TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Trivial Artificial Neural Network in Assembly Language

51 点作者 fogus大约 13 年前

3 条评论

dave_sullivan大约 13 年前
That's really cool, a good breakdown on what's actually going on with an ann. But I don't think using assembly will help with speed, wouldn't you still be better off using matrix multiplication on a gpu, written with something like theano in python? Then again, maybe you just used assembly to explain things at a very low level, rather than for any speed boost. Either way, very cool article.
评论 #3667267 未加载
评论 #3667168 未加载
kruhft大约 13 年前
I don't think I had more fun and felt more control than when I was professionally programming in assembly language. The more I use high level languages the more I just want to step back and learn x86-64 assembler and put scoping and default allocation and language lawyers behind me. Of course there's nothing stopping me except myself really...
xcallmejudasx大约 13 年前
How long did it take you to put this together and how much of that time was coding vs researching?