How does DARTS compare to ENAS? Which one is better or what advantages does they each have?


  • $\begingroup$ quote from paper1: “The training takes 1.5 days on a single 1080Ti GPU with our PyTorch implementation.” That sounds reasonable fast. I like the idea to benchmark different algorithms on an objective basis. Perhaps it would make sense to implement DARTS on an FPGA chip? $\endgroup$ – Manuel Rodriguez Jan 12 '19 at 23:43
  • $\begingroup$ @ManuelRodriguez But the ENAS paper says that it only takes 0.5 days to train under the same setting. It seems to be faster than DARTS. $\endgroup$ – Shen Zhuoran Jan 14 '19 at 6:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.