From 2afc4744e0f4231fdc6fc86b7af9c9c45e437dc7 Mon Sep 17 00:00:00 2001 From: rickstaa Date: Thu, 7 Mar 2024 17:56:09 +0100 Subject: [PATCH] chore(exp): fix CompOscillator learning length This commit changed the CompOscillator learning length from 1e5 to 2e5. This was done to deal with an inconsistency in Han et al.'s reserach. --- ...lpha3_tune_experiment_extra_seed26000.yml} | 2 +- ...pha3_tune_experiment_extra_seed388389.yml} | 2 +- ...pha3_tune_experiment_extra_seed408660.yml} | 2 +- ...lpha3_tune_experiment_extra_seed49672.yml} | 2 +- ...pha3_tune_experiment_extra_seed858762.yml} | 2 +- ...ted_alpha3_tune_experiment_extra_seeds.yml | 2 +- ..._tune_experiment_extra_seed26000_short.yml | 48 +++++++++++++++++++ ...tune_experiment_extra_seed388389_short.yml | 48 +++++++++++++++++++ ...tune_experiment_extra_seed408660_short.yml | 48 +++++++++++++++++++ ..._tune_experiment_extra_seed49672_short.yml | 48 +++++++++++++++++++ ...tune_experiment_extra_seed858762_short.yml | 48 +++++++++++++++++++ ...pha3_tune_experiment_extra_seeds_short.yml | 48 +++++++++++++++++++ ...tor_complicated_alpha3_tune_experiment.yml | 2 +- ...licated_alpha3_tune_experiment_seed234.yml | 2 +- ...icated_alpha3_tune_experiment_seed3658.yml | 2 +- ...cated_alpha3_tune_experiment_seed48104.yml | 2 +- ...licated_alpha3_tune_experiment_seed567.yml | 2 +- ...cated_alpha3_tune_experiment_seed78456.yml | 2 +- ...tune_experiment_lambda_lr_lambda_check.yml | 2 +- ...une_experiment_seed234_lr_lambda_check.yml | 2 +- ...ne_experiment_seed3658_lr_lambda_check.yml | 2 +- ...e_experiment_seed48104_lr_lambda_check.yml | 2 +- ...une_experiment_seed567_lr_lambda_check.yml | 2 +- ...e_experiment_seed78456_lr_lambda_check.yml | 2 +- ..._alpha3_tune_experiment_seed234_short.yml} | 6 +-- ...alpha3_tune_experiment_seed3658_short.yml} | 6 +-- ...lpha3_tune_experiment_seed48104_short.yml} | 6 +-- ..._alpha3_tune_experiment_seed567_short.yml} | 6 +-- ...lpha3_tune_experiment_seed78456_short.yml} | 6 +-- ...plicated_alpha3_tune_experiment_short.yml} | 6 +-- ...3_tune_experiment_seed234_small_critic.yml | 2 +- ..._tune_experiment_seed3658_small_critic.yml | 2 +- ...tune_experiment_seed48104_small_critic.yml | 2 +- ...3_tune_experiment_seed567_small_critic.yml | 2 +- ...tune_experiment_seed78456_small_critic.yml | 2 +- ...ed_alpha3_tune_experiment_small_critic.yml | 2 +- 36 files changed, 330 insertions(+), 42 deletions(-) rename experiments/staa_et_al_2024/comp_oscillator/extra_seeds/{han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds26000.yml => han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed26000.yml} (99%) rename experiments/staa_et_al_2024/comp_oscillator/extra_seeds/{han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds388389.yml => han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed388389.yml} (99%) rename experiments/staa_et_al_2024/comp_oscillator/extra_seeds/{han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds408660.yml => han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed408660.yml} (99%) rename experiments/staa_et_al_2024/comp_oscillator/extra_seeds/{han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds49672.yml => han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed49672.yml} (99%) rename experiments/staa_et_al_2024/comp_oscillator/extra_seeds/{han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds858762.yml => han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed858762.yml} (99%) create mode 100644 experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed26000_short.yml create mode 100644 experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed388389_short.yml create mode 100644 experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed408660_short.yml create mode 100644 experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed49672_short.yml create mode 100644 experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed858762_short.yml create mode 100644 experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds_short.yml rename experiments/staa_et_al_2024/comp_oscillator/{extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_extra_long.yml => short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_short.yml} (95%) rename experiments/staa_et_al_2024/comp_oscillator/{extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_extra_long.yml => short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_short.yml} (95%) rename experiments/staa_et_al_2024/comp_oscillator/{extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_extra_long.yml => short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_short.yml} (95%) rename experiments/staa_et_al_2024/comp_oscillator/{extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_extra_long.yml => short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_short.yml} (95%) rename experiments/staa_et_al_2024/comp_oscillator/{extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_extra_long.yml => short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_short.yml} (95%) rename experiments/staa_et_al_2024/comp_oscillator/{extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_long.yml => short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_short.yml} (95%) diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds26000.yml b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed26000.yml similarity index 99% rename from experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds26000.yml rename to experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed26000.yml index 03739374..5e55e1f9 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds26000.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed26000.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds388389.yml b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed388389.yml similarity index 99% rename from experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds388389.yml rename to experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed388389.yml index 1e038d5c..001d56ca 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds388389.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed388389.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds408660.yml b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed408660.yml similarity index 99% rename from experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds408660.yml rename to experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed408660.yml index e07a48cc..a2ff8019 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds408660.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed408660.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds49672.yml b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed49672.yml similarity index 99% rename from experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds49672.yml rename to experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed49672.yml index 85c63807..4be0b6a3 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds49672.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed49672.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds858762.yml b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed858762.yml similarity index 99% rename from experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds858762.yml rename to experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed858762.yml index de809d42..e4b5569f 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds858762.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed858762.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds.yml b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds.yml index dd59c037..19379dbf 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed26000_short.yml b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed26000_short.yml new file mode 100644 index 00000000..ba182156 --- /dev/null +++ b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed26000_short.yml @@ -0,0 +1,48 @@ +alg_name: lac +exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_short +env_name: "stable_gym:OscillatorComplicated-v1" +ac_kwargs: + hidden_sizes: + actor: [256, 256] # NOTE: Using [256, 256] for consistency with the article. + critic: [256, 256, 16] # NOTE: Using [256, 256, 16] for consistency with the article. + activation: + actor: "nn.ReLU" + critic: "nn.ReLU" + output_activation: + actor: "nn.ReLU" +opt_type: "minimize" +max_ep_len: 400 +epochs: 49 +steps_per_epoch: 2048 +start_steps: 0 +update_every: 100 +update_after: 1000 +steps_per_update: 80 +num_test_episodes: 10 +alpha: 2.0 +alpha3: 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 # NOTE: Tuning alpha3. +labda: 0.99 # NOTE: Decreased from 1.0 to 0.99 for stability. +# gamma: 0.995 # NOTE: Not used for finite horizon tasks. +polyak: 0.995 +adaptive_temperature: True +lr_a: "1e-4" +lr_c: "3e-4" +lr_alpha: "1e-4" +lr_labda: "3e-4" +lr_a_final: "1e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_c_final: "3e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_alpha_final: "1e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_labda_final: "3e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_a_decay_type: "linear" +lr_c_decay_type: "linear" +lr_alpha_decay_type: "linear" +lr_labda_decay_type: "constant" +lr_decay_ref: "step" +batch_size: 256 +replay_size: "int(1e6)" +horizon_length: 5 +seed: 26000 # 49672 858762 388389 # NOTE: Using 5 seeds for tuning. +save_freq: 10 +use_wandb: True +wandb_group: "han2020_reproduction_extra_seeds_short" +device: "gpu:1" diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed388389_short.yml b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed388389_short.yml new file mode 100644 index 00000000..a20ccee4 --- /dev/null +++ b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed388389_short.yml @@ -0,0 +1,48 @@ +alg_name: lac +exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_short +env_name: "stable_gym:OscillatorComplicated-v1" +ac_kwargs: + hidden_sizes: + actor: [256, 256] # NOTE: Using [256, 256] for consistency with the article. + critic: [256, 256, 16] # NOTE: Using [256, 256, 16] for consistency with the article. + activation: + actor: "nn.ReLU" + critic: "nn.ReLU" + output_activation: + actor: "nn.ReLU" +opt_type: "minimize" +max_ep_len: 400 +epochs: 49 +steps_per_epoch: 2048 +start_steps: 0 +update_every: 100 +update_after: 1000 +steps_per_update: 80 +num_test_episodes: 10 +alpha: 2.0 +alpha3: 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 # NOTE: Tuning alpha3. +labda: 0.99 # NOTE: Decreased from 1.0 to 0.99 for stability. +# gamma: 0.995 # NOTE: Not used for finite horizon tasks. +polyak: 0.995 +adaptive_temperature: True +lr_a: "1e-4" +lr_c: "3e-4" +lr_alpha: "1e-4" +lr_labda: "3e-4" +lr_a_final: "1e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_c_final: "3e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_alpha_final: "1e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_labda_final: "3e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_a_decay_type: "linear" +lr_c_decay_type: "linear" +lr_alpha_decay_type: "linear" +lr_labda_decay_type: "constant" +lr_decay_ref: "step" +batch_size: 256 +replay_size: "int(1e6)" +horizon_length: 5 +seed: 388389 # NOTE: Using 5 seeds for tuning. +save_freq: 10 +use_wandb: True +wandb_group: "han2020_reproduction_extra_seeds_short" +device: "gpu:1" diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed408660_short.yml b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed408660_short.yml new file mode 100644 index 00000000..4f9817fb --- /dev/null +++ b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed408660_short.yml @@ -0,0 +1,48 @@ +alg_name: lac +exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_short +env_name: "stable_gym:OscillatorComplicated-v1" +ac_kwargs: + hidden_sizes: + actor: [256, 256] # NOTE: Using [256, 256] for consistency with the article. + critic: [256, 256, 16] # NOTE: Using [256, 256, 16] for consistency with the article. + activation: + actor: "nn.ReLU" + critic: "nn.ReLU" + output_activation: + actor: "nn.ReLU" +opt_type: "minimize" +max_ep_len: 400 +epochs: 49 +steps_per_epoch: 2048 +start_steps: 0 +update_every: 100 +update_after: 1000 +steps_per_update: 80 +num_test_episodes: 10 +alpha: 2.0 +alpha3: 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 # NOTE: Tuning alpha3. +labda: 0.99 # NOTE: Decreased from 1.0 to 0.99 for stability. +# gamma: 0.995 # NOTE: Not used for finite horizon tasks. +polyak: 0.995 +adaptive_temperature: True +lr_a: "1e-4" +lr_c: "3e-4" +lr_alpha: "1e-4" +lr_labda: "3e-4" +lr_a_final: "1e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_c_final: "3e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_alpha_final: "1e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_labda_final: "3e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_a_decay_type: "linear" +lr_c_decay_type: "linear" +lr_alpha_decay_type: "linear" +lr_labda_decay_type: "constant" +lr_decay_ref: "step" +batch_size: 256 +replay_size: "int(1e6)" +horizon_length: 5 +seed: 408660 # 26000 49672 858762 388389 # NOTE: Using 5 seeds for tuning. +save_freq: 10 +use_wandb: True +wandb_group: "han2020_reproduction_extra_seeds_short" +device: "gpu:1" diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed49672_short.yml b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed49672_short.yml new file mode 100644 index 00000000..34ec4dd6 --- /dev/null +++ b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed49672_short.yml @@ -0,0 +1,48 @@ +alg_name: lac +exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_short +env_name: "stable_gym:OscillatorComplicated-v1" +ac_kwargs: + hidden_sizes: + actor: [256, 256] # NOTE: Using [256, 256] for consistency with the article. + critic: [256, 256, 16] # NOTE: Using [256, 256, 16] for consistency with the article. + activation: + actor: "nn.ReLU" + critic: "nn.ReLU" + output_activation: + actor: "nn.ReLU" +opt_type: "minimize" +max_ep_len: 400 +epochs: 49 +steps_per_epoch: 2048 +start_steps: 0 +update_every: 100 +update_after: 1000 +steps_per_update: 80 +num_test_episodes: 10 +alpha: 2.0 +alpha3: 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 # NOTE: Tuning alpha3. +labda: 0.99 # NOTE: Decreased from 1.0 to 0.99 for stability. +# gamma: 0.995 # NOTE: Not used for finite horizon tasks. +polyak: 0.995 +adaptive_temperature: True +lr_a: "1e-4" +lr_c: "3e-4" +lr_alpha: "1e-4" +lr_labda: "3e-4" +lr_a_final: "1e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_c_final: "3e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_alpha_final: "1e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_labda_final: "3e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_a_decay_type: "linear" +lr_c_decay_type: "linear" +lr_alpha_decay_type: "linear" +lr_labda_decay_type: "constant" +lr_decay_ref: "step" +batch_size: 256 +replay_size: "int(1e6)" +horizon_length: 5 +seed: 49672 # 858762 388389 # NOTE: Using 5 seeds for tuning. +save_freq: 10 +use_wandb: True +wandb_group: "han2020_reproduction_extra_seeds_short" +device: "gpu:1" diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed858762_short.yml b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed858762_short.yml new file mode 100644 index 00000000..96d89d23 --- /dev/null +++ b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seed858762_short.yml @@ -0,0 +1,48 @@ +alg_name: lac +exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_short +env_name: "stable_gym:OscillatorComplicated-v1" +ac_kwargs: + hidden_sizes: + actor: [256, 256] # NOTE: Using [256, 256] for consistency with the article. + critic: [256, 256, 16] # NOTE: Using [256, 256, 16] for consistency with the article. + activation: + actor: "nn.ReLU" + critic: "nn.ReLU" + output_activation: + actor: "nn.ReLU" +opt_type: "minimize" +max_ep_len: 400 +epochs: 49 +steps_per_epoch: 2048 +start_steps: 0 +update_every: 100 +update_after: 1000 +steps_per_update: 80 +num_test_episodes: 10 +alpha: 2.0 +alpha3: 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 # NOTE: Tuning alpha3. +labda: 0.99 # NOTE: Decreased from 1.0 to 0.99 for stability. +# gamma: 0.995 # NOTE: Not used for finite horizon tasks. +polyak: 0.995 +adaptive_temperature: True +lr_a: "1e-4" +lr_c: "3e-4" +lr_alpha: "1e-4" +lr_labda: "3e-4" +lr_a_final: "1e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_c_final: "3e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_alpha_final: "1e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_labda_final: "3e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_a_decay_type: "linear" +lr_c_decay_type: "linear" +lr_alpha_decay_type: "linear" +lr_labda_decay_type: "constant" +lr_decay_ref: "step" +batch_size: 256 +replay_size: "int(1e6)" +horizon_length: 5 +seed: 858762 # 388389 # NOTE: Using 5 seeds for tuning. +save_freq: 10 +use_wandb: True +wandb_group: "han2020_reproduction_extra_seeds_short" +device: "gpu:1" diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds_short.yml b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds_short.yml new file mode 100644 index 00000000..4c6a1588 --- /dev/null +++ b/experiments/staa_et_al_2024/comp_oscillator/extra_seeds/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_seeds_short.yml @@ -0,0 +1,48 @@ +alg_name: lac +exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_short +env_name: "stable_gym:OscillatorComplicated-v1" +ac_kwargs: + hidden_sizes: + actor: [256, 256] # NOTE: Using [256, 256] for consistency with the article. + critic: [256, 256, 16] # NOTE: Using [256, 256, 16] for consistency with the article. + activation: + actor: "nn.ReLU" + critic: "nn.ReLU" + output_activation: + actor: "nn.ReLU" +opt_type: "minimize" +max_ep_len: 400 +epochs: 49 +steps_per_epoch: 2048 +start_steps: 0 +update_every: 100 +update_after: 1000 +steps_per_update: 80 +num_test_episodes: 10 +alpha: 2.0 +alpha3: 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 # NOTE: Tuning alpha3. +labda: 0.99 # NOTE: Decreased from 1.0 to 0.99 for stability. +# gamma: 0.995 # NOTE: Not used for finite horizon tasks. +polyak: 0.995 +adaptive_temperature: True +lr_a: "1e-4" +lr_c: "3e-4" +lr_alpha: "1e-4" +lr_labda: "3e-4" +lr_a_final: "1e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_c_final: "3e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_alpha_final: "1e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_labda_final: "3e-9" # NOTE: Not in Han et al. 2020, but aligned with original code. +lr_a_decay_type: "linear" +lr_c_decay_type: "linear" +lr_alpha_decay_type: "linear" +lr_labda_decay_type: "constant" +lr_decay_ref: "step" +batch_size: 256 +replay_size: "int(1e6)" +horizon_length: 5 +seed: 408660 26000 49672 858762 388389 # NOTE: Using 5 seeds for tuning. +save_freq: 10 +use_wandb: True +wandb_group: "han2020_reproduction_extra_seeds_short" +device: "gpu:1" diff --git a/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment.yml b/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment.yml index bd193743..7fda4369 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234.yml b/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234.yml index 20e8e818..f53ae49f 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658.yml b/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658.yml index 83966c1d..e1d11fe3 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104.yml b/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104.yml index 666ab06e..dfc620f6 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567.yml b/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567.yml index f374af70..bfa235fb 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456.yml b/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456.yml index 169dfd61..1154590b 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_lambda_lr_lambda_check.yml b/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_lambda_lr_lambda_check.yml index 97f99d2b..a9f9c81b 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_lambda_lr_lambda_check.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_lambda_lr_lambda_check.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_lr_lambda_check.yml b/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_lr_lambda_check.yml index d7841fdf..07ff320e 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_lr_lambda_check.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_lr_lambda_check.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_lr_lambda_check.yml b/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_lr_lambda_check.yml index f2ea47ca..688b3d2e 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_lr_lambda_check.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_lr_lambda_check.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_lr_lambda_check.yml b/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_lr_lambda_check.yml index fbd3778a..20c7eb8b 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_lr_lambda_check.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_lr_lambda_check.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_lr_lambda_check.yml b/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_lr_lambda_check.yml index 6b0d6db5..c048d0d1 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_lr_lambda_check.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_lr_lambda_check.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_lr_lambda_check.yml b/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_lr_lambda_check.yml index faff9ae2..3f81ca8f 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_lr_lambda_check.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/lambda_lr_check/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_lr_lambda_check.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_extra_long.yml b/experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_short.yml similarity index 95% rename from experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_extra_long.yml rename to experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_short.yml index e43dc155..ca3edf00 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_extra_long.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_short.yml @@ -1,5 +1,5 @@ alg_name: lac -exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_extra_long +exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_short env_name: "stable_gym:OscillatorComplicated-v1" ac_kwargs: hidden_sizes: @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 98 +epochs: 49 steps_per_epoch: 2048 start_steps: 0 update_every: 100 @@ -44,5 +44,5 @@ horizon_length: 5 seed: 234 # NOTE: Using 5 seeds for tuning. save_freq: 10 use_wandb: True -wandb_group: "han2020_reproduction_extra_long" +wandb_group: "han2020_reproduction_short" device: "gpu:1" diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_extra_long.yml b/experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_short.yml similarity index 95% rename from experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_extra_long.yml rename to experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_short.yml index 771fdc93..3dcd6cb0 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_extra_long.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_short.yml @@ -1,5 +1,5 @@ alg_name: lac -exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_extra_long +exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_short env_name: "stable_gym:OscillatorComplicated-v1" ac_kwargs: hidden_sizes: @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 98 +epochs: 49 steps_per_epoch: 2048 start_steps: 0 update_every: 100 @@ -44,5 +44,5 @@ horizon_length: 5 seed: 3658 # 234 # NOTE: Using 5 seeds for tuning. save_freq: 10 use_wandb: True -wandb_group: "han2020_reproduction_extra_long" +wandb_group: "han2020_reproduction_short" device: "gpu:1" diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_extra_long.yml b/experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_short.yml similarity index 95% rename from experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_extra_long.yml rename to experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_short.yml index c1e2e496..4162470b 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_extra_long.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_short.yml @@ -1,5 +1,5 @@ alg_name: lac -exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_extra_long +exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_short env_name: "stable_gym:OscillatorComplicated-v1" ac_kwargs: hidden_sizes: @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 98 +epochs: 49 steps_per_epoch: 2048 start_steps: 0 update_every: 100 @@ -44,5 +44,5 @@ horizon_length: 5 seed: 48104 # 567 3658 234 # NOTE: Using 5 seeds for tuning. save_freq: 10 use_wandb: True -wandb_group: "han2020_reproduction_extra_long" +wandb_group: "han2020_reproduction_short" device: "gpu:1" diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_extra_long.yml b/experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_short.yml similarity index 95% rename from experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_extra_long.yml rename to experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_short.yml index c12f0ca4..66dc5b76 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_extra_long.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_short.yml @@ -1,5 +1,5 @@ alg_name: lac -exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_extra_long +exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_short env_name: "stable_gym:OscillatorComplicated-v1" ac_kwargs: hidden_sizes: @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 98 +epochs: 49 steps_per_epoch: 2048 start_steps: 0 update_every: 100 @@ -44,5 +44,5 @@ horizon_length: 5 seed: 567 # 3658 234 # NOTE: Using 5 seeds for tuning. save_freq: 10 use_wandb: True -wandb_group: "han2020_reproduction_extra_long" +wandb_group: "han2020_reproduction_short" device: "gpu:1" diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_extra_long.yml b/experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_short.yml similarity index 95% rename from experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_extra_long.yml rename to experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_short.yml index 5bfd2b00..61f38aa3 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_extra_long.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_short.yml @@ -1,5 +1,5 @@ alg_name: lac -exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_extra_long +exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_short env_name: "stable_gym:OscillatorComplicated-v1" ac_kwargs: hidden_sizes: @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 98 +epochs: 49 steps_per_epoch: 2048 start_steps: 0 update_every: 100 @@ -44,5 +44,5 @@ horizon_length: 5 seed: 78456 # 48104 567 3658 234 # NOTE: Using 5 seeds for tuning. save_freq: 10 use_wandb: True -wandb_group: "han2020_reproduction_extra_long" +wandb_group: "han2020_reproduction_short" device: "gpu:1" diff --git a/experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_long.yml b/experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_short.yml similarity index 95% rename from experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_long.yml rename to experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_short.yml index 0219fa05..284b88cc 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/extra_long/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_extra_long.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/short/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_short.yml @@ -1,5 +1,5 @@ alg_name: lac -exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_extra_long +exp_name: han2020_reproduction_lac_oscillator_complicated_alpha3_tune_exp_short env_name: "stable_gym:OscillatorComplicated-v1" ac_kwargs: hidden_sizes: @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 98 +epochs: 49 steps_per_epoch: 2048 start_steps: 0 update_every: 100 @@ -44,5 +44,5 @@ horizon_length: 5 seed: 78456 48104 567 3658 234 # NOTE: Using 5 seeds for tuning. save_freq: 10 use_wandb: True -wandb_group: "han2020_reproduction_extra_long" +wandb_group: "han2020_reproduction_short" device: "gpu:1" diff --git a/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_small_critic.yml b/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_small_critic.yml index 73a88249..142a22ea 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_small_critic.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed234_small_critic.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_small_critic.yml b/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_small_critic.yml index 438c3481..18922b39 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_small_critic.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed3658_small_critic.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_small_critic.yml b/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_small_critic.yml index 62fd3e51..e6f17a06 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_small_critic.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed48104_small_critic.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_small_critic.yml b/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_small_critic.yml index 8ea88207..75dd03f2 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_small_critic.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed567_small_critic.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_small_critic.yml b/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_small_critic.yml index 20df351f..bf25ff1d 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_small_critic.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_seed78456_small_critic.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100 diff --git a/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_small_critic.yml b/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_small_critic.yml index e652f328..4d75f53d 100644 --- a/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_small_critic.yml +++ b/experiments/staa_et_al_2024/comp_oscillator/small_critic/han2020_reproduction_lac_oscillator_complicated_alpha3_tune_experiment_small_critic.yml @@ -12,7 +12,7 @@ ac_kwargs: actor: "nn.ReLU" opt_type: "minimize" max_ep_len: 400 -epochs: 49 +epochs: 98 steps_per_epoch: 2048 start_steps: 0 update_every: 100