Getting bad results When I ran drqn experiments .

[Alex-zhai]

Hello, Ask you a question. When I ran exp11, the drqn agent seems not as good as dqn. I ran five games, and the drqn results are all lower than dqn. What caused this phenomenon.

[5vision]

You are right. We are unable to reproduce exact results from the paper by Matthew et al. Probably, we should remove a fully connect layer before LSTM or use the same hyper-parameters, etc. We decided not to waste time on it. For all the results of DRQN in our paper, we used the implementation on Caffe.

[Alex-zhai]

So, the implemention on Caffe has DRQN module?

[Alex-zhai]

Can you give me some instruction of installation of DRQN on caffe? Thank you!!!

[Alex-zhai]

And your environment of running DRQN on caffe is Ubantu??

[Alex-zhai]

One more question, when I ran Hard-Darqn, I get cannot open <pretrain_seaquest.t7> in mode r at /tmp/luarocks_torch-scm-1-5561/torch7/lib/TH/THDiskFile.c:484. So where should I put the pretrain_seaquest.t7 model.