implementation of distributed reinforcement learning with distributed tensorflow
distributed tensorflow
of server-client architecture.Recurrent Experience Replay in Distributed Reinforcement Learning
is implemented in Breakout-Deterministic-v4 with POMDP(Observation not provided with 20% probability)opencv-python
gym[atari]
tensorboardX
tensorflow==1.14.0
CUDA_VISIBLE_DEVICES=-1 python train_a3c.py --job_name --job_name actor --task 0
CUDA_VISIBLE_DEVICES=-1 python train_a3c.py --job_name --job_name actor --task 0
CUDA_VISIBLE_DEVICES=-1 python train_a3c.py --job_name --job_name actor --task 1
CUDA_VISIBLE_DEVICES=-1 python train_a3c.py --job_name --job_name actor --task 2
...
CUDA_VISIBLE_DEVICES=-1 python train_a3c.py --job_name --job_name actor --task 19
python train_apex.py --job_name learner --task 0
CUDA_VISIBLE_DEVICES=-1 python train_apex.py --job_name actor --task 0
CUDA_VISIBLE_DEVICES=-1 python train_apex.py --job_name actor --task 1
CUDA_VISIBLE_DEVICES=-1 python train_apex.py --job_name actor --task 2
...
CUDA_VISIBLE_DEVICES=-1 python train_apex.py --job_name actor --task 19
python train_impala.py --job_name learner --task 0
CUDA_VISIBLE_DEVICES=-1 python train_impala.py --job_name actor --task 0
CUDA_VISIBLE_DEVICES=-1 python train_impala.py --job_name actor --task 1
CUDA_VISIBLE_DEVICES=-1 python train_impala.py --job_name actor --task 2
...
CUDA_VISIBLE_DEVICES=-1 python train_impala.py --job_name actor --task 19
python train_r2d2.py --job_name learner --task 0
CUDA_VISIBLE_DEVICES=-1 python train_r2d2.py --job_name actor --task 0
CUDA_VISIBLE_DEVICES=-1 python train_r2d2.py --job_name actor --task 1
CUDA_VISIBLE_DEVICES=-1 python train_r2d2.py --job_name actor --task 2
...
CUDA_VISIBLE_DEVICES=-1 python train_r2d2.py --job_name actor --task 39