Any chance you will release (part?) of the code? RAPID would be really cool to try out!
Extremely balanced external blog post about our Rubik’s cube results:https://www.alexirpan.com/2019/10/29/openai-rubiks.html …
-
-
-
Have been on the look out for rapid ever since they released their “learning dexterity” paper But after interning at a startup(not to mention going to openai), I realized that this can somewhat easily be orchestrated with a ton of docker containers with and some background logic
Kraj razgovora
Novi razgovor -
-
-
Would it have an even higher success rate if the episodes were not limited to 50 trials? Even in simulation it peaks at ~45 consecutive successes, which is oddly close to 50! I'm asking because real life robots need ~1000 straight successes to be practical
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
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.