You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
TensorLayer is a deep learning and reinforcement learning library based on [TensorFlow](https://www.tensorflow.org). It provides rich data processing, model training and serving modules to help both **researchers** and **engineers** build practical machine learning workflows.
21
+
TensorLayer is a deep learning and reinforcement learning library on top of [TensorFlow](https://www.tensorflow.org). It provides rich neural layers and utility functions to help researchers and engineers build real-world AI applications. TensorLayer is awarded the 2017 Best Open Source Software Award by the prestigious [ACM Multimedia Society](http://www.acmmm.org/2017/mm-2017-awardees/).
*[18 Jan][《深度学习：一起玩转TensorLayer》](http://www.broadview.com.cn/book/5059) (Deep Learning using TensorLayer)
27
-
*[17 Dec]Officially support **Distributed Training** contributed by [TensorPort](https://tensorport.com), see[tiny example](https://github.com/zsdonghao/tensorlayer/blob/master/example/tutorial_mnist_distributed.py).
28
-
*[17 Nov] Release many APIs for data augmentation for object detection, see [tl.prepro](http://tensorlayer.readthedocs.io/en/latest/modules/prepro.html#object-detection).
27
+
*[17 Dec]Experimentally release APIs for distributed training (by [TensorPort](https://tensorport.com)). See[tiny example](https://github.com/zsdonghao/tensorlayer/blob/master/example/tutorial_mnist_distributed.py).
28
+
*[17 Nov] Release data augmentation APIs for object detection, see [tl.prepro](http://tensorlayer.readthedocs.io/en/latest/modules/prepro.html#object-detection).
29
29
*[17 Nov] Support [Convolutional LSTM](https://arxiv.org/abs/1506.04214), see [ConvLSTMLayer](http://tensorlayer.readthedocs.io/en/latest/modules/layers.html#conv-lstm-layer).
30
30
*[17 Nov] Support [Deformable Convolution](https://arxiv.org/abs/1703.06211), see [DeformableConv2dLayer](http://tensorlayer.readthedocs.io/en/latest/modules/layers.html#d-deformable-conv).
31
-
*[17 Nov] Download [VOC dataset](http://host.robots.ox.ac.uk/pascal/VOC/voc2012/) in one line, see [load\_voc_dataset](http://tensorlayer.readthedocs.io/en/latest/modules/files.html#voc-2007-2012).
32
-
*[17 Oct] We won the **Best Open Source Software Award**[@ACM MM](http://www.acmmm.org/2017/mm-2017-awardees/) 2017, here is the slide for the presentation [[click]](https://github.com/shorxp/tensorlayer-chinese/blob/master/docs/TensorLayer%20ACM%20MM.pdf).
33
31
*[17 Sep] New example [Chatbot in 200 lines of code](https://github.com/zsdonghao/seq2seq-chatbot) for [Seq2Seq](http://tensorlayer.readthedocs.io/en/latest/modules/layers.html#simple-seq2seq).
32
+
<!--
33
+
* [17 Nov] Download [VOC dataset](http://host.robots.ox.ac.uk/pascal/VOC/voc2012/) in one line, see [load\_voc_dataset](http://tensorlayer.readthedocs.io/en/latest/modules/files.html#voc-2007-2012).
34
34
* [17 Sep] Release [ROI layer](http://tensorlayer.readthedocs.io/en/latest/modules/layers.html#roi-layer) for Object Detection.
35
35
* [17 Jun] Release [SpatialTransformer2dAffineLayer](http://tensorlayer.readthedocs.io/en/latest/modules/layers.html#spatial-transformer) for [Spatial Transformer Networks](https://github.com/zsdonghao/Spatial-Transformer-Nets) see [example code](https://github.com/zsdonghao/Spatial-Transformer-Nets).
36
36
* [17 Jun] Release [Sub-pixel Convolution 2D](http://tensorlayer.readthedocs.io/en/latest/modules/layers.html#super-resolution-layer) for Super-resolution see [SRGAN code](https://github.com/zsdonghao/SRGAN).
37
37
* [17 May] You can now use TensorLayer with [TFSlim](https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_inceptionV3_tfslim.py) and [Keras](https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_keras.py) together!
38
+
-->
38
39
39
40
40
41
41
42
# Installation
42
43
43
-
TensorLayer has install prerequisites including TensorFlow, numpy, matplotlib and nltk(optional). For GPU support, CUDA and cuDNN are required. Please check [documentation](http://tensorlayer.readthedocs.io/en/latest/user/installation.html) for detailed instructions. The simplest way to install TensorLayer in your python program is:
44
+
TensorLayer has package pre-requisites including TensorFlow, numpy, matplotlib and nltk(optional). For GPU support, CUDA and cuDNN are required. Please check [documentation](http://tensorlayer.readthedocs.io/en/latest/user/installation.html) for detailed instructions. The simplest way to install TensorLayer in your python program is:
Examples can be found [in this folder](https://github.com/zsdonghao/tensorlayer/tree/master/example) and [Github topic](https://github.com/search?q=topic%3Atensorlayer&type=Repositories).
55
56
@@ -101,7 +102,7 @@ Examples can be found [in this folder](https://github.com/zsdonghao/tensorlayer/
101
102
-[DAGGER](https://www.cs.cmu.edu/%7Esross1/publications/Ross-AIStats11-NoRegret.pdf) for ([Gym Torcs](https://github.com/ugo-nama-kun/gym_torcs)) by [zsdonghao](https://github.com/zsdonghao/Imitation-Learning-Dagger-Torcs).
102
103
-[TRPO](https://arxiv.org/abs/1502.05477) for continuous and discrete action space by [jjkke88](https://github.com/jjkke88/RL_toolbox).
103
104
104
-
## Special Examples
105
+
## Miscellaneous
105
106
- Distributed Training. [mnist](https://github.com/zsdonghao/tensorlayer/blob/master/example/tutorial_mnist_distributed.py) and [imagenet](https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_imagenet_inceptionV3_distributed.py) by [jorgemf](https://github.com/jorgemf).
106
107
- Merge TF-Slim into TensorLayer. [tutorial\_inceptionV3_tfslim.py](https://github.com/zsdonghao/tensorlayer/blob/master/example/tutorial_inceptionV3_tfslim.py).
107
108
- Merge Keras into TensorLayer. [tutorial_keras.py](https://github.com/zsdonghao/tensorlayer/blob/master/example/tutorial_keras.py).
@@ -112,47 +113,42 @@ Examples can be found [in this folder](https://github.com/zsdonghao/tensorlayer/
112
113
- Float 16 half-precision model, see [tutorial\_mnist_float16.py](https://github.com/zsdonghao/tensorlayer/blob/master/example/tutorial_mnist_float16.py)
113
114
114
115
## Notes
115
-
* TensorLayer provides two set of Convolutional layer APIs, see [(Professional)](http://tensorlayer.readthedocs.io/en/latest/modules/layers.html#convolutional-layer-pro) and [(Simplified)](http://tensorlayer.readthedocs.io/en/latest/modules/layers.html#convolutional-layer-simplified) on readthedocs website.
116
+
TensorLayer provides two set of Convolutional layer APIs, see [(Professional)](http://tensorlayer.readthedocs.io/en/latest/modules/layers.html#convolutional-layer-pro) and [(Simplified)](http://tensorlayer.readthedocs.io/en/latest/modules/layers.html#convolutional-layer-simplified) on readthedocs website.
117
+
<!--
116
118
* If you get into trouble, you can start a discussion on [Slack](https://join.slack.com/t/tensorlayer/shared_invite/MjI1NjQ5NTUxOTY5LTE1MDI3MDYwNTItYzYwNmFiZmZkOA), [Gitter](https://gitter.im/tensorlayer/Lobby#?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge>),
[QQ group](https://github.com/zsdonghao/tensorlayer/blob/master/img/img_qq.png) and [Wechat group](https://github.com/shorxp/tensorlayer-chinese/blob/master/docs/wechat_group.md).
119
-
120
+
[QQ group](https://github.com/zsdonghao/tensorlayer/blob/master/img/img_qq.png) and [Wechat group](https://github.com/shorxp/tensorlayer-chinese/blob/master/docs/wechat_group.md).
121
+
-->
120
122
121
-
# Design Principles
122
123
123
-
As deep learning practitioners, we have been looking for a library that can serve for various development phases. This library shall be easy for beginners by providing rich neural network reference implementations. Later, it can be extended to address **real-world problems** by controlling training backends to exhibit low-level cognitive behaviours. In the end, it shall be able to serve in challenging **production environments**.
124
+
# Library Features
124
125
125
-
TensorLayer is designed for beginning, intermediate and professional deep learning users with following goals:
126
+
As deep learning practitioners, we have been looking for a library that can serve for various development phases. This library shall be easy for beginners by providing rich neural network reference implementations. Later, it can be extended to address *real-world problems* by controlling training backends to exhibit low-level cognitive behaviours. In the end, it shall be able to serve in challenging *production environments*. TensorLayer is developed with the following goals:
126
127
127
128
-*Simplicity* : TensorLayer lifts the low-level dataflow abstraction of TensorFlow to **high-level** deep learning modules.
128
129
A user often find it easy to bootstrap with TensorLayer, and then dive into low-level implementation only if need.
129
130
-*Transparency* : TensorLayer provides access to the **native APIs** of TensorFlow. This helps users achieve flexible controls within the training engine.
130
-
-*Composability* : If possible, deep learning modules are composed, not built. TensorLayer can glue existing pieces together (e.g., **connected with**[TF-Slim](http://tensorlayer.readthedocs.io/en/latest/modules/layers.html#connect-tf-slim) and [Keras](https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_keras.py)).
131
131
-*Performance* : TensorLayer provides **zero-cost** abstraction (see Benchmark below). It can run on distributed and heterogeneous TensorFlow platforms with full power.
132
132
133
-
#Benchmark
133
+
## Low Runtime Overhead
134
134
135
-
A common concern towards TensorLayer
136
-
is performance overhead. We investigate this by running classic
135
+
TensorLayer has negligible overhead. We show this by benchmarking classic deep learning
137
136
models using TensorLayer and native TensorFlow implementations
138
-
on a Titan X Pascal GPU. The following are the training throughputs of respective tasks:
137
+
on a Titan X Pascal GPU. Here are the training speeds of respective tasks:
A frequent question regarding TensorLayer is what is the different with other libraries like Keras, TFSlim and Tflearn.
146
+
A frequent question regarding TensorLayer is what is the different with other libraries like Keras and Tflearn.
148
147
These libraries are comfortable to start with. They provide imperative abstractions to lower adoption barrier;
149
148
but in turn mask the underlying engine from users. Though good for bootstrap,
150
-
it becomes hard to tune and modify from the bottom, which is quite necessary in tackling many real-world problems.
151
-
152
-
Without compromise in simplicity, TensorLayer advocates a more flexible and composable paradigm:
153
-
neural network libraries shall be used interchangeably with the native engine.
154
-
This allows users to enjoy the ease of pre-built modules without losing visibility to the deep.
155
-
This non-intrusive nature also makes it viable to consolidate with other TF's wrappers.
149
+
it becomes hard to leverage the advanced features of TensorFlow, which is quite necessary in tackling real-world problems.
150
+
Without compromise in simplicity, TensorLayer APIs are generally more flexible and transparent.
151
+
Users often find it easy to start with the examples and tutorials of TensorLayer, then dive into the TensorFlow low-level APIs only if need through the transparent API design.
156
152
157
153
158
154
# Documentation
@@ -231,29 +227,24 @@ We provide many helper functions (like `fit()` , `test()`) that is similar to Ke
231
227
-->
232
228
233
229
234
-
# Community
230
+
# Academic and Industry Users
235
231
236
-
TensorLayer is maintained by numerous Github contributors [here](https://github.com/zsdonghao/tensorlayer/graphs/contributors).
237
-
This project is in an active development stage and has received numerous contributions from an open community.
232
+
TensorLayer is in an active development stage and has received numerous contributions from an open community.
238
233
It has been widely used by researchers from Imperial College London, Carnegie Mellon University, Stanford University,
239
234
Tsinghua University, UCLA, Linköping University and etc.,
240
235
as well as engineers from Google, Microsoft, Alibaba, Tencent, Penguins Innovate, ReFULE4, Bloomberg, GoodAILab and many others.
241
236
242
-
<!--
243
-
TensorLayer started as an internal repository at Imperial College London, helping researchers to test their new methods. It now encourages researchers from all over the world to controbute their methods so as to promote the development of machine learning. You can either contact us to discuss your ideas, or fork our repository and make a pull request.
0 commit comments