search

harvitronix / reinforcement-learning-car

Using reinforcement learning to teach a car to avoid obstacles.
MIT License
488 stars 202 forks source link

improve process_minibatch function (50~60x faster) #13

Closed Microos closed 7 years ago

Microos commented 7 years ago

I have improved this batch processing function and gained 50~60x faster speed (tested on GTX 1080) that significantly increases the training FPS

instead of feeding a batch to the model one by one, feed the whole batch at once is much more efficient

harvitronix commented 7 years ago

Thanks for adding, @Microos!

xentity commented 6 years ago

Hey there!

I never saw this before: old_state_m[...] And I get an error called Ellipsion, caused by these three dots.

I'm using python 3.5.3