18:30)。
查看更多
* * * * 这些技术已经融合到了caffe中 * * * * * * * * * * * * * * 2.Theano 1.Overview: A Python library that allows to define, optimize and evaluate mathematical expression. From Yoshua Bengio’s group at University of Montreal. Embracing computation graphs, symbolic computation. High-level wrappers: Keras, Lasagne. 2.Github: /Theano/Theano 3.Tutorial: /tutorial/contents.html Pros and Cons of Theano 3.TensorFlow 1.Overview: Very similar to Theano - all about computation graphs. Easy visualizations (TensorBoard). Multi-GPU and multi-node training. 2.Tutorial: http://terryum.io/ml_practice/2016/05/28/TFIntroSlides/ Load data Define the NN structure Set optimization parameters Run! Basic Flow of TensorFlow 1.Load data 1.Load data 2. Define the NN structure 3.Set optimization parameters 4. RUN The Pros and Cons of TensorFlow Overview Caffe Theano TensorFlow language C++,Python,MATLAB Python Python Pretrained Yes++ Yes(Lasagne) Inception Multi-GPU: Data parallel Yes Yes Yes Multi-GPU: Model parallel No Experimental Yes(Best) Speed Very fast Quick Quick Platform All operation systems Linux, OSX Linux, OSX Readable source code Yes No No Good at RNN No Yes Yes(Best) Feature extraction / finetuning existing models: Use Caffe Complex uses of pretrained models: Use Lasagne(Theano) Crazy RNNs: Use Theano or Tensorflow Huge model, need model parallelism: Use TensorFlow Other popular deep learning tools 1.Matconvnet: From VGG: /matconvnet/ 2.Torch7: /torch/torch7 3. Mxnet http://mxnet.readthedocs.io/en/latest/ Reference /qiexingqieying/article/details/zer0n/deepframeworks/blob/master/README.md http://lchiffon.github.io/2015/11/16/long.html 深度学习:21天实战caffe /u011762313/article/category/5705779 * * * * 目前常用的有caffe,matconvnet(VGG),theano(Toronto),tensorflow(google),mxnet(DistributeMachine Learning Community),torch(facebook) * 重点讲caffe,caffe的全称是. Jia Yangqing 是清华毕业的,在Ucberkely 获得博士学位,师从Trevor Darrel。贾大牛跟倪老师也很熟,他之前去倪老师在新加坡的研究所待过 * * * * 这个就类似是一个搭积木的方式,每一层我要定义好输入与输
文档评论(0)