Visual Interpretability for Convolutional Neural Networks by Himanshu Rawlani medium.com/p/visual-inter…

deepai.org/publication/in…

Building Tensorflow-GPU from source for RTX 2080 by Saiteja Dommeti link.medium.com/aH3rrDD2NW

Benchmarking Edge Computing by Alasdair Allan link.medium.com/IcevY6wFRW

Say Hello to HACKcelerator 2019 👋 blog.angelhack.com/say-hello-to-h…

A Comprehensive Introduction to Different Types of Convolutions in Deep Learning by Kunlun Bai in Towards Data Science towardsdatascience.com/a-comprehensiv…

Coral is Google’s quiet initiative to enable AI without the cloud theverge.com/2020/1/14/2106… 来自 The Verge

The Next Generation of Machine Learning Tools inoryy.com/post/next-gen-…

Check this out! engineering.fb.com/ml-application… 来自 fbOpenSource

cv-tricks.com/object-detecti…

Object detection: speed and accuracy comparison (Faster R-CNN, R-FCN, SSD, FPN, RetinaNet and… by Jonathan Hui link.medium.com/oJNDw8Xq03

Possible link between Magecart group & Cerberus Banking Trojan C2 by Bank Security link.medium.com/3WKdOOMdu5

tiktok.com/@dixiedamelio

Perplexity AI: 1+1 =? perplexity.ai/?s=t&uuid=8e82…

"LLM as a Storage Tool Since LLMs can accomplish the compression of textual data from the entire internet,Training is the compression of data, inference is the decompression of data.has the LLM already surpassed the fundamental principles of information theory?" puttry amellya

LLM作为个人数据存储工具 LLM 既然可以完成对整个互联网的文本数据的压缩 那么就可以对一个人每天产生的 数据进行压缩，变成一个存储的介质 每天产生的新的数据，进入LLM 重新训练 训练是数据压缩 推理是数据解压缩 这样推理的话，LLM 是否已经突破了信息论的基础；

关于 > 世界经济论坛年会 cn.weforum.org/events/world-e… 来自 World Economic Forum

Check out what I built on Replit! replit.com/@zyxcambridge/…

Output is the real learning. Write a tweet and a content post to teach friends to do a small project and apply it to the customer application once. The value of output is threefold: to consolidate understanding (you must understand it before you can explain it), to get feedback