I have just got one 3080 GPU for about 885 dollars. I will test it for its deep learning performance.

3080 for the server! Awesome!

Currently, Cuda 11.0 does not fully support 3080 or 3090 GPUs. Therefore, if you want to use 3080 or 3090 for deep learning, you can use sm_80 and wait for the support to sm_86.

The deep learning perfomance test for 3080 GPU.

Awesome 3090!

This cooler design for 3080 GPUs and 3090 GPUs is powerful and quiet! The highest temperature of them is always below 75 ℃!

After seeing my paper is revised by my tutor word for word again, I feel so grateful. What makes me more excited is there is not a sentence deleted wholly. Hahaha

Welcome to our latest work "DomainMix: Learning Generalizable Person Re-Identification Without Human Annotations":arxiv.org/abs/2011.11953

The codes and models for the current state-of-the-art unsupervised domain adaptation for person re-identification are available at github.com/WangWenhao0716…. The related paper is updated at arxiv.org/abs/2006.06525.

You can now buy a Tesla with Bitcoin

Bag of Tricks and A Strong baseline for Image Copy Detection (3rd) Wenhao Wang, Weipu Zhang, Yifan Sun, Yi Yang tl;dr: BarlowTwins pretraining -> GeM. Descriptor stretching - un-normalizing the descriptor norm based on K nearest neighbors, black magic. arxiv.org/abs/2111.08004…

We are happy to announce that in the Facebook AI Image Similarity Challenge, we rank first in Matching Track and third in Descriptor Track. You can find the solutions from arxiv.org/abs/2111.07090 and arxiv.org/abs/2111.08004. The links to code are available in the papers.

Will the ECCV result be released after one and a half hours?

Get ready! The countdown has started 🔥 The wait is almost over as we gear up to launch our new Twitter account! Follow us for exclusive updates on the SmartMint platform and so much more 🚀 #SmartMint #LaunchingSoon #StayTuned