ray/rllib/tuned_examples/regression_tests/cartpole-ppo-torch.yaml

cartpole-ppo-torch:
    env: CartPole-v0
    run: PPO
    stop:
        episode_reward_mean: 150
        timesteps_total: 100000
    config:
        use_pytorch: true
        gamma: 0.99
        lr: 0.0003
        num_workers: 1
        observation_filter: MeanStdFilter
        num_sgd_iter: 6
        vf_share_layers: true
        vf_loss_coeff: 0.01
        model:
          fcnet_hiddens: [32]
          fcnet_activation: linear
[RLlib] PPO(torch) on CartPole not tuned well enough for consistent learning (#7556) 2020-03-12 04:31:27 +01:00			`cartpole-ppo-torch:`
Add cartpole PPO torch to regression (besides tf). (#7005) 2020-02-02 02:41:38 +01:00			`env: CartPole-v0`
			`run: PPO`
			`stop:`
			`episode_reward_mean: 150`
			`timesteps_total: 100000`
			`config:`
[RLlib] PPO(torch) on CartPole not tuned well enough for consistent learning (#7556) 2020-03-12 04:31:27 +01:00			`use_pytorch: true`
			`gamma: 0.99`
			`lr: 0.0003`
Add cartpole PPO torch to regression (besides tf). (#7005) 2020-02-02 02:41:38 +01:00			`num_workers: 1`
			`observation_filter: MeanStdFilter`
[RLlib] PPO(torch) on CartPole not tuned well enough for consistent learning (#7556) 2020-03-12 04:31:27 +01:00			`num_sgd_iter: 6`
			`vf_share_layers: true`
			`vf_loss_coeff: 0.01`
			`model:`
			`fcnet_hiddens: [32]`
			`fcnet_activation: linear`