DeepMind-Atari-Deep-Q-Learner icon indicating copy to clipboard operation
DeepMind-Atari-Deep-Q-Learner copied to clipboard
verified publisher icon kuz

Is there "trained" model?

Open gaoteng-git opened this issue 10 years ago • 7 comments

Since it costs a long time training (30hours on GPU) to get the smart model, would you please provide an already trained model file that I can run without training from beginning? I only have a CPU laptop so it can't afford such a long time training. Thanks a lot in advance!

gaoteng-git avatar Mar 11 '16 11:03 gaoteng-git

I've been running for ~8 hours on my GTX 970 and it can achieve high score of 10. The network file has grew to 400MiB.

baicaigithub avatar Mar 13 '16 05:03 baicaigithub

btw, it seems I cannot resume to train the model, even though I changed related fields in run_gpu.

baicaigithub avatar Mar 13 '16 10:03 baicaigithub

My bad, I set the save_freq too small, so that the network is overwritten before being analyzed.

baicaigithub avatar Mar 13 '16 21:03 baicaigithub

Could you specify how you resumed training on a network after stopping? Thanks

jtgebert avatar Aug 24 '16 17:08 jtgebert

So , do we have a trained model under some non commercial or open source license ?

scm-ns avatar Nov 03 '16 07:11 scm-ns

@jtgebert have you found anyway to resume training ?

kertansul avatar Dec 05 '16 00:12 kertansul

Nope, I set this aside for a bit.

jtgebert avatar Dec 05 '16 01:12 jtgebert