Auto ML is such an unfair advantage.. Seems like we should be focusing on generic open source auto ml implementations with this rate of progress. How many GPU hours does an efficient version of AutoML approximately take ? Is this something we can develop and train on Colab ?
Introducing EfficientNet-EdgeTPU: customized for mobile accelerators, with higher accuracy and 10x faster inference speed. blog post: https://ai.googleblog.com/2019/08/efficientnet-edgetpu-creating.html … Code and pertained models: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet/edgetpu … https://twitter.com/GoogleAI/status/1158804847488978944 …pic.twitter.com/Vbj6aRHQMi
-
-
-
Hi Karanbir, AutoML has evolved: recent algorithms (such as DARTS https://arxiv.org/abs/1806.09055 ) can finish a search in a couple of GPU hours.
- Još 2 druga odgovora
Novi razgovor -
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.