文件列表:

FD_RL (0, 2016-09-09)
FD_RL\donotdelete (0, 2016-09-09)
FeatureBased.py (5716, 2016-09-09)
NEAT_test (0, 2016-09-09)
NEAT_test\_stats.txt (6175, 2016-09-09)
NEAT_test\evolved_net.net (1357624, 2016-09-09)
NEAT_test\neatTest.py (7957, 2016-09-09)
RL_learning.py (15781, 2016-09-09)
Research (0, 2016-09-09)
Research\Bibliography (0, 2016-09-09)
Research\Bibliography\CNNs (0, 2016-09-09)
Research\Bibliography\CNNs\07410052.pdf (6313757, 2016-09-09)
Research\Bibliography\CNNs\CNN_hyperparameters.pdf (4319834, 2016-09-09)
Research\Bibliography\CNNs\Eval of CNNs visual.pdf (301354, 2016-09-09)
Research\Bibliography\CNNs\LeCun.pdf (910489, 2016-09-09)
Research\Bibliography\CNNs\SGD.pdf (113817, 2016-09-09)
Research\Bibliography\Deep reinforcement learning (0, 2016-09-09)
Research\Bibliography\Deep reinforcement learning\3D maze navegation deep reinforc.pdf (2374549, 2016-09-09)
Research\Bibliography\Deep reinforcement learning\Massively parallel deep reinforcement learning.pdf (2758553, 2016-09-09)
Research\Bibliography\Deep reinforcement learning\Mastering_the_game_of_Go_with_deep_neural_networks_and_tree_search.pdf (2682222, 2016-09-09)
Research\Bibliography\Deep reinforcement learning\ResNet.pdf (819383, 2016-09-09)
Research\Bibliography\Deep reinforcement learning\Training NN for RL without interference.pdf (86570, 2016-09-09)
Research\Bibliography\Deep reinforcement learning\ViZDoom paper.pdf (1344026, 2016-09-09)
Research\Bibliography\Deep reinforcement learning\deepgo.pdf (227605, 2016-09-09)
Research\Bibliography\Deep reinforcement learning\dqn - atari DRL.pdf (433921, 2016-09-09)
Research\Bibliography\Evolutionary deep learning (0, 2016-09-09)
Research\Bibliography\Evolutionary deep learning\Deep learning using genetic algorithms.pdf (470550, 2016-09-09)
Research\Bibliography\Evolutionary deep learning\Evolution for Deep learning.pdf (1773727, 2016-09-09)
Research\Bibliography\Evolutionary deep learning\Evolution vs TD in RL.pdf (241610, 2016-09-09)
Research\Bibliography\Evolutionary deep learning\Genetic algorithms for evolving DNN.pdf (476832, 2016-09-09)
Research\Bibliography\Evolutionary deep learning\Neuroevolution for Deep learning.pdf (707561, 2016-09-09)
Research\Bibliography\Evolutionary deep learning\Neuroevolution (0, 2016-09-09)
Research\Bibliography\Evolutionary deep learning\Neuroevolution\CoSyNE.pdf (449954, 2016-09-09)
Research\Bibliography\Evolutionary deep learning\Neuroevolution\ES-HyperNEAT (long).pdf (1766055, 2016-09-09)
Research\Bibliography\Evolutionary deep learning\Neuroevolution\ES-HyperNEAT (short).pdf (1110398, 2016-09-09)
Research\Bibliography\Evolutionary deep learning\Neuroevolution\HyperNEAT for image recognition.pdf (4112155, 2016-09-09)
Research\Bibliography\Evolutionary deep learning\Neuroevolution\HyperNEAT.pdf (2246918, 2016-09-09)
... ...

# ViZDoom_EA ViZDoom controller experiments using Evolutionary algorithms and Deep learning To use, make sure you have all ViZDoom and Keras dependencies installed. You will also need a link to the vizdoom.so library situated at ViZDoom/bin/python folder (place it on the root folder of this repository) Requirements: - Ubuntu 14.04 - ViZDoom installed and configured https://github.com/Marqt/ViZDoom - Keras installed and configured https://keras.io/ - MultiNEAT python installation https://github.com/peter-ch/MultiNEAT/tree/master/src - OpenCV 3.0 - Python concurrent.futures library - Theano (configure for GPU and cuDNN) Once ViZDoom is installed, remember to copy the vizdoom.so link library to the root folder of ViZDoom_EA before executing the code. The code expects ViZDoom to be in a sibling folder to ViZDoom_EA (so both folders should be contained in the same folder). To run a demo of one of the CNN evolved solutions on pursuit and gather scenario, simply run the python script evolve_controller.py (it is already configured). To run the solution on the health gathering scenario, edit the evolve_controller script and substitute "pursuit_and_gather" for "health_gathering_supreme" throughout lines 28 to 32. If you have any questions or want to see more, please contact me at carlos.fernandez.musoles@gmail.com

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：