4 changed lines results in a factor of 2 speedup for tensorflow policies

cbfinn / gps

Guided Policy Search

http://rll.berkeley.edu/gps/

Other

597 stars 239 forks source link

4 changed lines results in a factor of 2 speedup for tensorflow policies #11

Closed bstadie closed 8 years ago

bstadie commented 8 years ago

because now you do not need to compute everything twice.

@cbfinn since this is such a small change that only causes minor alteration in the internals of tensorflow policy opt, I'll just merge it immediately.

cbfinn commented 8 years ago

LGTM - thanks!