Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Tensorflow in JavaScript #6

Open
huan opened this issue Mar 17, 2019 · 4 comments
Open

Tensorflow in JavaScript #6

huan opened this issue Mar 17, 2019 · 4 comments
Assignees
Labels

Comments

@huan
Copy link
Collaborator

huan commented Mar 17, 2019

基本上比较全面的介绍了 Node.js 和 Browser 里面如何使用 Tensorflow.js ,同时提供了一个完整的 Seq2Seq 模型供读者更深度的了解和上手。

我的开发子 Repo: https://github.com/huan/javascript-concise-chitchat

cc: @wangtz

@huan
Copy link
Collaborator Author

huan commented Aug 25, 2019

Reviews from @pyu10055:

Overall looks great, couple suggestions.

  1. can you add the WeChat plugin support in the overview
  2. the converter params have changed, please check 1.x doc
  3. please also check node doc
  4. the performance numbers are old I will send you a new one.

Thanks for the reviews, I'll looking into them and looking forward to receiving your new performance charts!

@huan
Copy link
Collaborator Author

huan commented Sep 9, 2019

Reviews from @snowkylin

JavaScript

  • Use code reference in RST file instead of hardcoded. (submodule for huan's repo?)
  • Code example segment explanation
  • Replace the full example code with a link to repo/web
  • Make the chapter shorter by moving minor contents into tips box
  • Important to tell reader: J's can use to train model: linear regression
  • Compare lite(in edge) and serving(in server) with js(in browser)
  • JavaScript ES6 Syntax Introduction
    • Fat arrow explanation.
    • import explanation.
  • JavaScript in Browser Introduction
    • How to use camera
    • how to animate video
    • how to capture image,
  • Mini progromming move to 推荐阅读
  • Add a camera demo: interactive on phone.
  • Add source link to each image
  • The meaning of the JS performance numbers is milliseconds, not 100% percentage.

snowkylin added a commit that referenced this issue Nov 11, 2019
Javascript chapter improvements (#6)
@snowkylin
Copy link
Owner

训练一节

  1. 介绍一下链式表达式,这一块和Python区别很大,是必须要用链式表达式吗,还是说一般的表达式也可以用(比如tf.add(a, b))?
  2. 关于取值,介绍一下.dataSync()和.arraySync(),还有其他类似的.XXXSync的API吗,看起来是对应于传统的sess.run()或者eager execution的.numpy(),这一块可以系统介绍一下。

@huan
Copy link
Collaborator Author

huan commented Nov 12, 2019

  1. xsRaw.sub(...).div(...) 可以被改写成
    1. tf.sub(xsRaw, ...) https://js.tensorflow.org/api/latest/#sub
    2. tf.div(a, b) https://js.tensorflow.org/api/latest/#div
  2. 有三个: dataSync / arraySync / bufferSync 。他们的作用就是把数据从 GPU 里面取回来,和 session.run / .numpy() 的确是一类功能。
    1. A good article: The perils of tensor.dataSync()

我这两个都加一下。

huan added a commit that referenced this issue Nov 12, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

2 participants