You might like
Pinned
Critical batch size is crucial for reducing the wall-clock time of large-scale training runs with data parallelism. We find that it depends primarily on data size. 🧵 [1/n] Paper 📑: arxiv.org/abs/2410.21676 Blog 📝: kempnerinstitute.harvard.edu/research/deepe…
Hanlin Zhang reposted
Do language models have algorithmic creativity? To find out, we built AlgoTune, a benchmark challenging agents to optimize 100+ algorithms like gzip compression, AES encryption and PCA. Frontier models struggle, finding only surface-level wins. Lots of headroom here!🧵⬇️
United States Trends
- 1. Good Monday 27.4K posts
- 2. #MondayMotivation 31.5K posts
- 3. Rudy Giuliani 16.2K posts
- 4. Happy Birthday Marines 3,549 posts
- 5. The BBC 432K posts
- 6. 8 Democrats 11.3K posts
- 7. Steelers 54.2K posts
- 8. #ITZY_TUNNELVISION 38K posts
- 9. Semper Fi 3,991 posts
- 10. Resign 118K posts
- 11. Mr. 4 4,718 posts
- 12. Chargers 39.2K posts
- 13. Tim Kaine 25.3K posts
- 14. #Talus_Labs N/A
- 15. #USMC N/A
- 16. Tomlin 8,507 posts
- 17. Durbin 32.9K posts
- 18. 8 Dems 8,428 posts
- 19. Rodgers 21.9K posts
- 20. Angus King 20.5K posts
You might like
-
Ziniu Hu
@acbuller -
Dinghuai Zhang 张鼎怀
@zdhnarsil -
Yiping Lu
@2prime_PKU -
Jason Dou
@JasonDou9 -
Ruihan Yang
@RchalYang -
Meng Liu
@mengliu_1998 -
Tahoe Therapeutics
@tahoe_ai -
Oxford Protein Informatics Group (OPIG)
@OPIGlets -
Wangchunshu Zhou
@wangchunshu -
Enveda
@lifeschemistry -
Bonnie Berger Lab
@lab_berger -
Xiangchen Song
@XiangchenSong -
Zhiting Hu
@ZhitingHu -
Hongyi Wang
@HongyiWang10 -
Xiuyu Li
@xiuyu_l
Loading...
Something went wrong.
Something went wrong.