PyTorch implementation of Advantage Actor Critic (A2C), Proximal Policy Optimization (PPO), Scalable trust-region method for deep reinforcement learning using Kronecker-factored approximation (ACKTR) and Generative Adversarial Imitation Learning (GAIL).
I can run the PongNoFrameskip-v4 environment successfully by
python main.py --env-name "PongNoFrameskip-v4" --algo ppo --use-gae --lr 2.5e-4 --clip-param 0.1 --value-loss-coef 0.5 --num-processes 8 --num-steps 128 --num-mini-batch 4 --log-interval 1 --use-linear-lr-decay --entropy-coef 0.01
However, when I change the env-name to other name (AirRaid-ram-v0, DemonAttack-ram-v0, DemonAttack-v0 ...),
there is an error :
Traceback (most recent call last):
File "/home/lq/pytorch-a2c-ppo-acktr-gail/main.py", line 240, in <module>
main()
File "/home/lq/pytorch-a2c-ppo-acktr-gail/main.py", line 68, in main
args.gamma, args.log_dir, device, False)
File "/home/lq/pytorch-a2c-ppo-acktr-gail/a2c_ppo_acktr/envs.py", line 89, in make_vec_envs
envs = ShmemVecEnv(envs, context='fork')
File "/home/lq/pytorch-a2c-ppo-acktr-gail/baselines/baselines/common/vec_env/shmem_vec_env.py", line 36, in __init__
dummy = env_fns[0]()
File "/home/lq/pytorch-a2c-ppo-acktr-gail/a2c_ppo_acktr/envs.py", line 43, in _thunk
env = make_atari(env_id)
File "/home/lq/pytorch-a2c-ppo-acktr-gail/baselines/baselines/common/atari_wrappers.py", line 268, in make_atari
assert 'NoFrameskip' in env.spec.id
AssertionError
Thanks for your great work.
I'm facing a problem when I change the
evn-name
I can run the PongNoFrameskip-v4 environment successfully by
python main.py --env-name "PongNoFrameskip-v4" --algo ppo --use-gae --lr 2.5e-4 --clip-param 0.1 --value-loss-coef 0.5 --num-processes 8 --num-steps 128 --num-mini-batch 4 --log-interval 1 --use-linear-lr-decay --entropy-coef 0.01
However, when I change the env-name to other name (AirRaid-ram-v0, DemonAttack-ram-v0, DemonAttack-v0 ...), there is an error :
Can you give me some suggestions? Thanks a lot!