TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Deep Learning in a Single File for Smart Devices

40 pointsby sungeunsover 9 years ago

3 comments

IanCalover 9 years ago
This is a nice overview, but needs some editing as I guess this is not the author&#x27;s first language (but still written better than I could in my second language). The meat of it is there, but I think it needs a once-over from either a native speaker or someone more fluent.<p>I think there&#x27;s some content still to be added here:<p>&gt; Then we remove some layers. then further remove somethings, but now the results are too pool. more explains, and the results table.<p>&gt; After compiled library for smart platform, the last thing is call C-API in the target language (Jave&#x2F;Swift).<p>&quot;After the library is compiled for the smart platform, the last thing to do is call the C API from the target language (Java&#x2F;Swift)&quot;.<p>On the actual topic, I really like the ability now to take powerful recognition frameworks and put them onto &quot;small&quot; devices. Great work.
评论 #10555408 未加载
joshvmover 9 years ago
I saw a demo of a deep convnet (I assume) running on a tablet at a conference recently. While it&#x27;s limited to the vocabulary that the net is trained on, it&#x27;s seriously impressive seeing this stuff work in realtime.<p>One of the other speakers was giving a tech demo using his webcam. He was looking around for a mug to demonstrate that the classification was good and could work quickly. In the meantime the camera was looking behind him on stage and correctly classified the image as &quot;theatre curtains&quot;. It was particularly cool because image processing results are often cherrypicked to show optimal performance and you learn to be skeptical.
评论 #10553052 未加载
Aardwolfover 9 years ago
Link to the actual 30k line C++ file: <a href="https:&#x2F;&#x2F;raw.githubusercontent.com&#x2F;jdeng&#x2F;gomxnet&#x2F;master&#x2F;mxnet.cc" rel="nofollow">https:&#x2F;&#x2F;raw.githubusercontent.com&#x2F;jdeng&#x2F;gomxnet&#x2F;master&#x2F;mxnet...</a><p>Compiles with &quot;clang++ -lblas mxnet.cc -std=c++11&quot;, so I&#x27;m not disappointed :)