Top/lib/Chainer
English | Japanese
English | Japanese

Menu

  • Top
  • Search 検索
  • Writers 著者リスト
  • Online forum フォーラム
↑

Recent

2021-08-17
  • os/Linux
2019-01-17
  • eq/LulzBot
2018-12-10
  • dev/Motoman
2018-07-05
  • tool/Git
2018-05-02
  • lib/Eigen
  • writer/Ilya_Ardakani
2018-04-26
  • lib/Deformable_Simulators
  • Top
2018-04-23
  • writer/Naoya_Chiba
  • MenuBar
  • writer
2018-04-22
  • lang/Python
2018-04-21
  • lang/C++
2018-04-19
  • Editor(editor)/Assignment
  • Editor(editor)
  • lang/C++/Exercise
2018-04-18
  • SandBox2
  • lang/Python/Exercise
  • soft/FreeCAD
  • lib/Chainer
Access: 1/644 - Editor / Admin

Chainer

Akihiko Yamaguchi

Chainer (http://chainer.org/) is an easy-to-use implementation of neural networks, implemented in Python. If you want to start a deep learning thing, Chainer is a good starting point.

Here is an example of learning forward dynamics of 7-DoF (degrees of freedom) robot arm:

Deep learning for forward kinematics of 7-DoF arm (testing)
Learning forward kinematics (joint angles q1,...,q7 to end-effector pose x,y,z,qx,qy,qz,qw), with deep neural networks. Green box denotes an analytically computed pose (i.e. target), and purple box denotes an estimated pose with NNs.

Setup:
NNs:  6 hidden layers, 200 units for each.
Activation function: ReLU.
Loss (error/cost/objective) function: mean square error.
# of samples: 20k and 100k.
# of mini batches is 20.
# of training epochs: 100 and 200.
Hack: normalization of quaternion (qx,qy,qz,qw) is needed
when doing estimation.
Implementation: Chainer ( http://chainer.org/ ).



Last-modified:2018-04-18 (Wed) 05:01:31 (1800d)
Link: Top(1791d) writer/Akihiko_Yamaguchi(1800d)
Site admin: Akihiko Yamaguchi.
Written by: Akihiko Yamaguchi.
System: PukiWiki 1.5.0. PHP 5.2.17. HTML conversion time: 0.439 sec.