zaheersm / DeepRL

Modularized Implementation of Deep RL Algorithms in PyTorch

Run DQN with extra-replay steps for a fixed target

python run.py --id 0 --config-file experiment/config_files/valli/dqn_replay.json

Run DQN without extra replay steps

python run.py --id 0 --config-file experiment/config_files/valli/dqn_noreplay.json

Expected Sarsa with extra-replay steps for a fixed target

python run.py --id 0 --config-file experiment/config_files/valli/sarsa_target_epsilon.json

Softmax Agent

python run.py --id 0 --config-file experiment/config_files/valli/final.json