Sahil Sharma
Sahil Sharma
Hi Oogi, I am facing a similar issue. I have described it [here](https://groups.google.com/forum/#!topic/theano-users/ycmGFMDXhD0). Please let me know if you figured out the problem. Thanks
Thanks, I'll try that. Also roughly how many `episode_count` or `max_steps` did you have to train for before you arrived at the saved `.h5` models ?
Oh, thanks. Which tracks are simple? And which ones are more complicated? I usually run torcs in practice mode. Is that one of the simpler ones?
Hi. I don;t experience this error. Which dataset are you running your experiments on?
I used celebA as well. I use tensorflow 0.10 GPU version. I use pip install to install TF.
@apeterswu The reason for drawing from probability distribution over discrete action space is two-fold: 1. During Training this helps in exploration. Suppose you think that action 1 is the best...
@lforg37 Under the assumption that the observations could be aggregated to get a MDP-like state, my statement is true :) The case you mention is that of a POMDP. Typically...
@lforg37 thanks for pointing out that case :) I hadn't considered it.
Btw the code which @congling has used is not equivalent to resetting game on life lost. :p This is because all it says is that if you just lost a...