Currently training at our largest-ever scale (thanks to heroic efforts from the team), which means even we can't predict how much progress Five will make this week.

Aug 20, 2018 · 6:16 PM UTC

4
10
1
115
Replying to @gdb
Running on Azure or own data center? What hardware do you train on?
1
Running on GCP!
17
Replying to @gdb
Will the restrictions present for the casters’ matches still be in place?
1
Replying to @gdb
Could you provide the human team with the OpenAI draft simulator? It would be interesting to see how well humans do in an ~50% predicted winrate game, conditioned on the humans levels of comfort with the heroes.
1
2
Replying to @gdb
In your training can you recognize what kind of new strategies (macro/micro) it has learned so far? Is it even possible to define this strategies? How do you calculate progress?
1
This tweet is unavailable
They couldn't see it continuously, but they can see Five's chats about the win prob when it hits certain thresholds. Was a bit of an experiment to see how humans would react :).
3