Reduce the amount of calculation

andyzeng / visual-pushing-grasping

Train robotic agents to learn to plan pushing and grasping actions for manipulation with deep reinforcement learning.

http://vpg.cs.princeton.edu/

BSD 2-Clause "Simplified" License

898 stars 315 forks source link

Reduce the amount of calculation #46

Closed ChenyangRan closed 4 years ago

ChenyangRan commented 4 years ago

I want to train reactive policies while run out of memory. My GPU model is 1070Ti, 8G, which is just satisfy the memory. I want to reduce the amount of calculation but have no idea. At first I want to reduce the batchsize, but can't find. Has anyone reduced the amount of computation?

By the way, vrep3.6 can't work on Ubuntu mate16, the error is ERROR signals11. Can someone give advice?

Thanks.