Shuai Kyle Zheng

@bittnt

Research Scientist. Previously: eBay. U. Oxford, NLPR/CAS, BIT alumns. Interests in: computer vision and machine learning. All opinions are my own.

California, USA
Vrijeme pridruživanja: svibanj 2012.

Tweetovi

Blokirali ste korisnika/cu @bittnt

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @bittnt

  1. 29. sij

    Can all the open source frameworks tensorflow and pytorch stop using deprecationwarning hidden in the code? This practice does not help people to upgrades but force people to spend extra work to remove those warnings. There are must be better way to help users to upgrade.

    Poništi
  2. 19. sij

    Check out the 2nd 2020 workshop & challenge on learning from imperfect data (LID). Winners on weakly semantic segmentation & object detection who also provide opensource inference code & models will receive cash reward thanks to Baidu.

    Poništi
  3. 22. pro 2019.
    Poništi
  4. 7. stu 2019.

    How come arXiv website is down.... I just realize this is super important

    Poništi
  5. proslijedio/la je Tweet
    12. kol 2019.

    AAAI'20 () is seeking (self-)nominations of qualified individuals to serve as program committee (PC) members. With this call for help, we hope to expand and diversify the PC. See for details and to nominate

    Poništi
  6. 16. lip 2019.

    Rogerio Feris from IBM and MIT-IBM lab is presenting work on learning more from less at LID challenge

    Poništi
  7. 16. lip 2019.

    Yanping Huang from Google Brain presents the work GPipe a flexible library that enables large scale neural networks training. LID Challenge and workshop at

    Poništi
  8. 16. lip 2019.

    Professor Bolei Zhou is presenting Objects Disentangled from Classifying and Synthesizing Scenes at LID workshop – mjesto: Hyatt Regency Long Beach

    Poništi
  9. 16. lip 2019.

    Ming-Hsuan Yang from UC Merced & Google AI research is presenting the work in learning to segment objects in various condition. LID workshop at .

    Poništi
  10. proslijedio/la je Tweet

    James () is presenting their work on learning object structure with no supervision. Check out his talk at LID challenge for also a sneak peek of Slim Densepose which boosts annotations using video. 08:30-09:10, Hyatt Regency B

    Poništi
  11. proslijedio/la je Tweet

    Anurag () from is presenting joint work on "Learning from Weak Supervision: Panoptic Segmentation and 3D Human Pose Estimation" at LID challenge today (11:30-12:00, at Hyatt Regency B).

    Poništi
  12. 15. lip 2019.

    We also have 9 awesome accepted papers focusing on weakly supervised learning or related topics. Check out

    Prikaži ovu nit
    Poništi
  13. 15. lip 2019.

    Rogerio Feris (Learning More from Less: Weak Supervision and Beyond), (Enhancing U-Nets to Segment Ambiguous Images), following with the winners of two tasks: Object semantic segmentation with image-level supervision; Scene parsing with point-level supervision.

    Prikaži ovu nit
    Poništi
  14. 15. lip 2019.

    Then will have speakers Bolei Zhou (Objects Disentangled from Classifying and Synthesizing Scenes ), Yanping Huang (GPipe: Easy Scaling with Micro-Batch Pipeline Parallelism), Anurag Arnab (Learning from Weak Supervision: Panoptic Segmentation and 3D Human Pose Estimation)

    Prikaži ovu nit
    Poništi
  15. 15. lip 2019.

    Join the & workshop () Sunday (June 16, LA Time) at Hyatt Regency B Long Beach. We will kick off w learning from motion w/o supervision, Ming-Hsuan: Unseen Object Segmentation in Videos via Transferable Representations.

    Prikaži ovu nit
    Poništi
  16. 27. ožu 2019.

    Finally made the and Tensorflow work both with Jetson Nano, thanks to Dustin's work. The main block turns out to be the swapfile size is too small, which is to easy to address.

    Poništi
  17. 24. ožu 2019.

    Finally, The Nvidia Jetson Nano works with Pytorch (1.1.0a0+b457266) in Python 2.7 after it works with Tensorflow 1.13 in Python 3.6.

    Poništi
  18. 24. ožu 2019.

    Learning from imperfect data and annotation workshop will be held at Long Beach this year: welcome to contribute and participate ! We also hope to find out the SOTA weakly supervised methods for semantic segmentation from two challenges.

    Poništi
  19. 22. ožu 2019.

    It is awesome that everyone talks about the support of Tensorflow/pytorch for the ... well, if you went ahead and found that there is only a working official version of 1.13 (only for python 3.6). Anyone has good luck in getting pytorch working?

    Poništi
  20. 26. velj 2019.

    Feel awesome, running tensorflow code using Nvidia RTX 2070 in a new tensorbook after just start the Ubuntu. This should be in all laptop in the future.

    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·