Sherjil Ozair

Isn't `output_grads` a list (of one Theano tensor) ?

Cool. I'll test this with some other examples, and push it in. Thanks!

Hi @milesbrundage, Thanks for trying out the code. I like your ideas for improvement. I had been trying similar stuff myself, but didn't find time to complete the experiments. If...

it should be possible to use `tf.stop_gradient` to do this. However, this change should ideally belong to Keras backend, so that the DQN code becomes backend-agnostic.

Hi @ShibiHe, thanks for your comment. You're right. This is a bug. I should be using `np.arange(n)` instead of `:`.

You should try letting it run for more number of games. You should be able to get better performance, I think. Feel free to add this in either the root...

Thanks for the report. @Alicemargatroid @izqui. I need some help figuring out the right way to fix this problem. How big is the `data.npy` file? Is it 20 times large...

@leafo Is it possible to display correct error messages when using moonscript and love2d? Having to read compiled lua code to debug is not fun at all.