diff --git a/configs/S4/larva/seed_4.yaml b/configs/S4/larva/seed_4.yaml new file mode 100644 index 0000000000000000000000000000000000000000..8542e407c030d6450d7c8045af9995ca8c914c29 --- /dev/null +++ b/configs/S4/larva/seed_4.yaml @@ -0,0 +1,107 @@ +connectome: + frozen_graph_path: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/graphs/larva_row_l1.pkl + frozen_graph_sha256: 255dfe9d4dedba1b08d02bf4140a650de4cd80d6778b2e14d487fd1be2be6c76 + global_gain: 1.0 + kind: original + normalization: row_l1 + sensory_type: all + species: larva +continuous_gait_experiment: + action_dimension_matched: true + baseline: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/adult_seed4.yaml + cpg_freedom: duty_warping_plus_phase_family + primary_checkpoint: fixed_budget_final + target_sampling: uniform_episode + training_unit: one fresh Larva seed, matching Adult training protocol +env: + action_mode: command2 + command_adapter_config: + base_frequency_hz: 12.0 + continuous_gait: true + fast_duty_factor: 0.58 + fast_ipsilateral_lag_cycles: 0.5 + frequency_bounds: + - 0.25 + - 2.0 + gait_smoothing_alpha: 0.005 + k_drive: 0.75 + k_drive_step: 1.0 + k_step: 0.75 + slow_duty_factor: 0.72 + slow_ipsilateral_lag_cycles: 0.3333333333333333 + step_bounds: + - 0.4 + - 2.0 + swing_duration_s: 0.035 + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + speed_filter_alpha: 0.25 + speed_heading_rate_weight: 0.25 + speed_target_range_mm_s: + - 8.0 + - 16.0 + speed_target_sampling: uniform + speed_target_scale_mm_s: 20.0 + speed_targets_mm_s: + - 8.0 + - 16.0 + task: speed_control + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 +initialization: + critic_state_from: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/initialization/critic_seed4.pt +interface_hierarchy: + batch: EXP04_continuous_speed_larva6_20260928 + connectome: larva + interface: command2 + label: larva + smoke: false + task: speed_control + train_seed: 4 +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + environment_seed: 14000 + gamma: 0.99 + initial_hidden_mode: zero + initialization_seed: 4 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + minibatch_seed: 30004 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + policy_sampling_seed: 20004 + rollout_heartbeat: 32 + separate_rng_streams: true + seq_len: 32 + total_timesteps: 499200 + vf_clip_range: 0.2 + vf_coef: 0.5 +resume_policy: restart_only +seed: 4 +vec_env: subproc diff --git a/configs/S4/larva/seed_5.yaml b/configs/S4/larva/seed_5.yaml new file mode 100644 index 0000000000000000000000000000000000000000..329c56c1bc143097f9490f72097ff7e5c0fab21c --- /dev/null +++ b/configs/S4/larva/seed_5.yaml @@ -0,0 +1,107 @@ +connectome: + frozen_graph_path: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/graphs/larva_row_l1.pkl + frozen_graph_sha256: 255dfe9d4dedba1b08d02bf4140a650de4cd80d6778b2e14d487fd1be2be6c76 + global_gain: 1.0 + kind: original + normalization: row_l1 + sensory_type: all + species: larva +continuous_gait_experiment: + action_dimension_matched: true + baseline: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/adult_seed5.yaml + cpg_freedom: duty_warping_plus_phase_family + primary_checkpoint: fixed_budget_final + target_sampling: uniform_episode + training_unit: one fresh Larva seed, matching Adult training protocol +env: + action_mode: command2 + command_adapter_config: + base_frequency_hz: 12.0 + continuous_gait: true + fast_duty_factor: 0.58 + fast_ipsilateral_lag_cycles: 0.5 + frequency_bounds: + - 0.25 + - 2.0 + gait_smoothing_alpha: 0.005 + k_drive: 0.75 + k_drive_step: 1.0 + k_step: 0.75 + slow_duty_factor: 0.72 + slow_ipsilateral_lag_cycles: 0.3333333333333333 + step_bounds: + - 0.4 + - 2.0 + swing_duration_s: 0.035 + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + speed_filter_alpha: 0.25 + speed_heading_rate_weight: 0.25 + speed_target_range_mm_s: + - 8.0 + - 16.0 + speed_target_sampling: uniform + speed_target_scale_mm_s: 20.0 + speed_targets_mm_s: + - 8.0 + - 16.0 + task: speed_control + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 +initialization: + critic_state_from: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/initialization/critic_seed5.pt +interface_hierarchy: + batch: EXP04_continuous_speed_larva6_20260928 + connectome: larva + interface: command2 + label: larva + smoke: false + task: speed_control + train_seed: 5 +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + environment_seed: 15000 + gamma: 0.99 + initial_hidden_mode: zero + initialization_seed: 5 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + minibatch_seed: 30005 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + policy_sampling_seed: 20005 + rollout_heartbeat: 32 + separate_rng_streams: true + seq_len: 32 + total_timesteps: 499200 + vf_clip_range: 0.2 + vf_coef: 0.5 +resume_policy: restart_only +seed: 5 +vec_env: subproc diff --git a/configs/S5/banc_descending/seed_0.yaml b/configs/S5/banc_descending/seed_0.yaml new file mode 100644 index 0000000000000000000000000000000000000000..7f121d3df308e95df108c59bbaca2994f52f920d --- /dev/null +++ b/configs/S5/banc_descending/seed_0.yaml @@ -0,0 +1,120 @@ +interface_hierarchy: + batch: EXP04_continuous_speed_matched6_20260921 + connectome: banc_descending + interface: command2 + task: speed_control + train_seed: 0 + smoke: false + label: banc_descending +vec_env: subproc +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - descending + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc + frozen_graph_path: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/graphs/banc_descending.pkl + frozen_graph_sha256: 86468ba0ad24d42cb8507741c05830da720573cf3df188574d6ffa1309aad915 +env: + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 + command_adapter_config: + frequency_bounds: + - 0.25 + - 2.0 + k_drive: 0.75 + k_drive_step: 1.0 + k_step: 0.75 + step_bounds: + - 0.4 + - 2.0 + continuous_gait: true + slow_duty_factor: 0.72 + fast_duty_factor: 0.58 + slow_ipsilateral_lag_cycles: 0.3333333333333333 + fast_ipsilateral_lag_cycles: 0.5 + swing_duration_s: 0.035 + base_frequency_hz: 12.0 + gait_smoothing_alpha: 0.005 + speed_filter_alpha: 0.25 + speed_heading_rate_weight: 0.25 + speed_target_scale_mm_s: 20.0 + speed_targets_mm_s: + - 8.0 + - 16.0 + action_mode: command2 + task: speed_control + speed_target_sampling: uniform + speed_target_range_mm_s: + - 8.0 + - 16.0 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 32 + seq_len: 32 + total_timesteps: 499200 + vf_clip_range: 0.2 + vf_coef: 0.5 + separate_rng_streams: true + initialization_seed: 0 + environment_seed: 10000 + policy_sampling_seed: 20000 + minibatch_seed: 30000 + initial_hidden_mode: zero +continuous_gait_experiment: + baseline: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed/configs/speed_continuous_gait_banc_descending_command2_seed0.yaml + target_sampling: uniform_episode + cpg_freedom: duty_warping_plus_phase_family + action_dimension_matched: true + primary_checkpoint: fixed_budget_final + training_unit: one complete fresh train seed; ER also independent graph seed +seed: 0 +initialization: + critic_state_from: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/initialization/critic_seed0.pt +resume_policy: restart_only diff --git a/configs/S5/banc_descending/seed_1.yaml b/configs/S5/banc_descending/seed_1.yaml new file mode 100644 index 0000000000000000000000000000000000000000..b71b7f32f740e0db2484a1edd100df587e4aa8a4 --- /dev/null +++ b/configs/S5/banc_descending/seed_1.yaml @@ -0,0 +1,120 @@ +interface_hierarchy: + batch: EXP04_continuous_speed_matched6_20260921 + connectome: banc_descending + interface: command2 + task: speed_control + train_seed: 1 + smoke: false + label: banc_descending +vec_env: subproc +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - descending + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc + frozen_graph_path: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/graphs/banc_descending.pkl + frozen_graph_sha256: 86468ba0ad24d42cb8507741c05830da720573cf3df188574d6ffa1309aad915 +env: + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 + command_adapter_config: + frequency_bounds: + - 0.25 + - 2.0 + k_drive: 0.75 + k_drive_step: 1.0 + k_step: 0.75 + step_bounds: + - 0.4 + - 2.0 + continuous_gait: true + slow_duty_factor: 0.72 + fast_duty_factor: 0.58 + slow_ipsilateral_lag_cycles: 0.3333333333333333 + fast_ipsilateral_lag_cycles: 0.5 + swing_duration_s: 0.035 + base_frequency_hz: 12.0 + gait_smoothing_alpha: 0.005 + speed_filter_alpha: 0.25 + speed_heading_rate_weight: 0.25 + speed_target_scale_mm_s: 20.0 + speed_targets_mm_s: + - 8.0 + - 16.0 + action_mode: command2 + task: speed_control + speed_target_sampling: uniform + speed_target_range_mm_s: + - 8.0 + - 16.0 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 32 + seq_len: 32 + total_timesteps: 499200 + vf_clip_range: 0.2 + vf_coef: 0.5 + separate_rng_streams: true + initialization_seed: 1 + environment_seed: 11000 + policy_sampling_seed: 20001 + minibatch_seed: 30001 + initial_hidden_mode: zero +continuous_gait_experiment: + baseline: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed/configs/speed_continuous_gait_banc_descending_command2_seed0.yaml + target_sampling: uniform_episode + cpg_freedom: duty_warping_plus_phase_family + action_dimension_matched: true + primary_checkpoint: fixed_budget_final + training_unit: one complete fresh train seed; ER also independent graph seed +seed: 1 +initialization: + critic_state_from: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/initialization/critic_seed1.pt +resume_policy: restart_only diff --git a/configs/S5/banc_descending/seed_2.yaml b/configs/S5/banc_descending/seed_2.yaml new file mode 100644 index 0000000000000000000000000000000000000000..6d3f0556ed91716b0d26b4dee07a3342611037e1 --- /dev/null +++ b/configs/S5/banc_descending/seed_2.yaml @@ -0,0 +1,120 @@ +interface_hierarchy: + batch: EXP04_continuous_speed_matched6_20260921 + connectome: banc_descending + interface: command2 + task: speed_control + train_seed: 2 + smoke: false + label: banc_descending +vec_env: subproc +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - descending + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc + frozen_graph_path: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/graphs/banc_descending.pkl + frozen_graph_sha256: 86468ba0ad24d42cb8507741c05830da720573cf3df188574d6ffa1309aad915 +env: + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 + command_adapter_config: + frequency_bounds: + - 0.25 + - 2.0 + k_drive: 0.75 + k_drive_step: 1.0 + k_step: 0.75 + step_bounds: + - 0.4 + - 2.0 + continuous_gait: true + slow_duty_factor: 0.72 + fast_duty_factor: 0.58 + slow_ipsilateral_lag_cycles: 0.3333333333333333 + fast_ipsilateral_lag_cycles: 0.5 + swing_duration_s: 0.035 + base_frequency_hz: 12.0 + gait_smoothing_alpha: 0.005 + speed_filter_alpha: 0.25 + speed_heading_rate_weight: 0.25 + speed_target_scale_mm_s: 20.0 + speed_targets_mm_s: + - 8.0 + - 16.0 + action_mode: command2 + task: speed_control + speed_target_sampling: uniform + speed_target_range_mm_s: + - 8.0 + - 16.0 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 32 + seq_len: 32 + total_timesteps: 499200 + vf_clip_range: 0.2 + vf_coef: 0.5 + separate_rng_streams: true + initialization_seed: 2 + environment_seed: 12000 + policy_sampling_seed: 20002 + minibatch_seed: 30002 + initial_hidden_mode: zero +continuous_gait_experiment: + baseline: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed/configs/speed_continuous_gait_banc_descending_command2_seed0.yaml + target_sampling: uniform_episode + cpg_freedom: duty_warping_plus_phase_family + action_dimension_matched: true + primary_checkpoint: fixed_budget_final + training_unit: one complete fresh train seed; ER also independent graph seed +seed: 2 +initialization: + critic_state_from: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/initialization/critic_seed2.pt +resume_policy: restart_only diff --git a/configs/S5/banc_descending/seed_3.yaml b/configs/S5/banc_descending/seed_3.yaml new file mode 100644 index 0000000000000000000000000000000000000000..3159cf0ba6014852fe5725f7fced103e4134d59b --- /dev/null +++ b/configs/S5/banc_descending/seed_3.yaml @@ -0,0 +1,120 @@ +interface_hierarchy: + batch: EXP04_continuous_speed_matched6_20260921 + connectome: banc_descending + interface: command2 + task: speed_control + train_seed: 3 + smoke: false + label: banc_descending +vec_env: subproc +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - descending + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc + frozen_graph_path: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/graphs/banc_descending.pkl + frozen_graph_sha256: 86468ba0ad24d42cb8507741c05830da720573cf3df188574d6ffa1309aad915 +env: + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 + command_adapter_config: + frequency_bounds: + - 0.25 + - 2.0 + k_drive: 0.75 + k_drive_step: 1.0 + k_step: 0.75 + step_bounds: + - 0.4 + - 2.0 + continuous_gait: true + slow_duty_factor: 0.72 + fast_duty_factor: 0.58 + slow_ipsilateral_lag_cycles: 0.3333333333333333 + fast_ipsilateral_lag_cycles: 0.5 + swing_duration_s: 0.035 + base_frequency_hz: 12.0 + gait_smoothing_alpha: 0.005 + speed_filter_alpha: 0.25 + speed_heading_rate_weight: 0.25 + speed_target_scale_mm_s: 20.0 + speed_targets_mm_s: + - 8.0 + - 16.0 + action_mode: command2 + task: speed_control + speed_target_sampling: uniform + speed_target_range_mm_s: + - 8.0 + - 16.0 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 32 + seq_len: 32 + total_timesteps: 499200 + vf_clip_range: 0.2 + vf_coef: 0.5 + separate_rng_streams: true + initialization_seed: 3 + environment_seed: 13000 + policy_sampling_seed: 20003 + minibatch_seed: 30003 + initial_hidden_mode: zero +continuous_gait_experiment: + baseline: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed/configs/speed_continuous_gait_banc_descending_command2_seed0.yaml + target_sampling: uniform_episode + cpg_freedom: duty_warping_plus_phase_family + action_dimension_matched: true + primary_checkpoint: fixed_budget_final + training_unit: one complete fresh train seed; ER also independent graph seed +seed: 3 +initialization: + critic_state_from: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/initialization/critic_seed3.pt +resume_policy: restart_only diff --git a/configs/S5/banc_descending/seed_4.yaml b/configs/S5/banc_descending/seed_4.yaml new file mode 100644 index 0000000000000000000000000000000000000000..4cd01cacb8c4e1ad41f753771b0e8b8fdf01f40b --- /dev/null +++ b/configs/S5/banc_descending/seed_4.yaml @@ -0,0 +1,120 @@ +interface_hierarchy: + batch: EXP04_continuous_speed_matched6_20260921 + connectome: banc_descending + interface: command2 + task: speed_control + train_seed: 4 + smoke: false + label: banc_descending +vec_env: subproc +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - descending + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc + frozen_graph_path: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/graphs/banc_descending.pkl + frozen_graph_sha256: 86468ba0ad24d42cb8507741c05830da720573cf3df188574d6ffa1309aad915 +env: + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 + command_adapter_config: + frequency_bounds: + - 0.25 + - 2.0 + k_drive: 0.75 + k_drive_step: 1.0 + k_step: 0.75 + step_bounds: + - 0.4 + - 2.0 + continuous_gait: true + slow_duty_factor: 0.72 + fast_duty_factor: 0.58 + slow_ipsilateral_lag_cycles: 0.3333333333333333 + fast_ipsilateral_lag_cycles: 0.5 + swing_duration_s: 0.035 + base_frequency_hz: 12.0 + gait_smoothing_alpha: 0.005 + speed_filter_alpha: 0.25 + speed_heading_rate_weight: 0.25 + speed_target_scale_mm_s: 20.0 + speed_targets_mm_s: + - 8.0 + - 16.0 + action_mode: command2 + task: speed_control + speed_target_sampling: uniform + speed_target_range_mm_s: + - 8.0 + - 16.0 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 32 + seq_len: 32 + total_timesteps: 499200 + vf_clip_range: 0.2 + vf_coef: 0.5 + separate_rng_streams: true + initialization_seed: 4 + environment_seed: 14000 + policy_sampling_seed: 20004 + minibatch_seed: 30004 + initial_hidden_mode: zero +continuous_gait_experiment: + baseline: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed/configs/speed_continuous_gait_banc_descending_command2_seed0.yaml + target_sampling: uniform_episode + cpg_freedom: duty_warping_plus_phase_family + action_dimension_matched: true + primary_checkpoint: fixed_budget_final + training_unit: one complete fresh train seed; ER also independent graph seed +seed: 4 +initialization: + critic_state_from: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/initialization/critic_seed4.pt +resume_policy: restart_only diff --git a/configs/S5/banc_descending/seed_5.yaml b/configs/S5/banc_descending/seed_5.yaml new file mode 100644 index 0000000000000000000000000000000000000000..3490a8e8819e40449e48c3ba5ec8b31e20415fdb --- /dev/null +++ b/configs/S5/banc_descending/seed_5.yaml @@ -0,0 +1,120 @@ +interface_hierarchy: + batch: EXP04_continuous_speed_matched6_20260921 + connectome: banc_descending + interface: command2 + task: speed_control + train_seed: 5 + smoke: false + label: banc_descending +vec_env: subproc +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - descending + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc + frozen_graph_path: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/graphs/banc_descending.pkl + frozen_graph_sha256: 86468ba0ad24d42cb8507741c05830da720573cf3df188574d6ffa1309aad915 +env: + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 + command_adapter_config: + frequency_bounds: + - 0.25 + - 2.0 + k_drive: 0.75 + k_drive_step: 1.0 + k_step: 0.75 + step_bounds: + - 0.4 + - 2.0 + continuous_gait: true + slow_duty_factor: 0.72 + fast_duty_factor: 0.58 + slow_ipsilateral_lag_cycles: 0.3333333333333333 + fast_ipsilateral_lag_cycles: 0.5 + swing_duration_s: 0.035 + base_frequency_hz: 12.0 + gait_smoothing_alpha: 0.005 + speed_filter_alpha: 0.25 + speed_heading_rate_weight: 0.25 + speed_target_scale_mm_s: 20.0 + speed_targets_mm_s: + - 8.0 + - 16.0 + action_mode: command2 + task: speed_control + speed_target_sampling: uniform + speed_target_range_mm_s: + - 8.0 + - 16.0 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 32 + seq_len: 32 + total_timesteps: 499200 + vf_clip_range: 0.2 + vf_coef: 0.5 + separate_rng_streams: true + initialization_seed: 5 + environment_seed: 15000 + policy_sampling_seed: 20005 + minibatch_seed: 30005 + initial_hidden_mode: zero +continuous_gait_experiment: + baseline: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed/configs/speed_continuous_gait_banc_descending_command2_seed0.yaml + target_sampling: uniform_episode + cpg_freedom: duty_warping_plus_phase_family + action_dimension_matched: true + primary_checkpoint: fixed_budget_final + training_unit: one complete fresh train seed; ER also independent graph seed +seed: 5 +initialization: + critic_state_from: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/initialization/critic_seed5.pt +resume_policy: restart_only diff --git a/configs/S5/banc_motor/seed_0.yaml b/configs/S5/banc_motor/seed_0.yaml new file mode 100644 index 0000000000000000000000000000000000000000..b9d62f33a6d006ccda2796cfe1d4f1ee398d0541 --- /dev/null +++ b/configs/S5/banc_motor/seed_0.yaml @@ -0,0 +1,120 @@ +interface_hierarchy: + batch: EXP04_continuous_speed_matched6_20260921 + connectome: banc_motor + interface: command2 + task: speed_control + train_seed: 0 + smoke: false + label: banc_motor +vec_env: subproc +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - motor + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc + frozen_graph_path: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/graphs/banc_motor.pkl + frozen_graph_sha256: dde3a9e09030e301719ddf71c3eacc08bdaaeda009192310bcf0c39980859b59 +env: + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 + command_adapter_config: + frequency_bounds: + - 0.25 + - 2.0 + k_drive: 0.75 + k_drive_step: 1.0 + k_step: 0.75 + step_bounds: + - 0.4 + - 2.0 + continuous_gait: true + slow_duty_factor: 0.72 + fast_duty_factor: 0.58 + slow_ipsilateral_lag_cycles: 0.3333333333333333 + fast_ipsilateral_lag_cycles: 0.5 + swing_duration_s: 0.035 + base_frequency_hz: 12.0 + gait_smoothing_alpha: 0.005 + speed_filter_alpha: 0.25 + speed_heading_rate_weight: 0.25 + speed_target_scale_mm_s: 20.0 + speed_targets_mm_s: + - 8.0 + - 16.0 + action_mode: command2 + task: speed_control + speed_target_sampling: uniform + speed_target_range_mm_s: + - 8.0 + - 16.0 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 32 + seq_len: 32 + total_timesteps: 499200 + vf_clip_range: 0.2 + vf_coef: 0.5 + separate_rng_streams: true + initialization_seed: 0 + environment_seed: 10000 + policy_sampling_seed: 20000 + minibatch_seed: 30000 + initial_hidden_mode: zero +continuous_gait_experiment: + baseline: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed/configs/speed_continuous_gait_banc_motor_command2_seed0.yaml + target_sampling: uniform_episode + cpg_freedom: duty_warping_plus_phase_family + action_dimension_matched: true + primary_checkpoint: fixed_budget_final + training_unit: one complete fresh train seed; ER also independent graph seed +seed: 0 +initialization: + critic_state_from: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/initialization/critic_seed0.pt +resume_policy: restart_only diff --git a/configs/S5/banc_motor/seed_1.yaml b/configs/S5/banc_motor/seed_1.yaml new file mode 100644 index 0000000000000000000000000000000000000000..c76497416cc158f675468a7fb228baea29a8d202 --- /dev/null +++ b/configs/S5/banc_motor/seed_1.yaml @@ -0,0 +1,120 @@ +interface_hierarchy: + batch: EXP04_continuous_speed_matched6_20260921 + connectome: banc_motor + interface: command2 + task: speed_control + train_seed: 1 + smoke: false + label: banc_motor +vec_env: subproc +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - motor + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc + frozen_graph_path: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/graphs/banc_motor.pkl + frozen_graph_sha256: dde3a9e09030e301719ddf71c3eacc08bdaaeda009192310bcf0c39980859b59 +env: + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 + command_adapter_config: + frequency_bounds: + - 0.25 + - 2.0 + k_drive: 0.75 + k_drive_step: 1.0 + k_step: 0.75 + step_bounds: + - 0.4 + - 2.0 + continuous_gait: true + slow_duty_factor: 0.72 + fast_duty_factor: 0.58 + slow_ipsilateral_lag_cycles: 0.3333333333333333 + fast_ipsilateral_lag_cycles: 0.5 + swing_duration_s: 0.035 + base_frequency_hz: 12.0 + gait_smoothing_alpha: 0.005 + speed_filter_alpha: 0.25 + speed_heading_rate_weight: 0.25 + speed_target_scale_mm_s: 20.0 + speed_targets_mm_s: + - 8.0 + - 16.0 + action_mode: command2 + task: speed_control + speed_target_sampling: uniform + speed_target_range_mm_s: + - 8.0 + - 16.0 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 32 + seq_len: 32 + total_timesteps: 499200 + vf_clip_range: 0.2 + vf_coef: 0.5 + separate_rng_streams: true + initialization_seed: 1 + environment_seed: 11000 + policy_sampling_seed: 20001 + minibatch_seed: 30001 + initial_hidden_mode: zero +continuous_gait_experiment: + baseline: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed/configs/speed_continuous_gait_banc_motor_command2_seed0.yaml + target_sampling: uniform_episode + cpg_freedom: duty_warping_plus_phase_family + action_dimension_matched: true + primary_checkpoint: fixed_budget_final + training_unit: one complete fresh train seed; ER also independent graph seed +seed: 1 +initialization: + critic_state_from: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/initialization/critic_seed1.pt +resume_policy: restart_only diff --git a/configs/S5/banc_motor/seed_2.yaml b/configs/S5/banc_motor/seed_2.yaml new file mode 100644 index 0000000000000000000000000000000000000000..1bb9d5263bc8df708368670e4c9bead8754bd04f --- /dev/null +++ b/configs/S5/banc_motor/seed_2.yaml @@ -0,0 +1,120 @@ +interface_hierarchy: + batch: EXP04_continuous_speed_matched6_20260921 + connectome: banc_motor + interface: command2 + task: speed_control + train_seed: 2 + smoke: false + label: banc_motor +vec_env: subproc +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - motor + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc + frozen_graph_path: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/graphs/banc_motor.pkl + frozen_graph_sha256: dde3a9e09030e301719ddf71c3eacc08bdaaeda009192310bcf0c39980859b59 +env: + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 + command_adapter_config: + frequency_bounds: + - 0.25 + - 2.0 + k_drive: 0.75 + k_drive_step: 1.0 + k_step: 0.75 + step_bounds: + - 0.4 + - 2.0 + continuous_gait: true + slow_duty_factor: 0.72 + fast_duty_factor: 0.58 + slow_ipsilateral_lag_cycles: 0.3333333333333333 + fast_ipsilateral_lag_cycles: 0.5 + swing_duration_s: 0.035 + base_frequency_hz: 12.0 + gait_smoothing_alpha: 0.005 + speed_filter_alpha: 0.25 + speed_heading_rate_weight: 0.25 + speed_target_scale_mm_s: 20.0 + speed_targets_mm_s: + - 8.0 + - 16.0 + action_mode: command2 + task: speed_control + speed_target_sampling: uniform + speed_target_range_mm_s: + - 8.0 + - 16.0 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 32 + seq_len: 32 + total_timesteps: 499200 + vf_clip_range: 0.2 + vf_coef: 0.5 + separate_rng_streams: true + initialization_seed: 2 + environment_seed: 12000 + policy_sampling_seed: 20002 + minibatch_seed: 30002 + initial_hidden_mode: zero +continuous_gait_experiment: + baseline: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed/configs/speed_continuous_gait_banc_motor_command2_seed0.yaml + target_sampling: uniform_episode + cpg_freedom: duty_warping_plus_phase_family + action_dimension_matched: true + primary_checkpoint: fixed_budget_final + training_unit: one complete fresh train seed; ER also independent graph seed +seed: 2 +initialization: + critic_state_from: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/initialization/critic_seed2.pt +resume_policy: restart_only diff --git a/configs/S5/banc_motor/seed_3.yaml b/configs/S5/banc_motor/seed_3.yaml new file mode 100644 index 0000000000000000000000000000000000000000..874f0473b21f282abd8a781fe93dc01986e0c364 --- /dev/null +++ b/configs/S5/banc_motor/seed_3.yaml @@ -0,0 +1,120 @@ +interface_hierarchy: + batch: EXP04_continuous_speed_matched6_20260921 + connectome: banc_motor + interface: command2 + task: speed_control + train_seed: 3 + smoke: false + label: banc_motor +vec_env: subproc +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - motor + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc + frozen_graph_path: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/graphs/banc_motor.pkl + frozen_graph_sha256: dde3a9e09030e301719ddf71c3eacc08bdaaeda009192310bcf0c39980859b59 +env: + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 + command_adapter_config: + frequency_bounds: + - 0.25 + - 2.0 + k_drive: 0.75 + k_drive_step: 1.0 + k_step: 0.75 + step_bounds: + - 0.4 + - 2.0 + continuous_gait: true + slow_duty_factor: 0.72 + fast_duty_factor: 0.58 + slow_ipsilateral_lag_cycles: 0.3333333333333333 + fast_ipsilateral_lag_cycles: 0.5 + swing_duration_s: 0.035 + base_frequency_hz: 12.0 + gait_smoothing_alpha: 0.005 + speed_filter_alpha: 0.25 + speed_heading_rate_weight: 0.25 + speed_target_scale_mm_s: 20.0 + speed_targets_mm_s: + - 8.0 + - 16.0 + action_mode: command2 + task: speed_control + speed_target_sampling: uniform + speed_target_range_mm_s: + - 8.0 + - 16.0 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 32 + seq_len: 32 + total_timesteps: 499200 + vf_clip_range: 0.2 + vf_coef: 0.5 + separate_rng_streams: true + initialization_seed: 3 + environment_seed: 13000 + policy_sampling_seed: 20003 + minibatch_seed: 30003 + initial_hidden_mode: zero +continuous_gait_experiment: + baseline: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed/configs/speed_continuous_gait_banc_motor_command2_seed0.yaml + target_sampling: uniform_episode + cpg_freedom: duty_warping_plus_phase_family + action_dimension_matched: true + primary_checkpoint: fixed_budget_final + training_unit: one complete fresh train seed; ER also independent graph seed +seed: 3 +initialization: + critic_state_from: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/initialization/critic_seed3.pt +resume_policy: restart_only diff --git a/configs/S5/banc_motor/seed_4.yaml b/configs/S5/banc_motor/seed_4.yaml new file mode 100644 index 0000000000000000000000000000000000000000..9ac47f3384f179f3941b2415e6e1eaafc5477902 --- /dev/null +++ b/configs/S5/banc_motor/seed_4.yaml @@ -0,0 +1,120 @@ +interface_hierarchy: + batch: EXP04_continuous_speed_matched6_20260921 + connectome: banc_motor + interface: command2 + task: speed_control + train_seed: 4 + smoke: false + label: banc_motor +vec_env: subproc +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - motor + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc + frozen_graph_path: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/graphs/banc_motor.pkl + frozen_graph_sha256: dde3a9e09030e301719ddf71c3eacc08bdaaeda009192310bcf0c39980859b59 +env: + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 + command_adapter_config: + frequency_bounds: + - 0.25 + - 2.0 + k_drive: 0.75 + k_drive_step: 1.0 + k_step: 0.75 + step_bounds: + - 0.4 + - 2.0 + continuous_gait: true + slow_duty_factor: 0.72 + fast_duty_factor: 0.58 + slow_ipsilateral_lag_cycles: 0.3333333333333333 + fast_ipsilateral_lag_cycles: 0.5 + swing_duration_s: 0.035 + base_frequency_hz: 12.0 + gait_smoothing_alpha: 0.005 + speed_filter_alpha: 0.25 + speed_heading_rate_weight: 0.25 + speed_target_scale_mm_s: 20.0 + speed_targets_mm_s: + - 8.0 + - 16.0 + action_mode: command2 + task: speed_control + speed_target_sampling: uniform + speed_target_range_mm_s: + - 8.0 + - 16.0 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 32 + seq_len: 32 + total_timesteps: 499200 + vf_clip_range: 0.2 + vf_coef: 0.5 + separate_rng_streams: true + initialization_seed: 4 + environment_seed: 14000 + policy_sampling_seed: 20004 + minibatch_seed: 30004 + initial_hidden_mode: zero +continuous_gait_experiment: + baseline: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed/configs/speed_continuous_gait_banc_motor_command2_seed0.yaml + target_sampling: uniform_episode + cpg_freedom: duty_warping_plus_phase_family + action_dimension_matched: true + primary_checkpoint: fixed_budget_final + training_unit: one complete fresh train seed; ER also independent graph seed +seed: 4 +initialization: + critic_state_from: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/initialization/critic_seed4.pt +resume_policy: restart_only diff --git a/configs/S5/banc_motor/seed_5.yaml b/configs/S5/banc_motor/seed_5.yaml new file mode 100644 index 0000000000000000000000000000000000000000..f4b81bfd991f257257f1c3f6e8713e0c2a4a20bc --- /dev/null +++ b/configs/S5/banc_motor/seed_5.yaml @@ -0,0 +1,120 @@ +interface_hierarchy: + batch: EXP04_continuous_speed_matched6_20260921 + connectome: banc_motor + interface: command2 + task: speed_control + train_seed: 5 + smoke: false + label: banc_motor +vec_env: subproc +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - motor + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc + frozen_graph_path: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/graphs/banc_motor.pkl + frozen_graph_sha256: dde3a9e09030e301719ddf71c3eacc08bdaaeda009192310bcf0c39980859b59 +env: + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 + command_adapter_config: + frequency_bounds: + - 0.25 + - 2.0 + k_drive: 0.75 + k_drive_step: 1.0 + k_step: 0.75 + step_bounds: + - 0.4 + - 2.0 + continuous_gait: true + slow_duty_factor: 0.72 + fast_duty_factor: 0.58 + slow_ipsilateral_lag_cycles: 0.3333333333333333 + fast_ipsilateral_lag_cycles: 0.5 + swing_duration_s: 0.035 + base_frequency_hz: 12.0 + gait_smoothing_alpha: 0.005 + speed_filter_alpha: 0.25 + speed_heading_rate_weight: 0.25 + speed_target_scale_mm_s: 20.0 + speed_targets_mm_s: + - 8.0 + - 16.0 + action_mode: command2 + task: speed_control + speed_target_sampling: uniform + speed_target_range_mm_s: + - 8.0 + - 16.0 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 32 + seq_len: 32 + total_timesteps: 499200 + vf_clip_range: 0.2 + vf_coef: 0.5 + separate_rng_streams: true + initialization_seed: 5 + environment_seed: 15000 + policy_sampling_seed: 20005 + minibatch_seed: 30005 + initial_hidden_mode: zero +continuous_gait_experiment: + baseline: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed/configs/speed_continuous_gait_banc_motor_command2_seed0.yaml + target_sampling: uniform_episode + cpg_freedom: duty_warping_plus_phase_family + action_dimension_matched: true + primary_checkpoint: fixed_budget_final + training_unit: one complete fresh train seed; ER also independent graph seed +seed: 5 +initialization: + critic_state_from: /data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/initialization/critic_seed5.pt +resume_policy: restart_only diff --git a/configs/S6/banc_descending/seed_0.yaml b/configs/S6/banc_descending/seed_0.yaml new file mode 100644 index 0000000000000000000000000000000000000000..44dea192e7efeeb99f5a8d9316b8241752678bd5 --- /dev/null +++ b/configs/S6/banc_descending/seed_0.yaml @@ -0,0 +1,79 @@ +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - descending + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc +env: + action_mode: command2 + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + perturbation: + direction: lateral + duration_s: 0.04 + force: 10.0 + start_time_s: 0.75 + type: side_push + task: heading_hold + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 +interface_hierarchy: + batch: interface_hierarchy_original + connectome: banc_descending_original + interface: command2 + smoke: false + task: heading_hold + train_seed: 0 +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 8 + seq_len: 32 + total_timesteps: 500000 + vf_clip_range: 0.2 + vf_coef: 0.5 +seed: 0 +vec_env: subproc diff --git a/configs/S6/banc_descending/seed_1.yaml b/configs/S6/banc_descending/seed_1.yaml new file mode 100644 index 0000000000000000000000000000000000000000..4e165c14d7496d96334d505b1accd68012bb4d43 --- /dev/null +++ b/configs/S6/banc_descending/seed_1.yaml @@ -0,0 +1,79 @@ +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - descending + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc +env: + action_mode: command2 + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + perturbation: + direction: lateral + duration_s: 0.04 + force: 10.0 + start_time_s: 0.75 + type: side_push + task: heading_hold + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 +interface_hierarchy: + batch: interface_hierarchy_original + connectome: banc_descending_original + interface: command2 + smoke: false + task: heading_hold + train_seed: 1 +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 8 + seq_len: 32 + total_timesteps: 500000 + vf_clip_range: 0.2 + vf_coef: 0.5 +seed: 1 +vec_env: subproc diff --git a/configs/S6/banc_descending/seed_2.yaml b/configs/S6/banc_descending/seed_2.yaml new file mode 100644 index 0000000000000000000000000000000000000000..397b9bfc803a45432bf667856a156691d4721339 --- /dev/null +++ b/configs/S6/banc_descending/seed_2.yaml @@ -0,0 +1,79 @@ +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - descending + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc +env: + action_mode: command2 + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + perturbation: + direction: lateral + duration_s: 0.04 + force: 10.0 + start_time_s: 0.75 + type: side_push + task: heading_hold + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 +interface_hierarchy: + batch: interface_hierarchy_original + connectome: banc_descending_original + interface: command2 + smoke: false + task: heading_hold + train_seed: 2 +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 8 + seq_len: 32 + total_timesteps: 500000 + vf_clip_range: 0.2 + vf_coef: 0.5 +seed: 2 +vec_env: subproc diff --git a/configs/S6/banc_motor/seed_0.yaml b/configs/S6/banc_motor/seed_0.yaml new file mode 100644 index 0000000000000000000000000000000000000000..8a6be8c6ed641558e2c5fd445031de4f4866397a --- /dev/null +++ b/configs/S6/banc_motor/seed_0.yaml @@ -0,0 +1,79 @@ +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - motor + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc +env: + action_mode: command2 + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + perturbation: + direction: lateral + duration_s: 0.04 + force: 10.0 + start_time_s: 0.75 + type: side_push + task: heading_hold + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 +interface_hierarchy: + batch: interface_hierarchy_original + connectome: banc_original + interface: command2 + smoke: false + task: heading_hold + train_seed: 0 +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 8 + seq_len: 32 + total_timesteps: 500000 + vf_clip_range: 0.2 + vf_coef: 0.5 +seed: 0 +vec_env: subproc diff --git a/configs/S6/banc_motor/seed_1.yaml b/configs/S6/banc_motor/seed_1.yaml new file mode 100644 index 0000000000000000000000000000000000000000..3a94b76b727dd839ef18a823f748e7bedd772f95 --- /dev/null +++ b/configs/S6/banc_motor/seed_1.yaml @@ -0,0 +1,79 @@ +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - motor + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc +env: + action_mode: command2 + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + perturbation: + direction: lateral + duration_s: 0.04 + force: 10.0 + start_time_s: 0.75 + type: side_push + task: heading_hold + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 +interface_hierarchy: + batch: interface_hierarchy_original + connectome: banc_original + interface: command2 + smoke: false + task: heading_hold + train_seed: 1 +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 8 + seq_len: 32 + total_timesteps: 500000 + vf_clip_range: 0.2 + vf_coef: 0.5 +seed: 1 +vec_env: subproc diff --git a/configs/S6/banc_motor/seed_2.yaml b/configs/S6/banc_motor/seed_2.yaml new file mode 100644 index 0000000000000000000000000000000000000000..36c8080ccf8cac95e36bbb8ea2b3a6104895b546 --- /dev/null +++ b/configs/S6/banc_motor/seed_2.yaml @@ -0,0 +1,79 @@ +connectome: + banc_dir: larva/data/banc + chunksize: 250000 + global_gain: 1.0 + kind: original + min_syn_count: 1 + normalization: row_l1 + output_selector: + field: super_class + include: + - motor + polarity: + unknown_action: zero + randomize_seed: 0 + sensory_selector: + field: flow + include: + - afferent + species: banc +env: + action_mode: command2 + decision_interval: 100 + flip_penalty: 1000.0 + max_correction: 10.0 + max_time: 1.5 + observation_mode: proprio_task + perturbation: + direction: lateral + duration_s: 0.04 + force: 10.0 + start_time_s: 0.75 + type: side_push + task: heading_hold + terrain_seeds: + - 0 + - 1 + - 2 + - 3 + terrains: + - flat + timestep: 0.0001 +interface_hierarchy: + batch: interface_hierarchy_original + connectome: banc_original + interface: command2 + smoke: false + task: heading_hold + train_seed: 2 +policy: + actor_kwargs: + dt: 0.01 + input_target: sensory + mode: rate + tau_init: 0.02 + use_lora: false + log_std_init: -0.5 + name: connectome_cfc + value_hidden: + - 128 + - 128 +ppo: + clip_range: 0.2 + ent_coef: 0.0 + gamma: 0.99 + lam: 0.95 + lr: 0.0001 + max_grad_norm: 0.5 + n_envs: 16 + n_epochs: 2 + n_minibatches: 5 + n_steps: 160 + normalize_reward: true + rollout_heartbeat: 8 + seq_len: 32 + total_timesteps: 500000 + vf_clip_range: 0.2 + vf_coef: 0.5 +seed: 2 +vec_env: subproc diff --git a/configs/figure2/connectome/seed_2011.json b/configs/figure2/connectome/seed_2011.json new file mode 100644 index 0000000000000000000000000000000000000000..af19bd5af331c98afbb2e792bf58f9bdb2f77c99 --- /dev/null +++ b/configs/figure2/connectome/seed_2011.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2011, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2011, + "separate_rng_streams": true, + "initialization_seed": 2011, + "environment_seed": 5201100, + "policy_sampling_seed": 102011, + "minibatch_seed": 302011, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "663b9123cb0942fd2bbd98a8da7aaa14968116780a24042a28f6cb213dba4302", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2012.json b/configs/figure2/connectome/seed_2012.json new file mode 100644 index 0000000000000000000000000000000000000000..a971924de4bf0e39eb6be41501a832ffeb04cb38 --- /dev/null +++ b/configs/figure2/connectome/seed_2012.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2012, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2012, + "separate_rng_streams": true, + "initialization_seed": 2012, + "environment_seed": 5201200, + "policy_sampling_seed": 102012, + "minibatch_seed": 302012, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "27e7d4589ae2a02d49168ee3130056bc39e7d5acf80c93c5c78bec08901a3270", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2013.json b/configs/figure2/connectome/seed_2013.json new file mode 100644 index 0000000000000000000000000000000000000000..e42498981011d01b9b99387f012406a3f0401f17 --- /dev/null +++ b/configs/figure2/connectome/seed_2013.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2013, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2013, + "separate_rng_streams": true, + "initialization_seed": 2013, + "environment_seed": 5201300, + "policy_sampling_seed": 102013, + "minibatch_seed": 302013, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "e005816168be1f9a68a4e86d709e31c4746a591ad27106f580f1a8ac8146188d", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2014.json b/configs/figure2/connectome/seed_2014.json new file mode 100644 index 0000000000000000000000000000000000000000..34f846c2bbc45c47a075afbb4afb060dc66db616 --- /dev/null +++ b/configs/figure2/connectome/seed_2014.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2014, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2014, + "separate_rng_streams": true, + "initialization_seed": 2014, + "environment_seed": 5201400, + "policy_sampling_seed": 102014, + "minibatch_seed": 302014, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "1d01c86fdf54cee2689893193bad53dccd5ddc7662e3041d705c57aae05db9cf", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2015.json b/configs/figure2/connectome/seed_2015.json new file mode 100644 index 0000000000000000000000000000000000000000..8dbb1522bbefa99dbdff61b6042fb6e52a2cf83f --- /dev/null +++ b/configs/figure2/connectome/seed_2015.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2015, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2015, + "separate_rng_streams": true, + "initialization_seed": 2015, + "environment_seed": 5201500, + "policy_sampling_seed": 102015, + "minibatch_seed": 302015, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "0f0f10c0bb695923b99f909614ff2b05cf2d426a5fc8489cae9cba88bc4128a0", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2016.json b/configs/figure2/connectome/seed_2016.json new file mode 100644 index 0000000000000000000000000000000000000000..235205b3ba613824741a0376a1dbb11677d70a98 --- /dev/null +++ b/configs/figure2/connectome/seed_2016.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2016, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2016, + "separate_rng_streams": true, + "initialization_seed": 2016, + "environment_seed": 5201600, + "policy_sampling_seed": 102016, + "minibatch_seed": 302016, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "e7ac33a797fccb430fc12d7dc419a307a4bbe8148107f9394d6c324e4502e3cb", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2017.json b/configs/figure2/connectome/seed_2017.json new file mode 100644 index 0000000000000000000000000000000000000000..32bab2355c60726fc58e1118a419956a79b5578d --- /dev/null +++ b/configs/figure2/connectome/seed_2017.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2017, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2017, + "separate_rng_streams": true, + "initialization_seed": 2017, + "environment_seed": 5201700, + "policy_sampling_seed": 102017, + "minibatch_seed": 302017, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "af0e370c98101dae66fc3ae5f66047d573fb767917e2d74319069cc75f82ca6b", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2018.json b/configs/figure2/connectome/seed_2018.json new file mode 100644 index 0000000000000000000000000000000000000000..9f5d76bd7292c8aff736c4d3fa3c775d865e9cbb --- /dev/null +++ b/configs/figure2/connectome/seed_2018.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2018, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2018, + "separate_rng_streams": true, + "initialization_seed": 2018, + "environment_seed": 5201800, + "policy_sampling_seed": 102018, + "minibatch_seed": 302018, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "473c2abd5dbf0604b19594a05671430bbbea667f02e8df4e3c292cd0ea02f089", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2019.json b/configs/figure2/connectome/seed_2019.json new file mode 100644 index 0000000000000000000000000000000000000000..2a891a103fc0095d4049ff3db95542bbe9c96dec --- /dev/null +++ b/configs/figure2/connectome/seed_2019.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2019, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2019, + "separate_rng_streams": true, + "initialization_seed": 2019, + "environment_seed": 5201900, + "policy_sampling_seed": 102019, + "minibatch_seed": 302019, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "987b4cf36337646a2ede352a26900de76ac49af659af6519caa7809c692d1c4f", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2020.json b/configs/figure2/connectome/seed_2020.json new file mode 100644 index 0000000000000000000000000000000000000000..a60517794cba7d717f02360143c2fbf5d4a030ba --- /dev/null +++ b/configs/figure2/connectome/seed_2020.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2020, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2020, + "separate_rng_streams": true, + "initialization_seed": 2020, + "environment_seed": 5202000, + "policy_sampling_seed": 102020, + "minibatch_seed": 302020, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "9d7ce49cb0c96ed0ec9af1e5b73c9f93c7ee35a74182a249c5fdcc14375058ca", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2021.json b/configs/figure2/connectome/seed_2021.json new file mode 100644 index 0000000000000000000000000000000000000000..eb7d98ab2d30ee09312e18ae5b492ccec88c23ad --- /dev/null +++ b/configs/figure2/connectome/seed_2021.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2021, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2021, + "separate_rng_streams": true, + "initialization_seed": 2021, + "environment_seed": 5202100, + "policy_sampling_seed": 102021, + "minibatch_seed": 302021, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "a0121d8d6fb53980e9a2e0056a70efa9405e7894de5b84d34340b03b9b936cfe", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2022.json b/configs/figure2/connectome/seed_2022.json new file mode 100644 index 0000000000000000000000000000000000000000..18ce85fffbb6d776d03ee36a272b21c406bb0453 --- /dev/null +++ b/configs/figure2/connectome/seed_2022.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2022, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2022, + "separate_rng_streams": true, + "initialization_seed": 2022, + "environment_seed": 5202200, + "policy_sampling_seed": 102022, + "minibatch_seed": 302022, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "5cf8fe449402a9217e565ce53111bc58b2e6a2f3a8df2ebe904ec9afbd81665e", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2023.json b/configs/figure2/connectome/seed_2023.json new file mode 100644 index 0000000000000000000000000000000000000000..c1ed411c21bcb0acd3362af5ef5f45e449637daa --- /dev/null +++ b/configs/figure2/connectome/seed_2023.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2023, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2023, + "separate_rng_streams": true, + "initialization_seed": 2023, + "environment_seed": 5202300, + "policy_sampling_seed": 102023, + "minibatch_seed": 302023, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "40b3c46e8e20197007f9d394bb9a282d707df23f73e8aede75991084df8dd5e2", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2024.json b/configs/figure2/connectome/seed_2024.json new file mode 100644 index 0000000000000000000000000000000000000000..b2c025b3f3f81a316dc9e8eedfc23dd4fc2fd581 --- /dev/null +++ b/configs/figure2/connectome/seed_2024.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2024, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2024, + "separate_rng_streams": true, + "initialization_seed": 2024, + "environment_seed": 5202400, + "policy_sampling_seed": 102024, + "minibatch_seed": 302024, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "d3d564c0b8901fee3bdfccdce8fb678b4bda7ee8f7c9454ca2601eee8d8a8048", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2025.json b/configs/figure2/connectome/seed_2025.json new file mode 100644 index 0000000000000000000000000000000000000000..2a4f06b5ee6ccca2c7266aa22b91c55700a3c547 --- /dev/null +++ b/configs/figure2/connectome/seed_2025.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2025, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2025, + "separate_rng_streams": true, + "initialization_seed": 2025, + "environment_seed": 5202500, + "policy_sampling_seed": 102025, + "minibatch_seed": 302025, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "8e8ef57e505352aadb496afc69ff08ee15caf9e52cb82f6b7773aeff701e4617", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2026.json b/configs/figure2/connectome/seed_2026.json new file mode 100644 index 0000000000000000000000000000000000000000..a59b7dfc64041033ca17cb51ac6d2aaf30df06cb --- /dev/null +++ b/configs/figure2/connectome/seed_2026.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2026, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2026, + "separate_rng_streams": true, + "initialization_seed": 2026, + "environment_seed": 5202600, + "policy_sampling_seed": 102026, + "minibatch_seed": 302026, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "014bc75abbb7047ae99c521fd34b3179c0a5462b31f02ec6a15a06da44e9459f", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2027.json b/configs/figure2/connectome/seed_2027.json new file mode 100644 index 0000000000000000000000000000000000000000..97bb40172a608e86de7fbedeb561d08209a5c0fe --- /dev/null +++ b/configs/figure2/connectome/seed_2027.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2027, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2027, + "separate_rng_streams": true, + "initialization_seed": 2027, + "environment_seed": 5202700, + "policy_sampling_seed": 102027, + "minibatch_seed": 302027, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "4c4ad13c44d5d4df2a7347d2f954755c3f76955e122bd735a2967543f74a374f", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2028.json b/configs/figure2/connectome/seed_2028.json new file mode 100644 index 0000000000000000000000000000000000000000..a2e8738aaaff646dd5376a6bd151806d2ffdc5ff --- /dev/null +++ b/configs/figure2/connectome/seed_2028.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2028, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2028, + "separate_rng_streams": true, + "initialization_seed": 2028, + "environment_seed": 5202800, + "policy_sampling_seed": 102028, + "minibatch_seed": 302028, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "fee737e447bd4ab6542e53d21fe271887f6a88090634decd0b8925d253a15d4e", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2029.json b/configs/figure2/connectome/seed_2029.json new file mode 100644 index 0000000000000000000000000000000000000000..237522390f4c7b3465614b5004e77d08ea7e1b00 --- /dev/null +++ b/configs/figure2/connectome/seed_2029.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2029, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2029, + "separate_rng_streams": true, + "initialization_seed": 2029, + "environment_seed": 5202900, + "policy_sampling_seed": 102029, + "minibatch_seed": 302029, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "125c5cbcab296c4cd98a6af032c440e8eb82a722c43a6165045e00c0f4b601a8", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/connectome/seed_2030.json b/configs/figure2/connectome/seed_2030.json new file mode 100644 index 0000000000000000000000000000000000000000..8003fa11b38aa8a147e47d02d96a46669c483ddd --- /dev/null +++ b/configs/figure2/connectome/seed_2030.json @@ -0,0 +1,61 @@ +{ + "method": "connectome", + "seed": 2030, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2030, + "separate_rng_streams": true, + "initialization_seed": 2030, + "environment_seed": 5203000, + "policy_sampling_seed": 102030, + "minibatch_seed": 302030, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "8f953b7060aefdf6c736b37f2608446408210afc49072c99e0617134e0ccc0a3", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": "defe28edec2e000943bd7709e3439881e58a1e42bb4d9b870025e82826807ca7", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2011.json b/configs/figure2/er/seed_2011.json new file mode 100644 index 0000000000000000000000000000000000000000..bc60bee5e481bdc95cfda82002b3a80db36778c1 --- /dev/null +++ b/configs/figure2/er/seed_2011.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2011, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2011, + "separate_rng_streams": true, + "initialization_seed": 2011, + "environment_seed": 5201100, + "policy_sampling_seed": 102011, + "minibatch_seed": 302011, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "663b9123cb0942fd2bbd98a8da7aaa14968116780a24042a28f6cb213dba4302", + "graph_sha256": "af7bb7c42684e85f813d8fb53357cc5471a447f6b7088092822f4fc9ef4fb492", + "graph_buffers_sha256": "b42a1725f17c3740ce5a3b7db45688aca856797d05a1346a1917e7776d5080d2", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2012.json b/configs/figure2/er/seed_2012.json new file mode 100644 index 0000000000000000000000000000000000000000..c8998bd534b0c7bdd03c236904d7773c367af572 --- /dev/null +++ b/configs/figure2/er/seed_2012.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2012, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2012, + "separate_rng_streams": true, + "initialization_seed": 2012, + "environment_seed": 5201200, + "policy_sampling_seed": 102012, + "minibatch_seed": 302012, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "27e7d4589ae2a02d49168ee3130056bc39e7d5acf80c93c5c78bec08901a3270", + "graph_sha256": "1e31e4c7fb4a3d4e626e209f87c3bb84a5beae4b64c3b2db341769debe5c7cb0", + "graph_buffers_sha256": "61699a00c20da810d8be9c72eab5aabd5174da9e8a3a88e389fb50856e28ab4f", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2013.json b/configs/figure2/er/seed_2013.json new file mode 100644 index 0000000000000000000000000000000000000000..21242bbfa6a60ae96a4d46e46379929b5b20b315 --- /dev/null +++ b/configs/figure2/er/seed_2013.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2013, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2013, + "separate_rng_streams": true, + "initialization_seed": 2013, + "environment_seed": 5201300, + "policy_sampling_seed": 102013, + "minibatch_seed": 302013, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "e005816168be1f9a68a4e86d709e31c4746a591ad27106f580f1a8ac8146188d", + "graph_sha256": "8e512a5165514a154f3dc4e00b22584a515dabc02a84f6cadc53979aaf509ee1", + "graph_buffers_sha256": "501af264917a4ddefe0df64b2e2126ea0ac29385b218a3601bde830436cd990d", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2014.json b/configs/figure2/er/seed_2014.json new file mode 100644 index 0000000000000000000000000000000000000000..4392dc9e95e392e7aac6d1d61506b1fc971592ad --- /dev/null +++ b/configs/figure2/er/seed_2014.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2014, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2014, + "separate_rng_streams": true, + "initialization_seed": 2014, + "environment_seed": 5201400, + "policy_sampling_seed": 102014, + "minibatch_seed": 302014, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "1d01c86fdf54cee2689893193bad53dccd5ddc7662e3041d705c57aae05db9cf", + "graph_sha256": "721c7528d0969ad088566aa81a59658e5b0617ecb749a503ef14ebf686b259f4", + "graph_buffers_sha256": "8f061f242dab13ed8e9bfc61c7992dcfb102c563ce376c117e5be70675d948a9", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2015.json b/configs/figure2/er/seed_2015.json new file mode 100644 index 0000000000000000000000000000000000000000..3b232a03de6d2af97297352a6b38c944b301a8ce --- /dev/null +++ b/configs/figure2/er/seed_2015.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2015, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2015, + "separate_rng_streams": true, + "initialization_seed": 2015, + "environment_seed": 5201500, + "policy_sampling_seed": 102015, + "minibatch_seed": 302015, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "0f0f10c0bb695923b99f909614ff2b05cf2d426a5fc8489cae9cba88bc4128a0", + "graph_sha256": "0dd14788c19f7e4c007d2d0d900c499ffaf494747d15f6e4d5454c074aa7bff9", + "graph_buffers_sha256": "4677e7f045eb43036acd3cd54ccb2b4bb5a5137182168aee99260fe3276b63b2", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2016.json b/configs/figure2/er/seed_2016.json new file mode 100644 index 0000000000000000000000000000000000000000..2c09b7ad6315f8bf6f4aa7bd7e2b792379e766c4 --- /dev/null +++ b/configs/figure2/er/seed_2016.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2016, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2016, + "separate_rng_streams": true, + "initialization_seed": 2016, + "environment_seed": 5201600, + "policy_sampling_seed": 102016, + "minibatch_seed": 302016, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "e7ac33a797fccb430fc12d7dc419a307a4bbe8148107f9394d6c324e4502e3cb", + "graph_sha256": "e03df681a124f88b006169542334357aa3bafaf6b570cb765467a9769c93121b", + "graph_buffers_sha256": "dea28f4414b185e078e4a39930daae96480b1b96f367e5aba3d50e167eea1c1d", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2017.json b/configs/figure2/er/seed_2017.json new file mode 100644 index 0000000000000000000000000000000000000000..afa1b60545a228e99746b59095fd8be7f0d2ebb7 --- /dev/null +++ b/configs/figure2/er/seed_2017.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2017, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2017, + "separate_rng_streams": true, + "initialization_seed": 2017, + "environment_seed": 5201700, + "policy_sampling_seed": 102017, + "minibatch_seed": 302017, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "af0e370c98101dae66fc3ae5f66047d573fb767917e2d74319069cc75f82ca6b", + "graph_sha256": "8d89b097ea5df04a3cc90541f54f226b50ab64c457511a81c58bbe0eff0c432b", + "graph_buffers_sha256": "8fa55cb252067cda2d4212fb99731f15bd7817b762e2ea488f1158bb7a7e4d97", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2018.json b/configs/figure2/er/seed_2018.json new file mode 100644 index 0000000000000000000000000000000000000000..dc202571472f402ab47ab212011391e91df2f516 --- /dev/null +++ b/configs/figure2/er/seed_2018.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2018, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2018, + "separate_rng_streams": true, + "initialization_seed": 2018, + "environment_seed": 5201800, + "policy_sampling_seed": 102018, + "minibatch_seed": 302018, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "473c2abd5dbf0604b19594a05671430bbbea667f02e8df4e3c292cd0ea02f089", + "graph_sha256": "028b562392a4bceae6a679aaeb7aae26241adbba8fb587b2305f6da6d4c24a62", + "graph_buffers_sha256": "dc95cc3fb1c3a7c7a97eb9a489ef4784605ad56f2fe028fcb9585cc36558c8ce", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2019.json b/configs/figure2/er/seed_2019.json new file mode 100644 index 0000000000000000000000000000000000000000..754d325fea60e50e692e908cf3065be9547ba86b --- /dev/null +++ b/configs/figure2/er/seed_2019.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2019, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2019, + "separate_rng_streams": true, + "initialization_seed": 2019, + "environment_seed": 5201900, + "policy_sampling_seed": 102019, + "minibatch_seed": 302019, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "987b4cf36337646a2ede352a26900de76ac49af659af6519caa7809c692d1c4f", + "graph_sha256": "a17874dac6acffbbd0e6713b12a795b58fb7f01ffe57ea36955d3a7a24877e11", + "graph_buffers_sha256": "1042e9bc3309815321fbb82770280469c6aea80f335f36b6c593601eb79151f0", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2020.json b/configs/figure2/er/seed_2020.json new file mode 100644 index 0000000000000000000000000000000000000000..383af6d1fe4f00035c543aed26ba7b70caa02a0c --- /dev/null +++ b/configs/figure2/er/seed_2020.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2020, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2020, + "separate_rng_streams": true, + "initialization_seed": 2020, + "environment_seed": 5202000, + "policy_sampling_seed": 102020, + "minibatch_seed": 302020, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "9d7ce49cb0c96ed0ec9af1e5b73c9f93c7ee35a74182a249c5fdcc14375058ca", + "graph_sha256": "883b6a156e574babecdc4bc179651247d0faa6d3fee1f9d8236b6c099e94f6b2", + "graph_buffers_sha256": "e85e353744dba32fe3458d16832d657279d5c9e284a680b6e7f35a6cba84e952", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2021.json b/configs/figure2/er/seed_2021.json new file mode 100644 index 0000000000000000000000000000000000000000..dedc56558d43f3d11fe60d239b8ea1805c75dcb7 --- /dev/null +++ b/configs/figure2/er/seed_2021.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2021, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2021, + "separate_rng_streams": true, + "initialization_seed": 2021, + "environment_seed": 5202100, + "policy_sampling_seed": 102021, + "minibatch_seed": 302021, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "a0121d8d6fb53980e9a2e0056a70efa9405e7894de5b84d34340b03b9b936cfe", + "graph_sha256": "c338fddd5096b50f89f224a2adb082fbe1200040c10956606ad2ed9c89129f06", + "graph_buffers_sha256": "9d8871a98e8c5b3e5d749880b2f8f4ccf94e20d0668d65e7949f4a7132528997", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2022.json b/configs/figure2/er/seed_2022.json new file mode 100644 index 0000000000000000000000000000000000000000..35464444756f244902c9a9c1e9ae6e85efac2921 --- /dev/null +++ b/configs/figure2/er/seed_2022.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2022, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2022, + "separate_rng_streams": true, + "initialization_seed": 2022, + "environment_seed": 5202200, + "policy_sampling_seed": 102022, + "minibatch_seed": 302022, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "5cf8fe449402a9217e565ce53111bc58b2e6a2f3a8df2ebe904ec9afbd81665e", + "graph_sha256": "30233d35b8f6828ade0b3e7595bd978b571b43dd8eaaa0342643fb43d7efb791", + "graph_buffers_sha256": "4833a5b9d6ebf93bba5fcc029854bd2683b895e83fd732af534a47b650a55550", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2023.json b/configs/figure2/er/seed_2023.json new file mode 100644 index 0000000000000000000000000000000000000000..3b5df38eeb49f722d962e25cada7b34f54cdadac --- /dev/null +++ b/configs/figure2/er/seed_2023.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2023, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2023, + "separate_rng_streams": true, + "initialization_seed": 2023, + "environment_seed": 5202300, + "policy_sampling_seed": 102023, + "minibatch_seed": 302023, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "40b3c46e8e20197007f9d394bb9a282d707df23f73e8aede75991084df8dd5e2", + "graph_sha256": "f10f9ed7a58a7ef08ed19d2ad218b7cc2e9420edb92285490b43cdc997f65489", + "graph_buffers_sha256": "545255dcace7098c6a6c704a8ed4807780bd260cb0b19cb2e0e1065090064fed", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2024.json b/configs/figure2/er/seed_2024.json new file mode 100644 index 0000000000000000000000000000000000000000..3d314804370be1cf8d469202c99d99d5010ff2a0 --- /dev/null +++ b/configs/figure2/er/seed_2024.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2024, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2024, + "separate_rng_streams": true, + "initialization_seed": 2024, + "environment_seed": 5202400, + "policy_sampling_seed": 102024, + "minibatch_seed": 302024, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "d3d564c0b8901fee3bdfccdce8fb678b4bda7ee8f7c9454ca2601eee8d8a8048", + "graph_sha256": "bb5a600412f288a765328c03ce8f3cefa3279b89766e890453c29df4cd25747e", + "graph_buffers_sha256": "6e526035e1e1b2065219e882a5479808db3e14e40a39f4be903cea92bcdd5a12", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2025.json b/configs/figure2/er/seed_2025.json new file mode 100644 index 0000000000000000000000000000000000000000..94ff95df52f25c8ad6225b807b8ca9fa76da7dd2 --- /dev/null +++ b/configs/figure2/er/seed_2025.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2025, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2025, + "separate_rng_streams": true, + "initialization_seed": 2025, + "environment_seed": 5202500, + "policy_sampling_seed": 102025, + "minibatch_seed": 302025, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "8e8ef57e505352aadb496afc69ff08ee15caf9e52cb82f6b7773aeff701e4617", + "graph_sha256": "2016c3a6231485ad425c672ba5f01f42138ac9f7826e387dfd1f311f99c1b52f", + "graph_buffers_sha256": "fafac654da995677ba48cc43359cfe242818858891c429cfc34190e492094070", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2026.json b/configs/figure2/er/seed_2026.json new file mode 100644 index 0000000000000000000000000000000000000000..fab440585bed90c89e40878740e76666d47b6663 --- /dev/null +++ b/configs/figure2/er/seed_2026.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2026, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2026, + "separate_rng_streams": true, + "initialization_seed": 2026, + "environment_seed": 5202600, + "policy_sampling_seed": 102026, + "minibatch_seed": 302026, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "014bc75abbb7047ae99c521fd34b3179c0a5462b31f02ec6a15a06da44e9459f", + "graph_sha256": "0579494d7c0d70d680f4208328a58aa7e66b609889631df8a17f062ef5c19b08", + "graph_buffers_sha256": "3ad6c49fd3e925c582ee73c412abe10fc7cc10f22ebb586c104b631bc48874fe", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2027.json b/configs/figure2/er/seed_2027.json new file mode 100644 index 0000000000000000000000000000000000000000..8e3274b024bb77ac97c2524483ab09e30ec22a2e --- /dev/null +++ b/configs/figure2/er/seed_2027.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2027, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2027, + "separate_rng_streams": true, + "initialization_seed": 2027, + "environment_seed": 5202700, + "policy_sampling_seed": 102027, + "minibatch_seed": 302027, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "4c4ad13c44d5d4df2a7347d2f954755c3f76955e122bd735a2967543f74a374f", + "graph_sha256": "dfa77d8b8485258061214d9ffd88d5c4ebdb85ea35202969806e6cced1207a08", + "graph_buffers_sha256": "787df62c1b7582654d815274272886fa1ff08ad205c164e0915f9865e953a1ac", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2028.json b/configs/figure2/er/seed_2028.json new file mode 100644 index 0000000000000000000000000000000000000000..42fd85170b19cdfd05f3166d2bc81a19c4f5523c --- /dev/null +++ b/configs/figure2/er/seed_2028.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2028, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2028, + "separate_rng_streams": true, + "initialization_seed": 2028, + "environment_seed": 5202800, + "policy_sampling_seed": 102028, + "minibatch_seed": 302028, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "fee737e447bd4ab6542e53d21fe271887f6a88090634decd0b8925d253a15d4e", + "graph_sha256": "a73f9f18bbd89df49d65c85bfab8006f073eb33097f5e09bbd21b981105e8269", + "graph_buffers_sha256": "dc5ffa2f3379e3122e2e842344d2ff77749e6247b6a4f29267298b13d9266545", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2029.json b/configs/figure2/er/seed_2029.json new file mode 100644 index 0000000000000000000000000000000000000000..4b946e81b5dec6871e190bb96e78d3dd10915a01 --- /dev/null +++ b/configs/figure2/er/seed_2029.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2029, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2029, + "separate_rng_streams": true, + "initialization_seed": 2029, + "environment_seed": 5202900, + "policy_sampling_seed": 102029, + "minibatch_seed": 302029, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "125c5cbcab296c4cd98a6af032c440e8eb82a722c43a6165045e00c0f4b601a8", + "graph_sha256": "b6b0c3912ce536e5535f876b522d7a7e0fbf71496277d8d243dd5acc60a41c2e", + "graph_buffers_sha256": "5c2b69063542846f6573e2ad37492cfe2d6c357f25af34c5ceb7ebe870510342", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/er/seed_2030.json b/configs/figure2/er/seed_2030.json new file mode 100644 index 0000000000000000000000000000000000000000..e76960c601b20513ff678dacdab579ef033daad9 --- /dev/null +++ b/configs/figure2/er/seed_2030.json @@ -0,0 +1,61 @@ +{ + "method": "er", + "seed": 2030, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2030, + "separate_rng_streams": true, + "initialization_seed": 2030, + "environment_seed": 5203000, + "policy_sampling_seed": 102030, + "minibatch_seed": 302030, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404593, + "critic_parameter_count": 242433, + "critic_sha256": "8f953b7060aefdf6c736b37f2608446408210afc49072c99e0617134e0ccc0a3", + "graph_sha256": "f85a4fe5f42a4d858bcdaae2f3dd58a1fd3f5aa25ea71aab0709ce380a2a2760", + "graph_buffers_sha256": "d9db54d18d35b8ecb9335c08cf06a3ddaf77a3c0605b6bea340dc8086d00e472", + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2011.json b/configs/figure2/mlp/seed_2011.json new file mode 100644 index 0000000000000000000000000000000000000000..68517efca84d707e669da828c13db45f926a1386 --- /dev/null +++ b/configs/figure2/mlp/seed_2011.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2011, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2011, + "separate_rng_streams": true, + "initialization_seed": 2011, + "environment_seed": 5201100, + "policy_sampling_seed": 102011, + "minibatch_seed": 302011, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "663b9123cb0942fd2bbd98a8da7aaa14968116780a24042a28f6cb213dba4302", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2012.json b/configs/figure2/mlp/seed_2012.json new file mode 100644 index 0000000000000000000000000000000000000000..e46083a7d25748265ce905d96e8850e255a76c83 --- /dev/null +++ b/configs/figure2/mlp/seed_2012.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2012, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2012, + "separate_rng_streams": true, + "initialization_seed": 2012, + "environment_seed": 5201200, + "policy_sampling_seed": 102012, + "minibatch_seed": 302012, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "27e7d4589ae2a02d49168ee3130056bc39e7d5acf80c93c5c78bec08901a3270", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2013.json b/configs/figure2/mlp/seed_2013.json new file mode 100644 index 0000000000000000000000000000000000000000..0cfea5beffa2adf61074631c8de02d7c132a4173 --- /dev/null +++ b/configs/figure2/mlp/seed_2013.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2013, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2013, + "separate_rng_streams": true, + "initialization_seed": 2013, + "environment_seed": 5201300, + "policy_sampling_seed": 102013, + "minibatch_seed": 302013, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "e005816168be1f9a68a4e86d709e31c4746a591ad27106f580f1a8ac8146188d", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2014.json b/configs/figure2/mlp/seed_2014.json new file mode 100644 index 0000000000000000000000000000000000000000..29fec68aa83c4c5bdd2bc61562a6a97e36c765ce --- /dev/null +++ b/configs/figure2/mlp/seed_2014.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2014, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2014, + "separate_rng_streams": true, + "initialization_seed": 2014, + "environment_seed": 5201400, + "policy_sampling_seed": 102014, + "minibatch_seed": 302014, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "1d01c86fdf54cee2689893193bad53dccd5ddc7662e3041d705c57aae05db9cf", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2015.json b/configs/figure2/mlp/seed_2015.json new file mode 100644 index 0000000000000000000000000000000000000000..8878f7cc4ba3ccbf89376888069f237afd0245e4 --- /dev/null +++ b/configs/figure2/mlp/seed_2015.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2015, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2015, + "separate_rng_streams": true, + "initialization_seed": 2015, + "environment_seed": 5201500, + "policy_sampling_seed": 102015, + "minibatch_seed": 302015, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "0f0f10c0bb695923b99f909614ff2b05cf2d426a5fc8489cae9cba88bc4128a0", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2016.json b/configs/figure2/mlp/seed_2016.json new file mode 100644 index 0000000000000000000000000000000000000000..5fa680ff770fd521074460a96dd6f5a7e5591397 --- /dev/null +++ b/configs/figure2/mlp/seed_2016.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2016, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2016, + "separate_rng_streams": true, + "initialization_seed": 2016, + "environment_seed": 5201600, + "policy_sampling_seed": 102016, + "minibatch_seed": 302016, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "e7ac33a797fccb430fc12d7dc419a307a4bbe8148107f9394d6c324e4502e3cb", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2017.json b/configs/figure2/mlp/seed_2017.json new file mode 100644 index 0000000000000000000000000000000000000000..b402bd89c734b1ba3181f5aa933f200b98c310df --- /dev/null +++ b/configs/figure2/mlp/seed_2017.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2017, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2017, + "separate_rng_streams": true, + "initialization_seed": 2017, + "environment_seed": 5201700, + "policy_sampling_seed": 102017, + "minibatch_seed": 302017, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "af0e370c98101dae66fc3ae5f66047d573fb767917e2d74319069cc75f82ca6b", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2018.json b/configs/figure2/mlp/seed_2018.json new file mode 100644 index 0000000000000000000000000000000000000000..c9a57d802ccb60db5b01ad30322d4a71b331b235 --- /dev/null +++ b/configs/figure2/mlp/seed_2018.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2018, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2018, + "separate_rng_streams": true, + "initialization_seed": 2018, + "environment_seed": 5201800, + "policy_sampling_seed": 102018, + "minibatch_seed": 302018, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "473c2abd5dbf0604b19594a05671430bbbea667f02e8df4e3c292cd0ea02f089", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2019.json b/configs/figure2/mlp/seed_2019.json new file mode 100644 index 0000000000000000000000000000000000000000..0c6499fcec4aa3382840f9ff6f4afd7d724f98a5 --- /dev/null +++ b/configs/figure2/mlp/seed_2019.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2019, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2019, + "separate_rng_streams": true, + "initialization_seed": 2019, + "environment_seed": 5201900, + "policy_sampling_seed": 102019, + "minibatch_seed": 302019, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "987b4cf36337646a2ede352a26900de76ac49af659af6519caa7809c692d1c4f", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2020.json b/configs/figure2/mlp/seed_2020.json new file mode 100644 index 0000000000000000000000000000000000000000..94833335a6a0801a031fec51695a42e5331b16ca --- /dev/null +++ b/configs/figure2/mlp/seed_2020.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2020, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2020, + "separate_rng_streams": true, + "initialization_seed": 2020, + "environment_seed": 5202000, + "policy_sampling_seed": 102020, + "minibatch_seed": 302020, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "9d7ce49cb0c96ed0ec9af1e5b73c9f93c7ee35a74182a249c5fdcc14375058ca", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2021.json b/configs/figure2/mlp/seed_2021.json new file mode 100644 index 0000000000000000000000000000000000000000..5a285f8c40817cb6560f381a34dbf82d25d853b4 --- /dev/null +++ b/configs/figure2/mlp/seed_2021.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2021, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2021, + "separate_rng_streams": true, + "initialization_seed": 2021, + "environment_seed": 5202100, + "policy_sampling_seed": 102021, + "minibatch_seed": 302021, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "a0121d8d6fb53980e9a2e0056a70efa9405e7894de5b84d34340b03b9b936cfe", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2022.json b/configs/figure2/mlp/seed_2022.json new file mode 100644 index 0000000000000000000000000000000000000000..4906370d63824be07c718046f57521196b52a06c --- /dev/null +++ b/configs/figure2/mlp/seed_2022.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2022, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2022, + "separate_rng_streams": true, + "initialization_seed": 2022, + "environment_seed": 5202200, + "policy_sampling_seed": 102022, + "minibatch_seed": 302022, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "5cf8fe449402a9217e565ce53111bc58b2e6a2f3a8df2ebe904ec9afbd81665e", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2023.json b/configs/figure2/mlp/seed_2023.json new file mode 100644 index 0000000000000000000000000000000000000000..ccb7216beb325297f4321ff97a313a4a8af425e8 --- /dev/null +++ b/configs/figure2/mlp/seed_2023.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2023, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2023, + "separate_rng_streams": true, + "initialization_seed": 2023, + "environment_seed": 5202300, + "policy_sampling_seed": 102023, + "minibatch_seed": 302023, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "40b3c46e8e20197007f9d394bb9a282d707df23f73e8aede75991084df8dd5e2", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2024.json b/configs/figure2/mlp/seed_2024.json new file mode 100644 index 0000000000000000000000000000000000000000..a33dc7fbc36c21860dd0e73461a488352822d9bd --- /dev/null +++ b/configs/figure2/mlp/seed_2024.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2024, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2024, + "separate_rng_streams": true, + "initialization_seed": 2024, + "environment_seed": 5202400, + "policy_sampling_seed": 102024, + "minibatch_seed": 302024, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "d3d564c0b8901fee3bdfccdce8fb678b4bda7ee8f7c9454ca2601eee8d8a8048", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2025.json b/configs/figure2/mlp/seed_2025.json new file mode 100644 index 0000000000000000000000000000000000000000..ed0d476fa821d2b21279604b42c8c26e4463105e --- /dev/null +++ b/configs/figure2/mlp/seed_2025.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2025, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2025, + "separate_rng_streams": true, + "initialization_seed": 2025, + "environment_seed": 5202500, + "policy_sampling_seed": 102025, + "minibatch_seed": 302025, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "8e8ef57e505352aadb496afc69ff08ee15caf9e52cb82f6b7773aeff701e4617", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2026.json b/configs/figure2/mlp/seed_2026.json new file mode 100644 index 0000000000000000000000000000000000000000..7397aad2c45feda0cb6196bc7d32301eda576aae --- /dev/null +++ b/configs/figure2/mlp/seed_2026.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2026, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2026, + "separate_rng_streams": true, + "initialization_seed": 2026, + "environment_seed": 5202600, + "policy_sampling_seed": 102026, + "minibatch_seed": 302026, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "014bc75abbb7047ae99c521fd34b3179c0a5462b31f02ec6a15a06da44e9459f", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2027.json b/configs/figure2/mlp/seed_2027.json new file mode 100644 index 0000000000000000000000000000000000000000..e82698ea62b22d9d69f9240ba4fa9ef2894d67d8 --- /dev/null +++ b/configs/figure2/mlp/seed_2027.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2027, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2027, + "separate_rng_streams": true, + "initialization_seed": 2027, + "environment_seed": 5202700, + "policy_sampling_seed": 102027, + "minibatch_seed": 302027, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "4c4ad13c44d5d4df2a7347d2f954755c3f76955e122bd735a2967543f74a374f", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2028.json b/configs/figure2/mlp/seed_2028.json new file mode 100644 index 0000000000000000000000000000000000000000..0d81617fac13503af389508ecb61f67dba5ae545 --- /dev/null +++ b/configs/figure2/mlp/seed_2028.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2028, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2028, + "separate_rng_streams": true, + "initialization_seed": 2028, + "environment_seed": 5202800, + "policy_sampling_seed": 102028, + "minibatch_seed": 302028, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "fee737e447bd4ab6542e53d21fe271887f6a88090634decd0b8925d253a15d4e", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2029.json b/configs/figure2/mlp/seed_2029.json new file mode 100644 index 0000000000000000000000000000000000000000..e3c09c8a599f79460ca4c68ea372d5bb1fae03ad --- /dev/null +++ b/configs/figure2/mlp/seed_2029.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2029, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2029, + "separate_rng_streams": true, + "initialization_seed": 2029, + "environment_seed": 5202900, + "policy_sampling_seed": 102029, + "minibatch_seed": 302029, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "125c5cbcab296c4cd98a6af032c440e8eb82a722c43a6165045e00c0f4b601a8", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure2/mlp/seed_2030.json b/configs/figure2/mlp/seed_2030.json new file mode 100644 index 0000000000000000000000000000000000000000..515887bb0271a113a162524bfbcd8725dd04a5bc --- /dev/null +++ b/configs/figure2/mlp/seed_2030.json @@ -0,0 +1,61 @@ +{ + "method": "mlp", + "seed": 2030, + "config": { + "total_timesteps": 32768, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9920319148370607, + "lam": 0.951229424500714, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 2030, + "separate_rng_streams": true, + "initialization_seed": 2030, + "environment_seed": 5203000, + "policy_sampling_seed": 102030, + "minibatch_seed": 302030, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 4, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "actor_parameter_count": 404527, + "critic_parameter_count": 242433, + "critic_sha256": "8f953b7060aefdf6c736b37f2608446408210afc49072c99e0617134e0ccc0a3", + "graph_sha256": "547da3a28a9c7acb279cbc071f7582cbc45f3da3f07503e0d042980fb99eaa03", + "graph_buffers_sha256": null, + "action_storage": "pre-tanh Gaussian coordinates", + "observation_dim": 688, + "action_dim": 6, + "teacher_used": false, + "pretrained_actor_loaded": false, + "readout": "normalized", + "body": "NeuromechFly", + "exploration_sampling": "antithetic_environment_halves", + "initial_log_std": -1.0, + "leg_residual_l2_limit_rad": 0.25980762113533157, + "motor_interface": "normalized_multijoint_leg_synergy" +} diff --git a/configs/figure3/connectome/seed_4301.json b/configs/figure3/connectome/seed_4301.json new file mode 100644 index 0000000000000000000000000000000000000000..dfd54bdbac19049187fdc5ff241f146c9c3508b7 --- /dev/null +++ b/configs/figure3/connectome/seed_4301.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4301, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4301, + "separate_rng_streams": true, + "initialization_seed": 4301, + "environment_seed": 5430100, + "policy_sampling_seed": 104301, + "minibatch_seed": 304301, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "a613b6d001f7113171f833c44f85b733f7e0c6e3caba887b5a9a8777096e0794" +} diff --git a/configs/figure3/connectome/seed_4302.json b/configs/figure3/connectome/seed_4302.json new file mode 100644 index 0000000000000000000000000000000000000000..346f83bf0be0362cbde44c9057aae24cb478856c --- /dev/null +++ b/configs/figure3/connectome/seed_4302.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4302, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4302, + "separate_rng_streams": true, + "initialization_seed": 4302, + "environment_seed": 5430200, + "policy_sampling_seed": 104302, + "minibatch_seed": 304302, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "f0d505b318e0a7477600060a9699cc9c1ddb09850dc28b3fdfe17b0ffef8af52" +} diff --git a/configs/figure3/connectome/seed_4303.json b/configs/figure3/connectome/seed_4303.json new file mode 100644 index 0000000000000000000000000000000000000000..9994095c8e4a9b776bbbc5285ff1a77133549c5b --- /dev/null +++ b/configs/figure3/connectome/seed_4303.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4303, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4303, + "separate_rng_streams": true, + "initialization_seed": 4303, + "environment_seed": 5430300, + "policy_sampling_seed": 104303, + "minibatch_seed": 304303, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "648e1e4494b94d865fd407e590641c6c00d43c779cec46b933b6d43768b739f3" +} diff --git a/configs/figure3/connectome/seed_4304.json b/configs/figure3/connectome/seed_4304.json new file mode 100644 index 0000000000000000000000000000000000000000..3a190790ffbd34f716a060808c62fd41de9781d7 --- /dev/null +++ b/configs/figure3/connectome/seed_4304.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4304, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4304, + "separate_rng_streams": true, + "initialization_seed": 4304, + "environment_seed": 5430400, + "policy_sampling_seed": 104304, + "minibatch_seed": 304304, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "d32e591686ca064a27be2549d04358bfea2d1ffb78fe01552f0d9af06403484c" +} diff --git a/configs/figure3/connectome/seed_4305.json b/configs/figure3/connectome/seed_4305.json new file mode 100644 index 0000000000000000000000000000000000000000..8e9e047c87edd5d540c2d56ed3e9b3321e71c4e5 --- /dev/null +++ b/configs/figure3/connectome/seed_4305.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4305, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4305, + "separate_rng_streams": true, + "initialization_seed": 4305, + "environment_seed": 5430500, + "policy_sampling_seed": 104305, + "minibatch_seed": 304305, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "28d95bcf388e5dca37e33dc355bdc7a60bb8cdc65a6c1d8c53bee0a2bed33666" +} diff --git a/configs/figure3/connectome/seed_4306.json b/configs/figure3/connectome/seed_4306.json new file mode 100644 index 0000000000000000000000000000000000000000..e2a2b64094360cf961f9a2a575d929ee4b2cf460 --- /dev/null +++ b/configs/figure3/connectome/seed_4306.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4306, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4306, + "separate_rng_streams": true, + "initialization_seed": 4306, + "environment_seed": 5430600, + "policy_sampling_seed": 104306, + "minibatch_seed": 304306, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "736b079bab5880ad159362aedcd882ce04a5b82f0ff1930189af684efc84049b" +} diff --git a/configs/figure3/connectome/seed_4307.json b/configs/figure3/connectome/seed_4307.json new file mode 100644 index 0000000000000000000000000000000000000000..f21c3475cbff3dff43c8583a6c92379c232f2c1c --- /dev/null +++ b/configs/figure3/connectome/seed_4307.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4307, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4307, + "separate_rng_streams": true, + "initialization_seed": 4307, + "environment_seed": 5430700, + "policy_sampling_seed": 104307, + "minibatch_seed": 304307, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "d6f675d16bb31287b8dad9807624602bbb942767d4a61896c21985c1e167b491" +} diff --git a/configs/figure3/connectome/seed_4308.json b/configs/figure3/connectome/seed_4308.json new file mode 100644 index 0000000000000000000000000000000000000000..90511377c084e73efd23a32bcbc95dc34e159761 --- /dev/null +++ b/configs/figure3/connectome/seed_4308.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4308, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4308, + "separate_rng_streams": true, + "initialization_seed": 4308, + "environment_seed": 5430800, + "policy_sampling_seed": 104308, + "minibatch_seed": 304308, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "fc349ff7236d14963f8161d3737d0de9c9e8e281b32bdc813bed4a0f9b53f9a4" +} diff --git a/configs/figure3/connectome/seed_4309.json b/configs/figure3/connectome/seed_4309.json new file mode 100644 index 0000000000000000000000000000000000000000..7787d67eafba57178e0123b7548c717ffb7bf786 --- /dev/null +++ b/configs/figure3/connectome/seed_4309.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4309, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4309, + "separate_rng_streams": true, + "initialization_seed": 4309, + "environment_seed": 5430900, + "policy_sampling_seed": 104309, + "minibatch_seed": 304309, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "45fbb8853f3ec270aa06246d1847d27be03cd26a03a6c07e7a868001eabdea97" +} diff --git a/configs/figure3/connectome/seed_4310.json b/configs/figure3/connectome/seed_4310.json new file mode 100644 index 0000000000000000000000000000000000000000..ca70af4a6f8533572b13ca7dac09b4716836f30c --- /dev/null +++ b/configs/figure3/connectome/seed_4310.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4310, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4310, + "separate_rng_streams": true, + "initialization_seed": 4310, + "environment_seed": 5431000, + "policy_sampling_seed": 104310, + "minibatch_seed": 304310, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "db026d898da25d1847e474daafe21c464c24e1066f96f34040018f69d4b66bc6" +} diff --git a/configs/figure3/connectome/seed_4311.json b/configs/figure3/connectome/seed_4311.json new file mode 100644 index 0000000000000000000000000000000000000000..61bd06adc2aae507b612ed496f1411e086836d60 --- /dev/null +++ b/configs/figure3/connectome/seed_4311.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4311, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4311, + "separate_rng_streams": true, + "initialization_seed": 4311, + "environment_seed": 5431100, + "policy_sampling_seed": 104311, + "minibatch_seed": 304311, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "d46aa6b725ae696e716d870060a2ef37b4841f3214efe66e1a86edab33ca23af" +} diff --git a/configs/figure3/connectome/seed_4312.json b/configs/figure3/connectome/seed_4312.json new file mode 100644 index 0000000000000000000000000000000000000000..8d85a6cf3dc40a13b2b0ff83382d5022bb5b4afe --- /dev/null +++ b/configs/figure3/connectome/seed_4312.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4312, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4312, + "separate_rng_streams": true, + "initialization_seed": 4312, + "environment_seed": 5431200, + "policy_sampling_seed": 104312, + "minibatch_seed": 304312, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "71be9bd23fefbed71fabf3d30d14c51846b1cc5e72c9190dfa1f4f7c132ba38e" +} diff --git a/configs/figure3/connectome/seed_4313.json b/configs/figure3/connectome/seed_4313.json new file mode 100644 index 0000000000000000000000000000000000000000..bc5913081120defae2f56308b8ccfd715b5db2ed --- /dev/null +++ b/configs/figure3/connectome/seed_4313.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4313, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4313, + "separate_rng_streams": true, + "initialization_seed": 4313, + "environment_seed": 5431300, + "policy_sampling_seed": 104313, + "minibatch_seed": 304313, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "60f02dc38654802187bfdc7380ce0ed49d11f14904f22c4e6ac3b2268f0739c3" +} diff --git a/configs/figure3/connectome/seed_4314.json b/configs/figure3/connectome/seed_4314.json new file mode 100644 index 0000000000000000000000000000000000000000..7ed3302b3458379a553317f86e69555fd9f67837 --- /dev/null +++ b/configs/figure3/connectome/seed_4314.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4314, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4314, + "separate_rng_streams": true, + "initialization_seed": 4314, + "environment_seed": 5431400, + "policy_sampling_seed": 104314, + "minibatch_seed": 304314, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "b61aec65409ec6c977b2a792e9f2fdc8a44b331834516d57a7ff93c45f6ddcff" +} diff --git a/configs/figure3/connectome/seed_4315.json b/configs/figure3/connectome/seed_4315.json new file mode 100644 index 0000000000000000000000000000000000000000..22e0818d52b2c86ca79cf7cfb1a9ebae53f9788b --- /dev/null +++ b/configs/figure3/connectome/seed_4315.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4315, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4315, + "separate_rng_streams": true, + "initialization_seed": 4315, + "environment_seed": 5431500, + "policy_sampling_seed": 104315, + "minibatch_seed": 304315, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "947b0cba5cf97923970a543372437c5a75f7721ba13d3a1cd67ed1319e9044de" +} diff --git a/configs/figure3/connectome/seed_4316.json b/configs/figure3/connectome/seed_4316.json new file mode 100644 index 0000000000000000000000000000000000000000..67d6cd6751cdaf6f16e3e483afc5e73efdfde1c6 --- /dev/null +++ b/configs/figure3/connectome/seed_4316.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4316, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4316, + "separate_rng_streams": true, + "initialization_seed": 4316, + "environment_seed": 5431600, + "policy_sampling_seed": 104316, + "minibatch_seed": 304316, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "30681dcb743bb4b8d9776f3ba28cd7e629ce654bf1590926f78640893f3d43f6" +} diff --git a/configs/figure3/connectome/seed_4317.json b/configs/figure3/connectome/seed_4317.json new file mode 100644 index 0000000000000000000000000000000000000000..39d6ad1309a8f9877e832aaa1dd1a2c676642cd5 --- /dev/null +++ b/configs/figure3/connectome/seed_4317.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4317, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4317, + "separate_rng_streams": true, + "initialization_seed": 4317, + "environment_seed": 5431700, + "policy_sampling_seed": 104317, + "minibatch_seed": 304317, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "3d463a4df5827f4c01c0a331f853ce8f28243a77288909069e6bb4378c7b99fb" +} diff --git a/configs/figure3/connectome/seed_4318.json b/configs/figure3/connectome/seed_4318.json new file mode 100644 index 0000000000000000000000000000000000000000..54be99c0bf2c805a8aa013e6406e240c2973bc90 --- /dev/null +++ b/configs/figure3/connectome/seed_4318.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4318, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4318, + "separate_rng_streams": true, + "initialization_seed": 4318, + "environment_seed": 5431800, + "policy_sampling_seed": 104318, + "minibatch_seed": 304318, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "0bc9668db84a90346dbe8ba98f4340fd0604e1d3385e56123c4c0aef26ebfed6" +} diff --git a/configs/figure3/connectome/seed_4319.json b/configs/figure3/connectome/seed_4319.json new file mode 100644 index 0000000000000000000000000000000000000000..105e6b664a0014105b2da7fc2796a9e00d4de7bd --- /dev/null +++ b/configs/figure3/connectome/seed_4319.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4319, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4319, + "separate_rng_streams": true, + "initialization_seed": 4319, + "environment_seed": 5431900, + "policy_sampling_seed": 104319, + "minibatch_seed": 304319, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "fe1fd70d81e5e5e7d93ac5ee65dade69a79396ca4e3854990109498984360798" +} diff --git a/configs/figure3/connectome/seed_4320.json b/configs/figure3/connectome/seed_4320.json new file mode 100644 index 0000000000000000000000000000000000000000..41714550a9b65e47147664302503e9eb7927dc9c --- /dev/null +++ b/configs/figure3/connectome/seed_4320.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4320, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4320, + "separate_rng_streams": true, + "initialization_seed": 4320, + "environment_seed": 5432000, + "policy_sampling_seed": 104320, + "minibatch_seed": 304320, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "08bd2b9c11c538f6720183d66db4659e3dcb8769764f97cbc700bfdb2c270c21" +} diff --git a/configs/figure3/connectome/seed_4321.json b/configs/figure3/connectome/seed_4321.json new file mode 100644 index 0000000000000000000000000000000000000000..e63697fcb33e8ded8da920d3db4e65fdbb4f6f1d --- /dev/null +++ b/configs/figure3/connectome/seed_4321.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4321, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4321, + "separate_rng_streams": true, + "initialization_seed": 4321, + "environment_seed": 5432100, + "policy_sampling_seed": 104321, + "minibatch_seed": 304321, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "4a1c62130c9bff1f9860a928ff2c508bcdc6fdb1f5bd0f7103263ac080269f0b" +} diff --git a/configs/figure3/connectome/seed_4322.json b/configs/figure3/connectome/seed_4322.json new file mode 100644 index 0000000000000000000000000000000000000000..cd4aa180c13bb142a5be1d7363001aa24973190e --- /dev/null +++ b/configs/figure3/connectome/seed_4322.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4322, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4322, + "separate_rng_streams": true, + "initialization_seed": 4322, + "environment_seed": 5432200, + "policy_sampling_seed": 104322, + "minibatch_seed": 304322, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "f9694fd5f129efb276d7384f125f861d7f0c4ef55573da5ada164f93c82c727e" +} diff --git a/configs/figure3/connectome/seed_4323.json b/configs/figure3/connectome/seed_4323.json new file mode 100644 index 0000000000000000000000000000000000000000..02fe8feebfaaf24141a5acd3d4a6fc71863b541a --- /dev/null +++ b/configs/figure3/connectome/seed_4323.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4323, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4323, + "separate_rng_streams": true, + "initialization_seed": 4323, + "environment_seed": 5432300, + "policy_sampling_seed": 104323, + "minibatch_seed": 304323, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "e56ea095a13bc2557dc42189042018a1b452ecdd7f65c6f5f337ddf357b09696" +} diff --git a/configs/figure3/connectome/seed_4324.json b/configs/figure3/connectome/seed_4324.json new file mode 100644 index 0000000000000000000000000000000000000000..90393158a3fce92370a2c0734218278ac3df7858 --- /dev/null +++ b/configs/figure3/connectome/seed_4324.json @@ -0,0 +1,53 @@ +{ + "method": "connectome", + "seed": 4324, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4324, + "separate_rng_streams": true, + "initialization_seed": 4324, + "environment_seed": 5432400, + "policy_sampling_seed": 104324, + "minibatch_seed": 304324, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "15f8725bd5363025e572f2608aecab6d7541b0bdf9fc313bdb1448e96a60546d" +} diff --git a/configs/figure3/mlp/seed_4301.json b/configs/figure3/mlp/seed_4301.json new file mode 100644 index 0000000000000000000000000000000000000000..57d1a9b7264af14b1a41537c126f9746db721d9c --- /dev/null +++ b/configs/figure3/mlp/seed_4301.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4301, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4301, + "separate_rng_streams": true, + "initialization_seed": 4301, + "environment_seed": 5430100, + "policy_sampling_seed": 104301, + "minibatch_seed": 304301, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "24cec1644c72bae0487d2d317f5acccc971190f97071abef15b3111e1b727541" +} diff --git a/configs/figure3/mlp/seed_4302.json b/configs/figure3/mlp/seed_4302.json new file mode 100644 index 0000000000000000000000000000000000000000..9a3b78a2eff9bfdea615a0eeb81d16f9da7953e6 --- /dev/null +++ b/configs/figure3/mlp/seed_4302.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4302, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4302, + "separate_rng_streams": true, + "initialization_seed": 4302, + "environment_seed": 5430200, + "policy_sampling_seed": 104302, + "minibatch_seed": 304302, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "be1aba1ed01e2050b3f855854cdd5b020febed7a6944fa4e47f0775394d4dc3e" +} diff --git a/configs/figure3/mlp/seed_4303.json b/configs/figure3/mlp/seed_4303.json new file mode 100644 index 0000000000000000000000000000000000000000..cf1966c49363db59d39f2b58b34493b5fe7eb329 --- /dev/null +++ b/configs/figure3/mlp/seed_4303.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4303, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4303, + "separate_rng_streams": true, + "initialization_seed": 4303, + "environment_seed": 5430300, + "policy_sampling_seed": 104303, + "minibatch_seed": 304303, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "1a92b7c97dd76a13e66fa82f286017119e9943de366c8367e8e1f2788eebfa8f" +} diff --git a/configs/figure3/mlp/seed_4304.json b/configs/figure3/mlp/seed_4304.json new file mode 100644 index 0000000000000000000000000000000000000000..279b9d3db18e63205b7f8a1075d1e1fc99cd69e6 --- /dev/null +++ b/configs/figure3/mlp/seed_4304.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4304, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4304, + "separate_rng_streams": true, + "initialization_seed": 4304, + "environment_seed": 5430400, + "policy_sampling_seed": 104304, + "minibatch_seed": 304304, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "c6d28faa67c86b58b231dd2b2778fc4bced7e56abe4cb20177a18914089c1c6c" +} diff --git a/configs/figure3/mlp/seed_4305.json b/configs/figure3/mlp/seed_4305.json new file mode 100644 index 0000000000000000000000000000000000000000..68526aa9a4eb277a4d6dd0de5df2f9c43af63476 --- /dev/null +++ b/configs/figure3/mlp/seed_4305.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4305, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4305, + "separate_rng_streams": true, + "initialization_seed": 4305, + "environment_seed": 5430500, + "policy_sampling_seed": 104305, + "minibatch_seed": 304305, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "6c9a31134860eca5fbd241bd6077fa821db7b13ff74abbe856f5120b165d44a9" +} diff --git a/configs/figure3/mlp/seed_4306.json b/configs/figure3/mlp/seed_4306.json new file mode 100644 index 0000000000000000000000000000000000000000..2cf5586439dc36bf11a234a5cb7084666771589d --- /dev/null +++ b/configs/figure3/mlp/seed_4306.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4306, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4306, + "separate_rng_streams": true, + "initialization_seed": 4306, + "environment_seed": 5430600, + "policy_sampling_seed": 104306, + "minibatch_seed": 304306, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "64104db0b7be23edc62e0bf833a0c61230d47b74d5099d468a2fc828d05eb858" +} diff --git a/configs/figure3/mlp/seed_4307.json b/configs/figure3/mlp/seed_4307.json new file mode 100644 index 0000000000000000000000000000000000000000..74ab1872c4c9d332d08be45b805b32f96c0cf57c --- /dev/null +++ b/configs/figure3/mlp/seed_4307.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4307, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4307, + "separate_rng_streams": true, + "initialization_seed": 4307, + "environment_seed": 5430700, + "policy_sampling_seed": 104307, + "minibatch_seed": 304307, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "f12b7118850c6116532fcbe650fa18891a11a42f338e225b1bf91fc39acf7b12" +} diff --git a/configs/figure3/mlp/seed_4308.json b/configs/figure3/mlp/seed_4308.json new file mode 100644 index 0000000000000000000000000000000000000000..b22b4ba5f2df89495e3226b5627e8c5ba99ceccd --- /dev/null +++ b/configs/figure3/mlp/seed_4308.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4308, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4308, + "separate_rng_streams": true, + "initialization_seed": 4308, + "environment_seed": 5430800, + "policy_sampling_seed": 104308, + "minibatch_seed": 304308, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "e4bf60c1ec993dcd763e460f89bb1beacb9b955194d6c386a3f5a8574afb82f4" +} diff --git a/configs/figure3/mlp/seed_4309.json b/configs/figure3/mlp/seed_4309.json new file mode 100644 index 0000000000000000000000000000000000000000..0590d2c49b4a2f87f88f13398e5352fb148b082c --- /dev/null +++ b/configs/figure3/mlp/seed_4309.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4309, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4309, + "separate_rng_streams": true, + "initialization_seed": 4309, + "environment_seed": 5430900, + "policy_sampling_seed": 104309, + "minibatch_seed": 304309, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "d646078398b550975c4ed801df4dee7469378cfae8e9bf1d1ae7babc1b62ccee" +} diff --git a/configs/figure3/mlp/seed_4310.json b/configs/figure3/mlp/seed_4310.json new file mode 100644 index 0000000000000000000000000000000000000000..a7f7e1c33454f3d21263c69865b7d4b000744993 --- /dev/null +++ b/configs/figure3/mlp/seed_4310.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4310, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4310, + "separate_rng_streams": true, + "initialization_seed": 4310, + "environment_seed": 5431000, + "policy_sampling_seed": 104310, + "minibatch_seed": 304310, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "611ff81e028f7c9b9043d05208ff2bf4886b7cf17e01ea66c4cb27dc75127ef5" +} diff --git a/configs/figure3/mlp/seed_4311.json b/configs/figure3/mlp/seed_4311.json new file mode 100644 index 0000000000000000000000000000000000000000..40c52dbbc1224aa47eff5679d093d0e4407444e4 --- /dev/null +++ b/configs/figure3/mlp/seed_4311.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4311, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4311, + "separate_rng_streams": true, + "initialization_seed": 4311, + "environment_seed": 5431100, + "policy_sampling_seed": 104311, + "minibatch_seed": 304311, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "caf65d4c7ebb344fe5e290f981e4daeafc51c328d39f0eef1bdcea17629deaa2" +} diff --git a/configs/figure3/mlp/seed_4312.json b/configs/figure3/mlp/seed_4312.json new file mode 100644 index 0000000000000000000000000000000000000000..755bd4ca53e4005ed3f89cd8da19b78be7e2d08f --- /dev/null +++ b/configs/figure3/mlp/seed_4312.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4312, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4312, + "separate_rng_streams": true, + "initialization_seed": 4312, + "environment_seed": 5431200, + "policy_sampling_seed": 104312, + "minibatch_seed": 304312, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "dc6a0b58501335750fa096b3ca9e74510863a35b1d42c85cbbe39db8775311dc" +} diff --git a/configs/figure3/mlp/seed_4313.json b/configs/figure3/mlp/seed_4313.json new file mode 100644 index 0000000000000000000000000000000000000000..e2e6476e75f9b9f8b9a9ec148ddb04559b6270c9 --- /dev/null +++ b/configs/figure3/mlp/seed_4313.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4313, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4313, + "separate_rng_streams": true, + "initialization_seed": 4313, + "environment_seed": 5431300, + "policy_sampling_seed": 104313, + "minibatch_seed": 304313, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "6e5c0ebd50443cd36a7139648e10d1ef08407f11bbd822d19bf2786ea01f6919" +} diff --git a/configs/figure3/mlp/seed_4314.json b/configs/figure3/mlp/seed_4314.json new file mode 100644 index 0000000000000000000000000000000000000000..9b13fc0b8e91270f3b378eb8bef71a885750dcae --- /dev/null +++ b/configs/figure3/mlp/seed_4314.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4314, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4314, + "separate_rng_streams": true, + "initialization_seed": 4314, + "environment_seed": 5431400, + "policy_sampling_seed": 104314, + "minibatch_seed": 304314, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "58f9434c83d2fe7d00f87c23c5e48f934992f5ed0858711c861fc8dc538d2206" +} diff --git a/configs/figure3/mlp/seed_4315.json b/configs/figure3/mlp/seed_4315.json new file mode 100644 index 0000000000000000000000000000000000000000..e69e4f52c2306ba3d4d93b890b3f15e0f4379812 --- /dev/null +++ b/configs/figure3/mlp/seed_4315.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4315, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4315, + "separate_rng_streams": true, + "initialization_seed": 4315, + "environment_seed": 5431500, + "policy_sampling_seed": 104315, + "minibatch_seed": 304315, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "dd05b22f3973284160e5fa1692185594115da41d2abf7926ab9f2f296ff46fdf" +} diff --git a/configs/figure3/mlp/seed_4316.json b/configs/figure3/mlp/seed_4316.json new file mode 100644 index 0000000000000000000000000000000000000000..54cbca99be9891830588de11e98e8cc5f6c4046c --- /dev/null +++ b/configs/figure3/mlp/seed_4316.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4316, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4316, + "separate_rng_streams": true, + "initialization_seed": 4316, + "environment_seed": 5431600, + "policy_sampling_seed": 104316, + "minibatch_seed": 304316, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "86786cc70a833d90e00bcfdd60205b5bec279e88e389c1ba36688ce9ac806549" +} diff --git a/configs/figure3/mlp/seed_4317.json b/configs/figure3/mlp/seed_4317.json new file mode 100644 index 0000000000000000000000000000000000000000..ed0e09f922ef6fca1f8569d10bda6ee4dfcd0606 --- /dev/null +++ b/configs/figure3/mlp/seed_4317.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4317, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4317, + "separate_rng_streams": true, + "initialization_seed": 4317, + "environment_seed": 5431700, + "policy_sampling_seed": 104317, + "minibatch_seed": 304317, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "8d61b27de2776956d17cf4d198859c0016e58e7c0a92d6357ab70c95ab6af66f" +} diff --git a/configs/figure3/mlp/seed_4318.json b/configs/figure3/mlp/seed_4318.json new file mode 100644 index 0000000000000000000000000000000000000000..fa141572ec10e82d507fbafdc33fa68751ab63a7 --- /dev/null +++ b/configs/figure3/mlp/seed_4318.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4318, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4318, + "separate_rng_streams": true, + "initialization_seed": 4318, + "environment_seed": 5431800, + "policy_sampling_seed": 104318, + "minibatch_seed": 304318, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "931baed60aa14202ede0c9e2c4fcb44613fb59eb773b60d4800e7f4bde5440b9" +} diff --git a/configs/figure3/mlp/seed_4319.json b/configs/figure3/mlp/seed_4319.json new file mode 100644 index 0000000000000000000000000000000000000000..e3d9dd33128d8ab0723b5dc10dcad3d18e6e3552 --- /dev/null +++ b/configs/figure3/mlp/seed_4319.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4319, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4319, + "separate_rng_streams": true, + "initialization_seed": 4319, + "environment_seed": 5431900, + "policy_sampling_seed": 104319, + "minibatch_seed": 304319, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "5c31035770afb1b454e25455e2255c5b73e8c84789a42a9a39a2eda9f7f7856f" +} diff --git a/configs/figure3/mlp/seed_4320.json b/configs/figure3/mlp/seed_4320.json new file mode 100644 index 0000000000000000000000000000000000000000..ca217b77bee86720d539ee678660040f1924504b --- /dev/null +++ b/configs/figure3/mlp/seed_4320.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4320, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4320, + "separate_rng_streams": true, + "initialization_seed": 4320, + "environment_seed": 5432000, + "policy_sampling_seed": 104320, + "minibatch_seed": 304320, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "ded64f87b3e7ef858991726a96d0ebe8336e933ec2ce56585a1cf70e30597d41" +} diff --git a/configs/figure3/mlp/seed_4321.json b/configs/figure3/mlp/seed_4321.json new file mode 100644 index 0000000000000000000000000000000000000000..ad268c4937b5e2a4f9e234b02da8e048e661ecac --- /dev/null +++ b/configs/figure3/mlp/seed_4321.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4321, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4321, + "separate_rng_streams": true, + "initialization_seed": 4321, + "environment_seed": 5432100, + "policy_sampling_seed": 104321, + "minibatch_seed": 304321, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "758b01d8115d81f8619776d4c86f58ab7278f917c417104a4e514c03bf43ccb8" +} diff --git a/configs/figure3/mlp/seed_4322.json b/configs/figure3/mlp/seed_4322.json new file mode 100644 index 0000000000000000000000000000000000000000..99a6c73c1cc7c6ccd43d438bcc9154d1e2bf88cf --- /dev/null +++ b/configs/figure3/mlp/seed_4322.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4322, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4322, + "separate_rng_streams": true, + "initialization_seed": 4322, + "environment_seed": 5432200, + "policy_sampling_seed": 104322, + "minibatch_seed": 304322, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "734281ae00703cbfb815150e6d3631ca48458c3e9c7df65b30ed31a31ed8b53b" +} diff --git a/configs/figure3/mlp/seed_4323.json b/configs/figure3/mlp/seed_4323.json new file mode 100644 index 0000000000000000000000000000000000000000..54f329f2450021ea51c8dff20b8373ade6460829 --- /dev/null +++ b/configs/figure3/mlp/seed_4323.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4323, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4323, + "separate_rng_streams": true, + "initialization_seed": 4323, + "environment_seed": 5432300, + "policy_sampling_seed": 104323, + "minibatch_seed": 304323, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "e2687db2b9ae1e6900258e7091d8a964c200241967c51f1be6ccd7e8b5871389" +} diff --git a/configs/figure3/mlp/seed_4324.json b/configs/figure3/mlp/seed_4324.json new file mode 100644 index 0000000000000000000000000000000000000000..094a6056430cf322f0968935de49d8b621c0a806 --- /dev/null +++ b/configs/figure3/mlp/seed_4324.json @@ -0,0 +1,53 @@ +{ + "method": "mlp", + "seed": 4324, + "ppo": { + "total_timesteps": 8192, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4324, + "separate_rng_streams": true, + "initialization_seed": 4324, + "environment_seed": 5432400, + "policy_sampling_seed": 104324, + "minibatch_seed": 304324, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "fixed_budget": 8192, + "adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "original_initialization_sha256": "e48dfff74345994f61a98ae2731e38c4fea7ae7ec4fe411076769f7bdb146ae4" +} diff --git a/configs/figure4/connectome/seed_4401.json b/configs/figure4/connectome/seed_4401.json new file mode 100644 index 0000000000000000000000000000000000000000..b4ea8538f7b8baae807f72e422cf37543fb76d6e --- /dev/null +++ b/configs/figure4/connectome/seed_4401.json @@ -0,0 +1,104 @@ +{ + "seed": 4401, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4401, + "separate_rng_streams": true, + "initialization_seed": 4401, + "environment_seed": 6440100, + "policy_sampling_seed": 104401, + "minibatch_seed": 304401, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4401, + "separate_rng_streams": true, + "initialization_seed": 4401, + "environment_seed": 6440100, + "policy_sampling_seed": 104401, + "minibatch_seed": 304401, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "1514da1d5dbd96fc3006695fc4654c6a98b437d73c9e3b079dcd53530c5538bf", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 25600, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "lr_floor_20261010" +} diff --git a/configs/figure4/connectome/seed_4402.json b/configs/figure4/connectome/seed_4402.json new file mode 100644 index 0000000000000000000000000000000000000000..f54f4ff6619be702963f9845a50e31181d40c99b --- /dev/null +++ b/configs/figure4/connectome/seed_4402.json @@ -0,0 +1,104 @@ +{ + "seed": 4402, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4402, + "separate_rng_streams": true, + "initialization_seed": 4402, + "environment_seed": 6440200, + "policy_sampling_seed": 104402, + "minibatch_seed": 304402, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4402, + "separate_rng_streams": true, + "initialization_seed": 4402, + "environment_seed": 6440200, + "policy_sampling_seed": 104402, + "minibatch_seed": 304402, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "1514da1d5dbd96fc3006695fc4654c6a98b437d73c9e3b079dcd53530c5538bf", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/connectome/seed_4403.json b/configs/figure4/connectome/seed_4403.json new file mode 100644 index 0000000000000000000000000000000000000000..7244c5fe73f07fa421ac4d5477ea26a51a8cd49a --- /dev/null +++ b/configs/figure4/connectome/seed_4403.json @@ -0,0 +1,104 @@ +{ + "seed": 4403, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4403, + "separate_rng_streams": true, + "initialization_seed": 4403, + "environment_seed": 6440300, + "policy_sampling_seed": 104403, + "minibatch_seed": 304403, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4403, + "separate_rng_streams": true, + "initialization_seed": 4403, + "environment_seed": 6440300, + "policy_sampling_seed": 104403, + "minibatch_seed": 304403, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "1514da1d5dbd96fc3006695fc4654c6a98b437d73c9e3b079dcd53530c5538bf", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 25600, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "lr_floor_20261010" +} diff --git a/configs/figure4/connectome/seed_4404.json b/configs/figure4/connectome/seed_4404.json new file mode 100644 index 0000000000000000000000000000000000000000..ab97e594afcc610ff41a16d0c8f89d5f9eb062b7 --- /dev/null +++ b/configs/figure4/connectome/seed_4404.json @@ -0,0 +1,104 @@ +{ + "seed": 4404, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4404, + "separate_rng_streams": true, + "initialization_seed": 4404, + "environment_seed": 6440400, + "policy_sampling_seed": 104404, + "minibatch_seed": 304404, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4404, + "separate_rng_streams": true, + "initialization_seed": 4404, + "environment_seed": 6440400, + "policy_sampling_seed": 104404, + "minibatch_seed": 304404, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "1514da1d5dbd96fc3006695fc4654c6a98b437d73c9e3b079dcd53530c5538bf", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/connectome/seed_4405.json b/configs/figure4/connectome/seed_4405.json new file mode 100644 index 0000000000000000000000000000000000000000..ad69a0fb8abab74741bd7e8f0d95430fb2c02263 --- /dev/null +++ b/configs/figure4/connectome/seed_4405.json @@ -0,0 +1,104 @@ +{ + "seed": 4405, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4405, + "separate_rng_streams": true, + "initialization_seed": 4405, + "environment_seed": 6440500, + "policy_sampling_seed": 104405, + "minibatch_seed": 304405, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4405, + "separate_rng_streams": true, + "initialization_seed": 4405, + "environment_seed": 6440500, + "policy_sampling_seed": 104405, + "minibatch_seed": 304405, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "1514da1d5dbd96fc3006695fc4654c6a98b437d73c9e3b079dcd53530c5538bf", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/connectome/seed_4406.json b/configs/figure4/connectome/seed_4406.json new file mode 100644 index 0000000000000000000000000000000000000000..f98ddf94549d4e9024c521af24436ecc9ad682d5 --- /dev/null +++ b/configs/figure4/connectome/seed_4406.json @@ -0,0 +1,104 @@ +{ + "seed": 4406, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4406, + "separate_rng_streams": true, + "initialization_seed": 4406, + "environment_seed": 6440600, + "policy_sampling_seed": 104406, + "minibatch_seed": 304406, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4406, + "separate_rng_streams": true, + "initialization_seed": 4406, + "environment_seed": 6440600, + "policy_sampling_seed": 104406, + "minibatch_seed": 304406, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "1514da1d5dbd96fc3006695fc4654c6a98b437d73c9e3b079dcd53530c5538bf", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 25600, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "lr_floor_20261010" +} diff --git a/configs/figure4/connectome/seed_4407.json b/configs/figure4/connectome/seed_4407.json new file mode 100644 index 0000000000000000000000000000000000000000..967e4beb65b63ceaa5dc1b3e08b9836d36dab2ce --- /dev/null +++ b/configs/figure4/connectome/seed_4407.json @@ -0,0 +1,104 @@ +{ + "seed": 4407, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4407, + "separate_rng_streams": true, + "initialization_seed": 4407, + "environment_seed": 6440700, + "policy_sampling_seed": 104407, + "minibatch_seed": 304407, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4407, + "separate_rng_streams": true, + "initialization_seed": 4407, + "environment_seed": 6440700, + "policy_sampling_seed": 104407, + "minibatch_seed": 304407, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 25600, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "lr_floor_20261010" +} diff --git a/configs/figure4/connectome/seed_4408.json b/configs/figure4/connectome/seed_4408.json new file mode 100644 index 0000000000000000000000000000000000000000..3180e3780206e2f8b87b05bd3e7ac141ff62ca1e --- /dev/null +++ b/configs/figure4/connectome/seed_4408.json @@ -0,0 +1,104 @@ +{ + "seed": 4408, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4408, + "separate_rng_streams": true, + "initialization_seed": 4408, + "environment_seed": 6440800, + "policy_sampling_seed": 104408, + "minibatch_seed": 304408, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4408, + "separate_rng_streams": true, + "initialization_seed": 4408, + "environment_seed": 6440800, + "policy_sampling_seed": 104408, + "minibatch_seed": 304408, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 25600, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "lr_floor_20261010" +} diff --git a/configs/figure4/connectome/seed_4409.json b/configs/figure4/connectome/seed_4409.json new file mode 100644 index 0000000000000000000000000000000000000000..728199615c60856f46c486b1839e8629690b1a05 --- /dev/null +++ b/configs/figure4/connectome/seed_4409.json @@ -0,0 +1,104 @@ +{ + "seed": 4409, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4409, + "separate_rng_streams": true, + "initialization_seed": 4409, + "environment_seed": 6440900, + "policy_sampling_seed": 104409, + "minibatch_seed": 304409, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4409, + "separate_rng_streams": true, + "initialization_seed": 4409, + "environment_seed": 6440900, + "policy_sampling_seed": 104409, + "minibatch_seed": 304409, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/connectome/seed_4410.json b/configs/figure4/connectome/seed_4410.json new file mode 100644 index 0000000000000000000000000000000000000000..dfd9c740e992e196c9b71b68ebd979a3072fc464 --- /dev/null +++ b/configs/figure4/connectome/seed_4410.json @@ -0,0 +1,104 @@ +{ + "seed": 4410, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4410, + "separate_rng_streams": true, + "initialization_seed": 4410, + "environment_seed": 6441000, + "policy_sampling_seed": 104410, + "minibatch_seed": 304410, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4410, + "separate_rng_streams": true, + "initialization_seed": 4410, + "environment_seed": 6441000, + "policy_sampling_seed": 104410, + "minibatch_seed": 304410, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/connectome/seed_4411.json b/configs/figure4/connectome/seed_4411.json new file mode 100644 index 0000000000000000000000000000000000000000..1b28f6dae93cd75bc153c991d9d0f395193e8acd --- /dev/null +++ b/configs/figure4/connectome/seed_4411.json @@ -0,0 +1,104 @@ +{ + "seed": 4411, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4411, + "separate_rng_streams": true, + "initialization_seed": 4411, + "environment_seed": 6441100, + "policy_sampling_seed": 104411, + "minibatch_seed": 304411, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4411, + "separate_rng_streams": true, + "initialization_seed": 4411, + "environment_seed": 6441100, + "policy_sampling_seed": 104411, + "minibatch_seed": 304411, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/connectome/seed_4412.json b/configs/figure4/connectome/seed_4412.json new file mode 100644 index 0000000000000000000000000000000000000000..c51d72153e4c9e108b0e5c7c0ee9d42504e6aa7e --- /dev/null +++ b/configs/figure4/connectome/seed_4412.json @@ -0,0 +1,104 @@ +{ + "seed": 4412, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4412, + "separate_rng_streams": true, + "initialization_seed": 4412, + "environment_seed": 6441200, + "policy_sampling_seed": 104412, + "minibatch_seed": 304412, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4412, + "separate_rng_streams": true, + "initialization_seed": 4412, + "environment_seed": 6441200, + "policy_sampling_seed": 104412, + "minibatch_seed": 304412, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/connectome/seed_4413.json b/configs/figure4/connectome/seed_4413.json new file mode 100644 index 0000000000000000000000000000000000000000..b0e227b63567521cc36dc6716133c9f583061d53 --- /dev/null +++ b/configs/figure4/connectome/seed_4413.json @@ -0,0 +1,104 @@ +{ + "seed": 4413, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4413, + "separate_rng_streams": true, + "initialization_seed": 4413, + "environment_seed": 6441300, + "policy_sampling_seed": 104413, + "minibatch_seed": 304413, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4413, + "separate_rng_streams": true, + "initialization_seed": 4413, + "environment_seed": 6441300, + "policy_sampling_seed": 104413, + "minibatch_seed": 304413, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/connectome/seed_4414.json b/configs/figure4/connectome/seed_4414.json new file mode 100644 index 0000000000000000000000000000000000000000..180b91b055c32a82a899c10bd068dcb6436ba3cb --- /dev/null +++ b/configs/figure4/connectome/seed_4414.json @@ -0,0 +1,104 @@ +{ + "seed": 4414, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4414, + "separate_rng_streams": true, + "initialization_seed": 4414, + "environment_seed": 6441400, + "policy_sampling_seed": 104414, + "minibatch_seed": 304414, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4414, + "separate_rng_streams": true, + "initialization_seed": 4414, + "environment_seed": 6441400, + "policy_sampling_seed": 104414, + "minibatch_seed": 304414, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/connectome/seed_4415.json b/configs/figure4/connectome/seed_4415.json new file mode 100644 index 0000000000000000000000000000000000000000..6fa96bbccc5d77e060736e2546bf348f5ed51e7f --- /dev/null +++ b/configs/figure4/connectome/seed_4415.json @@ -0,0 +1,104 @@ +{ + "seed": 4415, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4415, + "separate_rng_streams": true, + "initialization_seed": 4415, + "environment_seed": 6441500, + "policy_sampling_seed": 104415, + "minibatch_seed": 304415, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4415, + "separate_rng_streams": true, + "initialization_seed": 4415, + "environment_seed": 6441500, + "policy_sampling_seed": 104415, + "minibatch_seed": 304415, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/connectome/seed_4416.json b/configs/figure4/connectome/seed_4416.json new file mode 100644 index 0000000000000000000000000000000000000000..284664a7332969a0becfa2bb6f23ce8f3466410c --- /dev/null +++ b/configs/figure4/connectome/seed_4416.json @@ -0,0 +1,104 @@ +{ + "seed": 4416, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4416, + "separate_rng_streams": true, + "initialization_seed": 4416, + "environment_seed": 6441600, + "policy_sampling_seed": 104416, + "minibatch_seed": 304416, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4416, + "separate_rng_streams": true, + "initialization_seed": 4416, + "environment_seed": 6441600, + "policy_sampling_seed": 104416, + "minibatch_seed": 304416, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 25600, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "lr_floor_20261010" +} diff --git a/configs/figure4/connectome/seed_4417.json b/configs/figure4/connectome/seed_4417.json new file mode 100644 index 0000000000000000000000000000000000000000..4a730c2c832af9b2cbc75a6c6390535c42e3234b --- /dev/null +++ b/configs/figure4/connectome/seed_4417.json @@ -0,0 +1,104 @@ +{ + "seed": 4417, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4417, + "separate_rng_streams": true, + "initialization_seed": 4417, + "environment_seed": 6441700, + "policy_sampling_seed": 104417, + "minibatch_seed": 304417, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4417, + "separate_rng_streams": true, + "initialization_seed": 4417, + "environment_seed": 6441700, + "policy_sampling_seed": 104417, + "minibatch_seed": 304417, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 25600, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "lr_floor_20261010" +} diff --git a/configs/figure4/connectome/seed_4418.json b/configs/figure4/connectome/seed_4418.json new file mode 100644 index 0000000000000000000000000000000000000000..c5e200c9a8ee983dc73e9aba25e2f4c5e2f4462c --- /dev/null +++ b/configs/figure4/connectome/seed_4418.json @@ -0,0 +1,104 @@ +{ + "seed": 4418, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4418, + "separate_rng_streams": true, + "initialization_seed": 4418, + "environment_seed": 6441800, + "policy_sampling_seed": 104418, + "minibatch_seed": 304418, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4418, + "separate_rng_streams": true, + "initialization_seed": 4418, + "environment_seed": 6441800, + "policy_sampling_seed": 104418, + "minibatch_seed": 304418, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/connectome/seed_4419.json b/configs/figure4/connectome/seed_4419.json new file mode 100644 index 0000000000000000000000000000000000000000..588b3c22295e1e9ac35404fb8995b6e584b877ae --- /dev/null +++ b/configs/figure4/connectome/seed_4419.json @@ -0,0 +1,104 @@ +{ + "seed": 4419, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4419, + "separate_rng_streams": true, + "initialization_seed": 4419, + "environment_seed": 6441900, + "policy_sampling_seed": 104419, + "minibatch_seed": 304419, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4419, + "separate_rng_streams": true, + "initialization_seed": 4419, + "environment_seed": 6441900, + "policy_sampling_seed": 104419, + "minibatch_seed": 304419, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 25600, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "lr_floor_20261010" +} diff --git a/configs/figure4/connectome/seed_4420.json b/configs/figure4/connectome/seed_4420.json new file mode 100644 index 0000000000000000000000000000000000000000..19196345eb31b4e3cdf01f0c641a94b2ba01a3a4 --- /dev/null +++ b/configs/figure4/connectome/seed_4420.json @@ -0,0 +1,104 @@ +{ + "seed": 4420, + "method": "connectome", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4420, + "separate_rng_streams": true, + "initialization_seed": 4420, + "environment_seed": 6442000, + "policy_sampling_seed": 104420, + "minibatch_seed": 304420, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4420, + "separate_rng_streams": true, + "initialization_seed": 4420, + "environment_seed": 6442000, + "policy_sampling_seed": 104420, + "minibatch_seed": 304420, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/mlp/seed_4401.json b/configs/figure4/mlp/seed_4401.json new file mode 100644 index 0000000000000000000000000000000000000000..4cb3948db4025f8b7f832ce1a119d5d264309fa7 --- /dev/null +++ b/configs/figure4/mlp/seed_4401.json @@ -0,0 +1,104 @@ +{ + "seed": 4401, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4401, + "separate_rng_streams": true, + "initialization_seed": 4401, + "environment_seed": 6440100, + "policy_sampling_seed": 104401, + "minibatch_seed": 304401, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4401, + "separate_rng_streams": true, + "initialization_seed": 4401, + "environment_seed": 6440100, + "policy_sampling_seed": 104401, + "minibatch_seed": 304401, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "1514da1d5dbd96fc3006695fc4654c6a98b437d73c9e3b079dcd53530c5538bf", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 25600, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "lr_floor_20261010" +} diff --git a/configs/figure4/mlp/seed_4402.json b/configs/figure4/mlp/seed_4402.json new file mode 100644 index 0000000000000000000000000000000000000000..915a8e7ff86acc2a3eb2cfaa00e5328577b54e6f --- /dev/null +++ b/configs/figure4/mlp/seed_4402.json @@ -0,0 +1,104 @@ +{ + "seed": 4402, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4402, + "separate_rng_streams": true, + "initialization_seed": 4402, + "environment_seed": 6440200, + "policy_sampling_seed": 104402, + "minibatch_seed": 304402, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4402, + "separate_rng_streams": true, + "initialization_seed": 4402, + "environment_seed": 6440200, + "policy_sampling_seed": 104402, + "minibatch_seed": 304402, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "1514da1d5dbd96fc3006695fc4654c6a98b437d73c9e3b079dcd53530c5538bf", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/mlp/seed_4403.json b/configs/figure4/mlp/seed_4403.json new file mode 100644 index 0000000000000000000000000000000000000000..27078cf9133792b0ab5caff4d18a5c2ce7a4ef49 --- /dev/null +++ b/configs/figure4/mlp/seed_4403.json @@ -0,0 +1,105 @@ +{ + "seed": 4403, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4403, + "separate_rng_streams": true, + "initialization_seed": 4403, + "environment_seed": 6440300, + "policy_sampling_seed": 104403, + "minibatch_seed": 304403, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4403, + "separate_rng_streams": true, + "initialization_seed": 4403, + "environment_seed": 6440300, + "policy_sampling_seed": 104403, + "minibatch_seed": 304403, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "1514da1d5dbd96fc3006695fc4654c6a98b437d73c9e3b079dcd53530c5538bf", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "learning_rate_matched_20261016", + "previous_origin_stage": "matched_budget_20261010" +} diff --git a/configs/figure4/mlp/seed_4404.json b/configs/figure4/mlp/seed_4404.json new file mode 100644 index 0000000000000000000000000000000000000000..2a4d9627800c9423353d850983deb1c97f8a65e2 --- /dev/null +++ b/configs/figure4/mlp/seed_4404.json @@ -0,0 +1,104 @@ +{ + "seed": 4404, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4404, + "separate_rng_streams": true, + "initialization_seed": 4404, + "environment_seed": 6440400, + "policy_sampling_seed": 104404, + "minibatch_seed": 304404, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4404, + "separate_rng_streams": true, + "initialization_seed": 4404, + "environment_seed": 6440400, + "policy_sampling_seed": 104404, + "minibatch_seed": 304404, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "1514da1d5dbd96fc3006695fc4654c6a98b437d73c9e3b079dcd53530c5538bf", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/mlp/seed_4405.json b/configs/figure4/mlp/seed_4405.json new file mode 100644 index 0000000000000000000000000000000000000000..d7c84d3140e6aca9655dc1208db715c22beed350 --- /dev/null +++ b/configs/figure4/mlp/seed_4405.json @@ -0,0 +1,104 @@ +{ + "seed": 4405, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4405, + "separate_rng_streams": true, + "initialization_seed": 4405, + "environment_seed": 6440500, + "policy_sampling_seed": 104405, + "minibatch_seed": 304405, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4405, + "separate_rng_streams": true, + "initialization_seed": 4405, + "environment_seed": 6440500, + "policy_sampling_seed": 104405, + "minibatch_seed": 304405, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "1514da1d5dbd96fc3006695fc4654c6a98b437d73c9e3b079dcd53530c5538bf", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/mlp/seed_4406.json b/configs/figure4/mlp/seed_4406.json new file mode 100644 index 0000000000000000000000000000000000000000..fd9affc1a6f2073f9493676952d37cfc2228dfd5 --- /dev/null +++ b/configs/figure4/mlp/seed_4406.json @@ -0,0 +1,105 @@ +{ + "seed": 4406, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4406, + "separate_rng_streams": true, + "initialization_seed": 4406, + "environment_seed": 6440600, + "policy_sampling_seed": 104406, + "minibatch_seed": 304406, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4406, + "separate_rng_streams": true, + "initialization_seed": 4406, + "environment_seed": 6440600, + "policy_sampling_seed": 104406, + "minibatch_seed": 304406, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "1514da1d5dbd96fc3006695fc4654c6a98b437d73c9e3b079dcd53530c5538bf", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "learning_rate_matched_20261016", + "previous_origin_stage": "matched_budget_20261010" +} diff --git a/configs/figure4/mlp/seed_4407.json b/configs/figure4/mlp/seed_4407.json new file mode 100644 index 0000000000000000000000000000000000000000..587475693de40749f7d6c9ba778f1f7bb9d1cfb9 --- /dev/null +++ b/configs/figure4/mlp/seed_4407.json @@ -0,0 +1,105 @@ +{ + "seed": 4407, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4407, + "separate_rng_streams": true, + "initialization_seed": 4407, + "environment_seed": 6440700, + "policy_sampling_seed": 104407, + "minibatch_seed": 304407, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4407, + "separate_rng_streams": true, + "initialization_seed": 4407, + "environment_seed": 6440700, + "policy_sampling_seed": 104407, + "minibatch_seed": 304407, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "learning_rate_matched_20261016", + "previous_origin_stage": "matched_budget_20261010" +} diff --git a/configs/figure4/mlp/seed_4408.json b/configs/figure4/mlp/seed_4408.json new file mode 100644 index 0000000000000000000000000000000000000000..7c15160ded2f6bd43a23aec4cc929d72b96c3054 --- /dev/null +++ b/configs/figure4/mlp/seed_4408.json @@ -0,0 +1,105 @@ +{ + "seed": 4408, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4408, + "separate_rng_streams": true, + "initialization_seed": 4408, + "environment_seed": 6440800, + "policy_sampling_seed": 104408, + "minibatch_seed": 304408, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4408, + "separate_rng_streams": true, + "initialization_seed": 4408, + "environment_seed": 6440800, + "policy_sampling_seed": 104408, + "minibatch_seed": 304408, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "learning_rate_matched_20261016", + "previous_origin_stage": "matched_budget_20261010" +} diff --git a/configs/figure4/mlp/seed_4409.json b/configs/figure4/mlp/seed_4409.json new file mode 100644 index 0000000000000000000000000000000000000000..41aebe58259fab0f8a0d53186463879e87d60165 --- /dev/null +++ b/configs/figure4/mlp/seed_4409.json @@ -0,0 +1,104 @@ +{ + "seed": 4409, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4409, + "separate_rng_streams": true, + "initialization_seed": 4409, + "environment_seed": 6440900, + "policy_sampling_seed": 104409, + "minibatch_seed": 304409, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4409, + "separate_rng_streams": true, + "initialization_seed": 4409, + "environment_seed": 6440900, + "policy_sampling_seed": 104409, + "minibatch_seed": 304409, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/mlp/seed_4410.json b/configs/figure4/mlp/seed_4410.json new file mode 100644 index 0000000000000000000000000000000000000000..ea7398618ee05dbf5691b5a7cebff854f9cf2fa5 --- /dev/null +++ b/configs/figure4/mlp/seed_4410.json @@ -0,0 +1,104 @@ +{ + "seed": 4410, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4410, + "separate_rng_streams": true, + "initialization_seed": 4410, + "environment_seed": 6441000, + "policy_sampling_seed": 104410, + "minibatch_seed": 304410, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4410, + "separate_rng_streams": true, + "initialization_seed": 4410, + "environment_seed": 6441000, + "policy_sampling_seed": 104410, + "minibatch_seed": 304410, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/mlp/seed_4411.json b/configs/figure4/mlp/seed_4411.json new file mode 100644 index 0000000000000000000000000000000000000000..8d29a8ea2947ac49bf3f0541499ff68ec7205576 --- /dev/null +++ b/configs/figure4/mlp/seed_4411.json @@ -0,0 +1,104 @@ +{ + "seed": 4411, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4411, + "separate_rng_streams": true, + "initialization_seed": 4411, + "environment_seed": 6441100, + "policy_sampling_seed": 104411, + "minibatch_seed": 304411, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4411, + "separate_rng_streams": true, + "initialization_seed": 4411, + "environment_seed": 6441100, + "policy_sampling_seed": 104411, + "minibatch_seed": 304411, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/mlp/seed_4412.json b/configs/figure4/mlp/seed_4412.json new file mode 100644 index 0000000000000000000000000000000000000000..725c881c7228670d38a855844b1a4e5bbc73ab04 --- /dev/null +++ b/configs/figure4/mlp/seed_4412.json @@ -0,0 +1,104 @@ +{ + "seed": 4412, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4412, + "separate_rng_streams": true, + "initialization_seed": 4412, + "environment_seed": 6441200, + "policy_sampling_seed": 104412, + "minibatch_seed": 304412, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4412, + "separate_rng_streams": true, + "initialization_seed": 4412, + "environment_seed": 6441200, + "policy_sampling_seed": 104412, + "minibatch_seed": 304412, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/mlp/seed_4413.json b/configs/figure4/mlp/seed_4413.json new file mode 100644 index 0000000000000000000000000000000000000000..2bbfcca804fea615c1054b8dd594ca6816fa074c --- /dev/null +++ b/configs/figure4/mlp/seed_4413.json @@ -0,0 +1,104 @@ +{ + "seed": 4413, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4413, + "separate_rng_streams": true, + "initialization_seed": 4413, + "environment_seed": 6441300, + "policy_sampling_seed": 104413, + "minibatch_seed": 304413, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4413, + "separate_rng_streams": true, + "initialization_seed": 4413, + "environment_seed": 6441300, + "policy_sampling_seed": 104413, + "minibatch_seed": 304413, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/mlp/seed_4414.json b/configs/figure4/mlp/seed_4414.json new file mode 100644 index 0000000000000000000000000000000000000000..57df47b807b7d42f58138d2060e5ffe047b1b541 --- /dev/null +++ b/configs/figure4/mlp/seed_4414.json @@ -0,0 +1,104 @@ +{ + "seed": 4414, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4414, + "separate_rng_streams": true, + "initialization_seed": 4414, + "environment_seed": 6441400, + "policy_sampling_seed": 104414, + "minibatch_seed": 304414, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4414, + "separate_rng_streams": true, + "initialization_seed": 4414, + "environment_seed": 6441400, + "policy_sampling_seed": 104414, + "minibatch_seed": 304414, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/mlp/seed_4415.json b/configs/figure4/mlp/seed_4415.json new file mode 100644 index 0000000000000000000000000000000000000000..93905bb466302ddc41f444b7b3b19bcd633eceb9 --- /dev/null +++ b/configs/figure4/mlp/seed_4415.json @@ -0,0 +1,104 @@ +{ + "seed": 4415, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4415, + "separate_rng_streams": true, + "initialization_seed": 4415, + "environment_seed": 6441500, + "policy_sampling_seed": 104415, + "minibatch_seed": 304415, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4415, + "separate_rng_streams": true, + "initialization_seed": 4415, + "environment_seed": 6441500, + "policy_sampling_seed": 104415, + "minibatch_seed": 304415, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/mlp/seed_4416.json b/configs/figure4/mlp/seed_4416.json new file mode 100644 index 0000000000000000000000000000000000000000..e926c30d0015ce3cf0b1383c4f929d1f7cd3a096 --- /dev/null +++ b/configs/figure4/mlp/seed_4416.json @@ -0,0 +1,105 @@ +{ + "seed": 4416, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4416, + "separate_rng_streams": true, + "initialization_seed": 4416, + "environment_seed": 6441600, + "policy_sampling_seed": 104416, + "minibatch_seed": 304416, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4416, + "separate_rng_streams": true, + "initialization_seed": 4416, + "environment_seed": 6441600, + "policy_sampling_seed": 104416, + "minibatch_seed": 304416, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "learning_rate_matched_20261016", + "previous_origin_stage": "matched_budget_20261010" +} diff --git a/configs/figure4/mlp/seed_4417.json b/configs/figure4/mlp/seed_4417.json new file mode 100644 index 0000000000000000000000000000000000000000..50483ac327129ac6b3aa6fefe0ea6b3772dd5f35 --- /dev/null +++ b/configs/figure4/mlp/seed_4417.json @@ -0,0 +1,105 @@ +{ + "seed": 4417, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4417, + "separate_rng_streams": true, + "initialization_seed": 4417, + "environment_seed": 6441700, + "policy_sampling_seed": 104417, + "minibatch_seed": 304417, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4417, + "separate_rng_streams": true, + "initialization_seed": 4417, + "environment_seed": 6441700, + "policy_sampling_seed": 104417, + "minibatch_seed": 304417, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "learning_rate_matched_20261016", + "previous_origin_stage": "matched_budget_20261010" +} diff --git a/configs/figure4/mlp/seed_4418.json b/configs/figure4/mlp/seed_4418.json new file mode 100644 index 0000000000000000000000000000000000000000..6228bebedc4d021ae8285925dae73260018bf19f --- /dev/null +++ b/configs/figure4/mlp/seed_4418.json @@ -0,0 +1,104 @@ +{ + "seed": 4418, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4418, + "separate_rng_streams": true, + "initialization_seed": 4418, + "environment_seed": 6441800, + "policy_sampling_seed": 104418, + "minibatch_seed": 304418, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4418, + "separate_rng_streams": true, + "initialization_seed": 4418, + "environment_seed": 6441800, + "policy_sampling_seed": 104418, + "minibatch_seed": 304418, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/configs/figure4/mlp/seed_4419.json b/configs/figure4/mlp/seed_4419.json new file mode 100644 index 0000000000000000000000000000000000000000..29753735993655e18d6be869dd24754c94533137 --- /dev/null +++ b/configs/figure4/mlp/seed_4419.json @@ -0,0 +1,105 @@ +{ + "seed": 4419, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4419, + "separate_rng_streams": true, + "initialization_seed": 4419, + "environment_seed": 6441900, + "policy_sampling_seed": 104419, + "minibatch_seed": 304419, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4419, + "separate_rng_streams": true, + "initialization_seed": 4419, + "environment_seed": 6441900, + "policy_sampling_seed": 104419, + "minibatch_seed": 304419, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "learning_rate_matched_20261016", + "previous_origin_stage": "matched_budget_20261010" +} diff --git a/configs/figure4/mlp/seed_4420.json b/configs/figure4/mlp/seed_4420.json new file mode 100644 index 0000000000000000000000000000000000000000..7af35e0070ef09528f2bd171cb173d51bb785c33 --- /dev/null +++ b/configs/figure4/mlp/seed_4420.json @@ -0,0 +1,104 @@ +{ + "seed": 4420, + "method": "mlp", + "base_ppo": { + "total_timesteps": 12800, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4420, + "separate_rng_streams": true, + "initialization_seed": 4420, + "environment_seed": 6442000, + "policy_sampling_seed": 104420, + "minibatch_seed": 304420, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "continuation_ppo": { + "total_timesteps": 15872, + "n_envs": 8, + "n_steps": 32, + "seq_len": 8, + "n_epochs": 4, + "n_minibatches": 8, + "gamma": 0.9801986733067554, + "lam": 0.8187307530779819, + "clip_range": 0.2, + "vf_clip_range": 0.2, + "lr": 0.0001, + "actor_lr": 0.0001, + "critic_lr": 0.0001, + "log_std_lr": 0.001, + "vf_coef": 0.5, + "ent_coef": 0.0, + "max_grad_norm": 0.5, + "separate_actor_critic_grad_clipping": true, + "target_kl": 0.03, + "normalize_reward": true, + "normalize_observation": false, + "observation_clip": 10.0, + "log_interval": 1, + "seed": 4420, + "separate_rng_streams": true, + "initialization_seed": 4420, + "environment_seed": 6442000, + "policy_sampling_seed": 104420, + "minibatch_seed": 304420, + "initial_hidden_mode": "zero", + "rollout_heartbeat": 32, + "best_checkpoint_mode": "return", + "best_checkpoint_window": 100, + "best_checkpoint_min_episodes": 10, + "best_checkpoint_velocity_key": "mean_x_velocity", + "best_checkpoint_min_forward_velocity": 0.0, + "best_checkpoint_max_high_z_termination_rate": 1.0, + "best_checkpoint_max_inverted_completion_rate": 1.0, + "best_checkpoint_max_functional_failure_rate": 1.0 + }, + "base_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-05, + "max_lr": 0.001 + }, + "continuation_adaptive_policy_kl": { + "target": 0.015, + "factor": 1.5, + "min_lr": 1e-06, + "max_lr": 0.001 + }, + "base_protocol_hash": "8fedbec32a824c6acc018cc99b534957af52b2f76a0991a2c45cb5f6ca34e4e1", + "base_steps": 12800, + "final_steps": 15872, + "original_requested_final_steps": 15872, + "resume": "Restore optimizer, policy and RNG; reset physics once at update50, then run12updates.", + "origin_stage": "short_budget_twenty_20261010" +} diff --git a/figure_map.csv b/figure_map.csv new file mode 100644 index 0000000000000000000000000000000000000000..f693da6fb757d088ff63d815e99b9c458b965ce6 --- /dev/null +++ b/figure_map.csv @@ -0,0 +1,8 @@ +paper_item,data_archive,model_id_prefix,checkpoint_files,historical_data_location +figure2,archives/figure2.tar.gz,figure2/,60,data/figure2 +figure3,archives/figure3.tar.gz,figure3/,72,data/figure3 +figure4,archives/figure4.tar.gz,figure4/,40,data/figure4 +S3,archives/S3.tar.gz,S3/,12,data/supplementary/development +S4,archives/S4.tar.gz,S4/,12,data/supplementary/development +S5,archives/S5.tar.gz,S5/,24,data/supplementary/s06 +S6,archives/S6.tar.gz,S6/,6,data/supplementary/s07 diff --git a/load_policy.py b/load_policy.py new file mode 100644 index 0000000000000000000000000000000000000000..90433a1a5058d702700c61783a2eba0244ad8113 --- /dev/null +++ b/load_policy.py @@ -0,0 +1,98 @@ +"""Load the archived policy with the matching FlyControl engine (no simulator).""" +import argparse +import hashlib +import json +import os +from pathlib import Path +import sys + + +def load_policy(model_id, model_dir, code_dir, data_dir=None, device='cpu'): + import torch + import yaml + + model_dir, code_dir = Path(model_dir).resolve(), Path(code_dir).resolve() + entries = json.loads((model_dir / 'model_index.json').read_text())['models'] + entry = next(x for x in entries if x['id'] == model_id) + checkpoint = model_dir / entry['path'] + hasher = hashlib.sha256() + with checkpoint.open('rb') as stream: + for chunk in iter(lambda: stream.read(8 * 1024 * 1024), b''): + hasher.update(chunk) + digest = hasher.hexdigest() + if digest != entry['sha256']: + raise ValueError('Checkpoint checksum mismatch: ' + model_id) + # These are original research checkpoints, including legacy training state. + saved = torch.load(checkpoint, map_location='cpu', weights_only=False) + state = saved['policy_state'] + engine = code_dir / 'src/flycontrol/engines' + for folder in [code_dir / 'src', engine / 'common', engine / 'terrain', engine / 'supplemental']: + sys.path.insert(0, str(folder)) + os.environ.setdefault('FLYCONTROL_RUN', str(Path(data_dir or '.').resolve())) + group, method = entry['paper_item'], entry['method'] + if group.startswith('figure'): + if data_dir is None: + raise ValueError('Main-figure policies require the extracted dataset --data-dir') + obs_dim, action_dim = {'figure2': (688, 6), 'figure3': (64, 2), 'figure4': (51, 2)}[group] + root = Path(data_dir) / 'data' / group + graph = root / (f'er_inputs/er_seed{entry["seed"]}.npz' if method == 'er' else 'graph.npz') + if group == 'figure2': + from terrain_control.policy import build_policy + else: + from fly_control.policy import build_policy + policy = build_policy(graph, obs_dim, action_dim, 'connectome' if method == 'er' else method, + entry['seed'], device, 'normalized', -1.0) + if group in ('figure3', 'figure4') and method == 'connectome': + policy.actor.core.dt = 0.04 + policy.actor.core.internal_substeps = 4 + # FixedGraph's nonpersistent reduction buffers must match these edges. + if method != 'mlp': + for key in ('edge_pre', 'edge_post', 'edge_weight'): + if not torch.equal(getattr(policy.actor.core, key).cpu(), state['actor.core.' + key]): + raise ValueError('Checkpoint and graph differ: ' + key) + else: + from larva.locomotion.connectome import ConnectomeData + from larva.locomotion.models import build_policy + config = saved.get('config') or yaml.safe_load((model_dir / entry['config']).read_text()) + cfg = config['policy'] + inp, out = state['actor.core.input_proj.weight'], state['actor.core.output_layer.weight'] + sensory, obs_dim = inp.shape + action_dim, output = out.shape + n = 2952 if method == 'larva' else 138584 if method == 'adult' else 112885 + dense = state.get('actor.core.W') + edges = [state.get('actor.core.' + k) for k in ('edge_pre', 'edge_post', 'edge_weight')] + graph = ConnectomeData(None if dense is None else dense.numpy(), sensory, n-sensory-output, + output, (), (), (), *[None if x is None else x.numpy() for x in edges], + metadata={'source': 'checkpoint', 'root_ids_available': False}) + policy = build_policy(cfg['name'], obs_dim, action_dim, actor_kwargs=cfg.get('actor_kwargs'), + value_hidden=cfg.get('value_hidden'), + value_activation=cfg.get('value_activation', 'tanh'), + log_std_init=cfg.get('log_std_init', -0.5), connectome=graph).to(device) + policy.load_state_dict(state, strict=True) + policy.eval() + for key, value in policy.state_dict().items(): + if not torch.equal(value.cpu(), state[key]): + raise ValueError('Loaded state differs: ' + key) + return policy, {'entry': entry, 'obs_dim': obs_dim, 'action_dim': action_dim, + 'obs_rms_state': saved.get('obs_rms_state')} + + +if __name__ == '__main__': + parser = argparse.ArgumentParser(description=__doc__) + parser.add_argument('--model-id', required=True) + parser.add_argument('--model-dir', default=str(Path(__file__).parent)) + parser.add_argument('--code-dir', required=True) + parser.add_argument('--data-dir') + args = parser.parse_args() + import torch + torch.set_num_threads(1) + torch.manual_seed(0) + policy, meta = load_policy(args.model_id, args.model_dir, args.code_dir, args.data_dir) + obs = torch.zeros(1, meta['obs_dim']) + hidden = policy.init_state(1, torch.device('cpu')) + with torch.no_grad(): + mean, hidden, value = policy(obs, hidden) + assert all(torch.isfinite(x).all() for x in (mean, hidden, value)) + print(json.dumps({'model_id': args.model_id, 'strict_load': True, + 'finite_forward': True, 'observation_dim': meta['obs_dim'], + 'action_dim': meta['action_dim']})) diff --git a/model_index.json b/model_index.json new file mode 100644 index 0000000000000000000000000000000000000000..c6b326c581dc57d2874a7473773a6def92089f03 --- /dev/null +++ b/model_index.json @@ -0,0 +1,4990 @@ +{ + "scope": [ + "figure2", + "figure3", + "figure4", + "S3", + "S4", + "S5", + "S6" + ], + "models": [ + { + "id": "figure2/connectome/seed_2011/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2011, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2011/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2011/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2011.json", + "config": "configs/figure2/connectome/seed_2011.json", + "sha256": "2cb0aa0cd82fa215858a86e345ab430d13de4a97cebe13f504ca3b819cd194ae", + "bytes": 77244141, + "config_sha256": "5746e0a4a1ac2ff082473139caee81dcd66bb4cfce842e38d9130fb7a5f1540e", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2011/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2011, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2011/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2011/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2011.json", + "config": "configs/figure2/mlp/seed_2011.json", + "sha256": "fe11d0988013d624dca67dcba4ad0b0e90ef93e24c8f5dcadbd9d85abd7bc853", + "bytes": 2593549, + "config_sha256": "179835689f6bce89e7e1c14d2dac2a0a88a10361d4e0d59797d3c34b952834a3", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2012/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2012, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2012/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2012/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2012.json", + "config": "configs/figure2/connectome/seed_2012.json", + "sha256": "406a20c02ea272dd9e6c5cc13193bb1eb3beb23293ee17284efa57a03e4f9a32", + "bytes": 77244141, + "config_sha256": "0c3fe10477f6d6a0da9e6d0cf89e54162e79d33685ea3335063604f69a37b990", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2012/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2012, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2012/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2012/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2012.json", + "config": "configs/figure2/mlp/seed_2012.json", + "sha256": "34b0094d197bbc07489ffdc2ecd47b73d9f1c9bd671003c03845bf1832d5eb7f", + "bytes": 2593549, + "config_sha256": "9ab6654ca9e2fc5e41aa1379b7503dbb845c7aa638c9e0884933beef6475d303", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2013/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2013, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2013/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2013/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2013.json", + "config": "configs/figure2/connectome/seed_2013.json", + "sha256": "5b85aef5f6a03e1d202464f9bc190987f7bab0ac91e9b8b667dcbfc84efef6c1", + "bytes": 77244141, + "config_sha256": "740267b0f9e1f716cd76ba732ae1beb5cf1017b43874645505b38377c93f43f9", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2013/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2013, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2013/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2013/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2013.json", + "config": "configs/figure2/mlp/seed_2013.json", + "sha256": "5026b33dcfd91bac7b0b1c09f717fa90835a6f81f7d82e19fbda3315e5ac91d5", + "bytes": 2593549, + "config_sha256": "1f5ac45bac96fb12373474285eb7845219812d10d68a6c4eacd9ebc24a4a15de", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2014/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2014, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2014/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2014/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2014.json", + "config": "configs/figure2/connectome/seed_2014.json", + "sha256": "630fd1e978de57cac762b5a607cee84d03e2afeb4004b543dbb5f2c5b3a80c13", + "bytes": 77244141, + "config_sha256": "72cdc988f2f761ef0dcae1c0edaa89126f68e507e9cb9d27aea1d1a35e200aa4", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2014/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2014, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2014/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2014/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2014.json", + "config": "configs/figure2/mlp/seed_2014.json", + "sha256": "05e4ae0abea0e8d0c2d6f04ba469580c7acd9a254511be328f02234fbe2475ab", + "bytes": 2593549, + "config_sha256": "5a8f59e3dbf687df6da30b536865550de1a104e8e7e0cf8dfd723061c4d0b739", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2015/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2015, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2015/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2015/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2015.json", + "config": "configs/figure2/connectome/seed_2015.json", + "sha256": "1a45be7c2aa221a050cc8b60c6e70964060f9b7037d627fa16cf61ae8bcbd713", + "bytes": 77244141, + "config_sha256": "76095c80759d74e109c6c698d4d38030c244ba2a7de0cee2b353e3478b7ef11c", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2015/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2015, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2015/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2015/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2015.json", + "config": "configs/figure2/mlp/seed_2015.json", + "sha256": "830319390b07a9f6cfefe35edde1a58bdd1d6476c590a72fbe0eb2b66beaf018", + "bytes": 2593549, + "config_sha256": "5af8f98e1f475956560a1a2322499e6551ea1b4820f4113f86fde885bf89e3ba", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2016/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2016, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2016/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2016/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2016.json", + "config": "configs/figure2/connectome/seed_2016.json", + "sha256": "1b6f677e2328f9885e3c71d94fc2ca33c4723fbf47b1e849409d4c1336c49218", + "bytes": 77244141, + "config_sha256": "7560dd3cf2136990acb257a0c1e3701fd883eea88306f9c44fd2f54555df7283", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2016/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2016, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2016/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2016/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2016.json", + "config": "configs/figure2/mlp/seed_2016.json", + "sha256": "64ea22299c0a25ed9f41006934ebe4c08eca7d5c15378a7f75173c13ea858757", + "bytes": 2593549, + "config_sha256": "bde6fd055c57bb7187c11ea76fc0d12978815c0ae6c5940b40eb6dc52e0d2cd2", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2017/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2017, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2017/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2017/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2017.json", + "config": "configs/figure2/connectome/seed_2017.json", + "sha256": "51632081b08b6cfde6d890c0b968fd37e35aaea86c97ebb31a36f458bf37426b", + "bytes": 77244141, + "config_sha256": "7c07c50d97b54eb01f93fe6bf134fd8154018894b70baf0bcdcc24b3675b8e42", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2017/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2017, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2017/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2017/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2017.json", + "config": "configs/figure2/mlp/seed_2017.json", + "sha256": "10999a20b6ccb3e5576434a666573716e13dc0cd188eb32df296fb83b6791a86", + "bytes": 2593549, + "config_sha256": "fc4983a319d46027ec4fc7680b7bedb80cfcbcb3401cfe240fa69216bc195fd6", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2018/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2018, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2018/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2018/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2018.json", + "config": "configs/figure2/connectome/seed_2018.json", + "sha256": "e7a35e8515b32831656dc7343a3064719e1648f352ceb508cfe57045d9506036", + "bytes": 77244141, + "config_sha256": "0d597202f7ca31ce40f9e4574a7be21ab47936329d07bae068620791eab092d9", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2018/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2018, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2018/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2018/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2018.json", + "config": "configs/figure2/mlp/seed_2018.json", + "sha256": "2300b45eb62614e4d14a369ac776528ed1c1233abb38a03c9e357b843a1ec5f0", + "bytes": 2593549, + "config_sha256": "850ecf408180a6176a2350a7872be3bd10350f303325822902026c70b47e1b11", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2019/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2019, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2019/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2019/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2019.json", + "config": "configs/figure2/connectome/seed_2019.json", + "sha256": "f51f3cdd91c657668f73961d0c9aca2df8610e7db430f776329524edc2bf142c", + "bytes": 77244141, + "config_sha256": "a318a0fc3b417b5821de0d7a9bda6c940cf858431cffaaf5fa2e54b1bf7b2983", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2019/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2019, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2019/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2019/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2019.json", + "config": "configs/figure2/mlp/seed_2019.json", + "sha256": "67e40eebf10f46235e9914361f16b9889d3df4645faaee20a1585774c691489d", + "bytes": 2593549, + "config_sha256": "20afc9c7322f12bc389b7c3fe12bb708794a0788d526d8fc850abe8e8d62ef43", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2020/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2020, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2020/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2020/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2020.json", + "config": "configs/figure2/connectome/seed_2020.json", + "sha256": "065e72465366e81e66984c9e06a544f65e8fc9898fd5df8ce109da82a7bcbec8", + "bytes": 77244141, + "config_sha256": "10ffc53a4196beaf252b11bfe3beec23881ca4490751b52f2796342c7b0a1cad", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2020/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2020, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2020/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2020/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2020.json", + "config": "configs/figure2/mlp/seed_2020.json", + "sha256": "2b15d05104770b51d0c6bf63ad37a905e878d577fca87ff09fbf73cc8ded9394", + "bytes": 2593549, + "config_sha256": "2a59d32b780a075ed37fa1a8dd002bed05df1035c0726a8f8b90bda8b3b81e95", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2021/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2021, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2021/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2021/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2021.json", + "config": "configs/figure2/connectome/seed_2021.json", + "sha256": "e75069d5cc70f43e3767438121ca25836073ca936a5c76a333eb8fe7d78a16d7", + "bytes": 77244141, + "config_sha256": "b633c1f417ad39af7e5ade42e4eef7b1c55ed274b40a37736d341f1edd4ea2b6", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2021/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2021, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2021/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2021/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2021.json", + "config": "configs/figure2/mlp/seed_2021.json", + "sha256": "4c441d16ae664d1fc2e4fbfd6758f53951a30ce958c33082ddca01d314af64fd", + "bytes": 2593549, + "config_sha256": "5cf81c2ece306c4e3e81ba3bbc7e67f6267b3d382a19d9b4fb45433fc28a9376", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2022/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2022, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2022/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2022/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2022.json", + "config": "configs/figure2/connectome/seed_2022.json", + "sha256": "370607f06b475769229bc1715aafca7bd9075b599f50c8a9f0d0843c35010b51", + "bytes": 77244141, + "config_sha256": "770608dc793af911f7f218a99591e7450c04b5513a1ccde8dc64f4eafcdc3d54", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2022/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2022, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2022/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2022/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2022.json", + "config": "configs/figure2/mlp/seed_2022.json", + "sha256": "8813281170a649db1ed4f9a042e090ab9ff44a43845d9c688bbdd3481a624d3d", + "bytes": 2593549, + "config_sha256": "945d37fae201ab78f43ca5792f4a19cf65c215fb7656fe8be32fb35fe9947442", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2023/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2023, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2023/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2023/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2023.json", + "config": "configs/figure2/connectome/seed_2023.json", + "sha256": "1fc4cfb8d7e9844c1747a9a5b6f6de98def7e9bba32740d6dbc7eaa7a889160f", + "bytes": 77244141, + "config_sha256": "4e1411362af2414142f34488598e59893ea03b477b38e00bda24cc04ab273146", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2023/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2023, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2023/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2023/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2023.json", + "config": "configs/figure2/mlp/seed_2023.json", + "sha256": "63ac963d1cda3d36e93926ef0ca89713819e497f2190e12910fabe222f09f35b", + "bytes": 2593549, + "config_sha256": "02b17cf782a4f5fa73805a3efbc3b406fd4ba60f792c6704ce90ced787616a26", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2024/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2024, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2024/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2024/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2024.json", + "config": "configs/figure2/connectome/seed_2024.json", + "sha256": "3183bbd9ec2ee07e38abd5d5aaf36e02b5d98bf91e9557c4728ee466deac1b26", + "bytes": 77244141, + "config_sha256": "013e11e13f120ea0910f1e14295665df50aec43c6ec74711c311d4148b8903a3", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2024/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2024, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2024/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2024/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2024.json", + "config": "configs/figure2/mlp/seed_2024.json", + "sha256": "55556986ae3a3238c6145468ba46d4531aba5aa6ef5e32297c27c7aa96ee2408", + "bytes": 2593549, + "config_sha256": "da67e862d6794d76b391649c44a8d0afdf6c1f157fc396de0c445bb6be450be9", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2025/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2025, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2025/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2025/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2025.json", + "config": "configs/figure2/connectome/seed_2025.json", + "sha256": "a10fa9953bd640065955c621ac1d0bb743639987ce57cbc53294d8a9c2996744", + "bytes": 77244141, + "config_sha256": "d879b2abd1d002d1f8f0b56caadd429c134d1151eab55ab377060173139f5249", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2025/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2025, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2025/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2025/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2025.json", + "config": "configs/figure2/mlp/seed_2025.json", + "sha256": "a30735d645080ba4d9ab7e63e5fcebd00d2d2885b81e9d977f562c254f856213", + "bytes": 2593549, + "config_sha256": "c55e10f9d8f6deb82e68a69e2facb676ce101bca71481124a8d9fdbcbb5d94ec", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2026/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2026, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2026/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2026/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2026.json", + "config": "configs/figure2/connectome/seed_2026.json", + "sha256": "18583e4c699763368cc8869a909344b8b748b1c67fab0f1a3def80837895d673", + "bytes": 77244141, + "config_sha256": "de45d4b8227d2eb6aa8728a65cf9a2305e4556ac57c1c1e97c7987ec108a4485", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2026/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2026, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2026/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2026/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2026.json", + "config": "configs/figure2/mlp/seed_2026.json", + "sha256": "a2c2df5f71b28d8c59ed3a6539e0aa569101a3c3f5542a899b1c1432e6f4c183", + "bytes": 2593549, + "config_sha256": "7f143654a24a9bf628d3d01e2ea0b0b6e43ed29236bee63cd20307107299bdde", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2027/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2027, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2027/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2027/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2027.json", + "config": "configs/figure2/connectome/seed_2027.json", + "sha256": "be4cd921e7dac858fd55b340481134cb07103ae25adabca6e5c0e14920b9adde", + "bytes": 77244141, + "config_sha256": "1ac74c89eb6c523ade06adbf489ed2b28ae8d9b2cd188c3d60c21097ba98376c", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2027/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2027, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2027/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2027/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2027.json", + "config": "configs/figure2/mlp/seed_2027.json", + "sha256": "92cf42e5451dd021e944b9871347bb56e123e16d18e378c3025694550471e1a8", + "bytes": 2593549, + "config_sha256": "9e763da4ae6ac87603b4420a819023895a3891f51540e3e6fd013e8deac2e9d1", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2028/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2028, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2028/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2028/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2028.json", + "config": "configs/figure2/connectome/seed_2028.json", + "sha256": "6281b027e8f954150636fee4d58619f8f530d684c74d860f6be5c6ea6b229209", + "bytes": 77244141, + "config_sha256": "5826df4b8d371022b5ec05773a9549d8cf4d18a5c34e1cbfa4c5e1a49450f5b2", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2028/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2028, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2028/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2028/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2028.json", + "config": "configs/figure2/mlp/seed_2028.json", + "sha256": "0a6cadea13e78830ad6c85dfe6cc29d9a3ec46ef44d43ee8cf41b43c439b5b89", + "bytes": 2593549, + "config_sha256": "0543c2b248bf8c8d524446c1b86b66c34414864204b90dd7930f2f6178b9458f", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2029/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2029, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2029/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2029/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2029.json", + "config": "configs/figure2/connectome/seed_2029.json", + "sha256": "dbbf4e7d45f2fde367f28f112f1a9b9bc987006881d3f25c6f524a4eaf0f8a18", + "bytes": 77244141, + "config_sha256": "70cf9e939a4dbbabbdd90b40fbd4fcd52176f384c5e674c48b5908cfe1b3abb7", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2029/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2029, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2029/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2029/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2029.json", + "config": "configs/figure2/mlp/seed_2029.json", + "sha256": "cf262552d16613edc1e4b40f0e3641519f8b1720a9a06db30076951e2df442a4", + "bytes": 2593549, + "config_sha256": "0a0392ee9402d7d4de5c673f78e17016972ab37960fa4c1ed62fbabf9ea533ef", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/connectome/seed_2030/final", + "paper_item": "figure2", + "method": "connectome", + "seed": 2030, + "stage": "final", + "path": "checkpoints/figure2/connectome/seed_2030/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/connectome_seed2030/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/connectome_seed2030.json", + "config": "configs/figure2/connectome/seed_2030.json", + "sha256": "967e1af99f480eccd80579dbc3f282ec5c95c5574dc1bb1210a10b277dad4f80", + "bytes": 77244141, + "config_sha256": "20156444d5b791c716fdaf0dd8d531a634796763ff1edf1496bc014b3c074dbd", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/mlp/seed_2030/final", + "paper_item": "figure2", + "method": "mlp", + "seed": 2030, + "stage": "final", + "path": "checkpoints/figure2/mlp/seed_2030/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/mlp_seed2030/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/mlp_seed2030.json", + "config": "configs/figure2/mlp/seed_2030.json", + "sha256": "4f1a4c01e614d43e87cb0b08c471513db2b126ecb00937d5eb2898db3ea6db48", + "bytes": 2593549, + "config_sha256": "e7a35caeac1fdec7bdfb90254478f026d4588eef2a29814e2da40bde0448d932", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/graph.npz" + }, + { + "id": "figure2/er/seed_2011/final", + "paper_item": "figure2", + "method": "er", + "seed": 2011, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2011/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2011/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2011.json", + "config": "configs/figure2/er/seed_2011.json", + "sha256": "c6b1ee58d4b0b1a718a177b6fecd9f787dfd50056618c8b22187d6b22b9191e3", + "bytes": 77244141, + "config_sha256": "92a45120b73964ac766ed63fe124ec66844a7518ee4558fee27bb10b89e46a69", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2011.npz" + }, + { + "id": "figure2/er/seed_2012/final", + "paper_item": "figure2", + "method": "er", + "seed": 2012, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2012/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2012/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2012.json", + "config": "configs/figure2/er/seed_2012.json", + "sha256": "3a35973f31f34e653e37c457b2eda69fb5c1d82313fbcdea0abe512c149df746", + "bytes": 77244141, + "config_sha256": "a67cf35af075426857c7c2ef86cfafd16c1e87564a3c4065c6e14f488cbe95c6", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2012.npz" + }, + { + "id": "figure2/er/seed_2013/final", + "paper_item": "figure2", + "method": "er", + "seed": 2013, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2013/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2013/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2013.json", + "config": "configs/figure2/er/seed_2013.json", + "sha256": "b538897fbcccb36327c1b53ba1045983db4aa40be28f4dfc50977ca415f477d3", + "bytes": 77244141, + "config_sha256": "cfaa02c7a483acb568a40aceda45ed7232a69ed8d18a36b0292acaaf88070c60", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2013.npz" + }, + { + "id": "figure2/er/seed_2014/final", + "paper_item": "figure2", + "method": "er", + "seed": 2014, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2014/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2014/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2014.json", + "config": "configs/figure2/er/seed_2014.json", + "sha256": "0bd322b9855c4765f36d3fb50ef3837a7465e2ff0d115433ac8754ee4451dd3e", + "bytes": 77244141, + "config_sha256": "440f81a376217555c3d653474564f2b223414cd7dcbfd9133f4e1e34c63a68b3", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2014.npz" + }, + { + "id": "figure2/er/seed_2015/final", + "paper_item": "figure2", + "method": "er", + "seed": 2015, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2015/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2015/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2015.json", + "config": "configs/figure2/er/seed_2015.json", + "sha256": "13439a2c984264a135bafdf2442da65b615b2fa88cd3526b9561d23811baf62c", + "bytes": 77244141, + "config_sha256": "dd739594e41ef2194e894b36d1323d0c6afb72074683516cbe03d23bc561e545", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2015.npz" + }, + { + "id": "figure2/er/seed_2016/final", + "paper_item": "figure2", + "method": "er", + "seed": 2016, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2016/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2016/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2016.json", + "config": "configs/figure2/er/seed_2016.json", + "sha256": "f4921679e3c2a04e6708831294d022414e757f4d8c0a63c1616a63aa9f476a71", + "bytes": 77244141, + "config_sha256": "825be07b5d2d4edf21d187acf9409212a4af902bee8f88be16a4951d62d9d86e", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2016.npz" + }, + { + "id": "figure2/er/seed_2017/final", + "paper_item": "figure2", + "method": "er", + "seed": 2017, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2017/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2017/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2017.json", + "config": "configs/figure2/er/seed_2017.json", + "sha256": "f10df291156bd5a10ecf45774f84f2d7a874ab5e41c7c198e6715ad4881da540", + "bytes": 77244141, + "config_sha256": "39fd0964d56b28ce9dab1553d6483fb16a81e563e8a04fd76774221656f70f67", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2017.npz" + }, + { + "id": "figure2/er/seed_2018/final", + "paper_item": "figure2", + "method": "er", + "seed": 2018, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2018/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2018/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2018.json", + "config": "configs/figure2/er/seed_2018.json", + "sha256": "6e8be79a9276e1854fc65a2ee7f7946f396aa0ab17e1c62b579c7fefae7a93ae", + "bytes": 77244141, + "config_sha256": "3ce8f71538e301c030ab7ab1cc7f3f3027b80da4e2943b7b2a7f6139d66b6ac9", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2018.npz" + }, + { + "id": "figure2/er/seed_2019/final", + "paper_item": "figure2", + "method": "er", + "seed": 2019, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2019/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2019/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2019.json", + "config": "configs/figure2/er/seed_2019.json", + "sha256": "79acccdc5a9d03d4b45ef58a21ab503bc54f4ff23f8e0e527ec38a5b741c5d3a", + "bytes": 77244141, + "config_sha256": "d634c7a37e3e591973a87f58491c0dc6be990bd63170469f581adb48fd543ee5", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2019.npz" + }, + { + "id": "figure2/er/seed_2020/final", + "paper_item": "figure2", + "method": "er", + "seed": 2020, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2020/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2020/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2020.json", + "config": "configs/figure2/er/seed_2020.json", + "sha256": "c43f8bb9dc97e273b13d3e906c6a98d1f52e05e0fb6831729530f403cf4dc87d", + "bytes": 77244141, + "config_sha256": "195d8ba102530d673dafeccaad0bc6375432c568372d0c16aad77d2e364c78f7", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2020.npz" + }, + { + "id": "figure2/er/seed_2021/final", + "paper_item": "figure2", + "method": "er", + "seed": 2021, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2021/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2021/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2021.json", + "config": "configs/figure2/er/seed_2021.json", + "sha256": "774366bbbc7b71f8b6eb8965940fdef89e0812639a7ddf13eff7f18b93432ead", + "bytes": 77244141, + "config_sha256": "ec55d40af79e24ab8bd60fca97c8071560d9e084fd467c623f7143b160ab1aa4", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2021.npz" + }, + { + "id": "figure2/er/seed_2022/final", + "paper_item": "figure2", + "method": "er", + "seed": 2022, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2022/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2022/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2022.json", + "config": "configs/figure2/er/seed_2022.json", + "sha256": "bfe2ad1f212fa5e59d5a99641d42ffc3d9ecc310b59cbc13dad03bb579bbc090", + "bytes": 77244141, + "config_sha256": "341adf6fcc05bfb850355280f48d5630b5d5cba9bbcccfce05948db039c8b176", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2022.npz" + }, + { + "id": "figure2/er/seed_2023/final", + "paper_item": "figure2", + "method": "er", + "seed": 2023, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2023/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2023/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2023.json", + "config": "configs/figure2/er/seed_2023.json", + "sha256": "15ea924c475604796a38d6e594cff3125e9a6e761e30494f53fc443a5753543b", + "bytes": 77244141, + "config_sha256": "088d67b1d637d386b7a08d0a39b45c14c93f1e23bf401888dd973cb5c769f7e2", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2023.npz" + }, + { + "id": "figure2/er/seed_2024/final", + "paper_item": "figure2", + "method": "er", + "seed": 2024, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2024/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2024/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2024.json", + "config": "configs/figure2/er/seed_2024.json", + "sha256": "d4ab08a63affc0c646e03b6bdabc5f438296d8572846dba080d30c2cb2863e52", + "bytes": 77244141, + "config_sha256": "6ef296265a90a43349a6acb39134ff231cac32c7610256cb3de3222e971a4532", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2024.npz" + }, + { + "id": "figure2/er/seed_2025/final", + "paper_item": "figure2", + "method": "er", + "seed": 2025, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2025/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2025/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2025.json", + "config": "configs/figure2/er/seed_2025.json", + "sha256": "77059081f78ddd050ce14a9f59825f77cc86805e5166995bf4ef927950ad1e1f", + "bytes": 77244141, + "config_sha256": "ab56d536c47a5b96930f166909a93222570aa900f04180a5c964c4cf6c329c5c", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2025.npz" + }, + { + "id": "figure2/er/seed_2026/final", + "paper_item": "figure2", + "method": "er", + "seed": 2026, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2026/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2026/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2026.json", + "config": "configs/figure2/er/seed_2026.json", + "sha256": "6eec57a0dcaec01c95b13f8a93d276cf454cededa28829ac896a9577a1873a37", + "bytes": 77244141, + "config_sha256": "79adbeb65d1faa05c66d87c82997868f4e1216a5d5624e6d980b38b42676792b", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2026.npz" + }, + { + "id": "figure2/er/seed_2027/final", + "paper_item": "figure2", + "method": "er", + "seed": 2027, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2027/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2027/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2027.json", + "config": "configs/figure2/er/seed_2027.json", + "sha256": "799b3122bde32c74b1421b9c4943615e961ed80f6b86deaa7eb17fb133d112d4", + "bytes": 77244141, + "config_sha256": "7750f9ade4dad1e728eb8da80545005c5a401f1b5d2cd3234fdb3b7240864dd8", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2027.npz" + }, + { + "id": "figure2/er/seed_2028/final", + "paper_item": "figure2", + "method": "er", + "seed": 2028, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2028/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2028/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2028.json", + "config": "configs/figure2/er/seed_2028.json", + "sha256": "8b1bc9a82eed616f8ad67ab081354f776b05da3f85938227d4f81180e312ec57", + "bytes": 77244141, + "config_sha256": "1ef3aabe80ae3635661647c406857402a80081dae7ccf4ef36da340b3e6b67f2", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2028.npz" + }, + { + "id": "figure2/er/seed_2029/final", + "paper_item": "figure2", + "method": "er", + "seed": 2029, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2029/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2029/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2029.json", + "config": "configs/figure2/er/seed_2029.json", + "sha256": "49fbb045134ab89a60d04ec497a8e7479b6155913313587ac21dd817ece3e0f7", + "bytes": 77244141, + "config_sha256": "c0788d959b1199d05daf64fcdb171fda7482b01c003a31de1851304987879b47", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2029.npz" + }, + { + "id": "figure2/er/seed_2030/final", + "paper_item": "figure2", + "method": "er", + "seed": 2030, + "stage": "final", + "path": "checkpoints/figure2/er/seed_2030/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure2/current/er_seed2030/policy_00032768.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure2/experiments/er_seed2030.json", + "config": "configs/figure2/er/seed_2030.json", + "sha256": "4f4350717b6a7e72425cc417215e1eddf431d998a8999fd71c744885c5b0a7f3", + "bytes": 77244141, + "config_sha256": "a2f899320c265452dda1ef3f2701fb31bbdb0ef34db78a395a49eb4bd50fd52e", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 688, + "action_dim": 6, + "timesteps_in_checkpoint": 32768, + "observation_normalization_present": false, + "policy_engine": "terrain", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure2/er_inputs/er_seed2030.npz" + }, + { + "id": "figure3/connectome/seed_4301/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4301, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4301/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4301/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4301.json", + "config": "configs/figure3/connectome/seed_4301.json", + "sha256": "93eee6f179a89f3147cf43df5603d9a216649e2d978b0db4bbba7239cb729e2c", + "bytes": 79907592, + "config_sha256": "6d23a2e5b2a33ddc2b1d1aeeafedb61b6a77d9a6116567543f10cfb9421d9ed1", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4301/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4301, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4301/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4301/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4301.json", + "config": "configs/figure3/connectome/seed_4301.json", + "sha256": "1227ad4c864ffc254184b726c4f28e25afe36f56b4ccdc24a78930b14165d25d", + "bytes": 79884013, + "config_sha256": "6d23a2e5b2a33ddc2b1d1aeeafedb61b6a77d9a6116567543f10cfb9421d9ed1", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4302/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4302, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4302/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4302/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4302.json", + "config": "configs/figure3/connectome/seed_4302.json", + "sha256": "64ce096daa059ea101659317c0cf876bd28ed78c516bb48dc20dd4edc64a38cc", + "bytes": 79907592, + "config_sha256": "d070f0b238e224a211fac3cff08a1543fcb1cae3ff3f2aae61618666a3023b9c", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4302/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4302, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4302/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4302/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4302.json", + "config": "configs/figure3/connectome/seed_4302.json", + "sha256": "e3beef2577826c378a7f00e5cd38a9bdfef8119c045dffb7aff508d1b2dd60d7", + "bytes": 79884013, + "config_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4303/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4303, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4303/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4303/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4303.json", + "config": "configs/figure3/connectome/seed_4303.json", + "sha256": "56b0f77742f8933f990f15afbf94982c7d044b212edc72d8dd197c41f7663ce7", + "bytes": 79907592, + "config_sha256": "428d045676941f84ef99e3104a4496c311cbb0ea7c8f8d450519f18eec8e8578", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4303/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4303, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4303/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4303/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4303.json", + "config": "configs/figure3/connectome/seed_4303.json", + "sha256": "ffea34274fa2c6aac1b6dc8c00f17ea773cfa47de6cb2bf1f718a0008f361a53", + "bytes": 79884013, + "config_sha256": "428d045676941f84ef99e3104a4496c311cbb0ea7c8f8d450519f18eec8e8578", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4304/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4304, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4304/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4304/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4304.json", + "config": "configs/figure3/connectome/seed_4304.json", + "sha256": "a0e5ff41c66a67614c13b25b812fba1fb2e3d5223d4ccdd0bca3325494f9320b", + "bytes": 79907592, + "config_sha256": "3665fb22f1511e54247fbe106e5bc27810a9a7640ba3fcdfc358c34af4f166e2", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4304/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4304, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4304/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4304/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4304.json", + "config": "configs/figure3/connectome/seed_4304.json", + "sha256": "ba952979c39b2daf30ba098c98f441c413fb907c80e3a4678f487a8c6c41e213", + "bytes": 79884013, + "config_sha256": "3665fb22f1511e54247fbe106e5bc27810a9a7640ba3fcdfc358c34af4f166e2", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4305/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4305, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4305/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4305/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4305.json", + "config": "configs/figure3/connectome/seed_4305.json", + "sha256": "c96ea61c5117ec7e62b2376fb6b28ded161718941d07c576f2bb403e6268aaa4", + "bytes": 79907592, + "config_sha256": "aa901a35e8e58b01246def88deb6e96154b5b3df06fe26cc06368b01951a5f09", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4305/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4305, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4305/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4305/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4305.json", + "config": "configs/figure3/connectome/seed_4305.json", + "sha256": "8747d4a16141d1433fc55b98ed95696b4ae10929eb0390cdcf2a847226c05bec", + "bytes": 79884013, + "config_sha256": "aa901a35e8e58b01246def88deb6e96154b5b3df06fe26cc06368b01951a5f09", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4306/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4306, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4306/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4306/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4306.json", + "config": "configs/figure3/connectome/seed_4306.json", + "sha256": "5425ca1fd4ad068365395b07bface60fc4485cbd3ea6b662031b236b35162ae6", + "bytes": 79907592, + "config_sha256": "0334019da5c71e333ed548a9b904e07ee509be32566abb065354144a9801bd5e", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4306/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4306, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4306/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4306/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4306.json", + "config": "configs/figure3/connectome/seed_4306.json", + "sha256": "81bac3d26ee53140759c32f95b04f681211f3b5480540764f14d7643824e08be", + "bytes": 79884013, + "config_sha256": "0334019da5c71e333ed548a9b904e07ee509be32566abb065354144a9801bd5e", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4307/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4307, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4307/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4307/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4307.json", + "config": "configs/figure3/connectome/seed_4307.json", + "sha256": "649527404ea6f699028f0c296ad124d1f60a15db958d28a5e10d6d18eb101bfe", + "bytes": 79907528, + "config_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4307/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4307, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4307/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4307/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4307.json", + "config": "configs/figure3/connectome/seed_4307.json", + "sha256": "3e066c2da792ede0db54e3f7fcd8c29e821dc5ac4a5f4623de0c761415855136", + "bytes": 79884013, + "config_sha256": "4aed8484f49e3ed26f675c974c50e7ccbd0b89ae70d98c35a8b5157c413a0c3d", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4308/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4308, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4308/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4308/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4308.json", + "config": "configs/figure3/connectome/seed_4308.json", + "sha256": "01b41c8ce209904cc18b5aa71a120a0eb48733fb30d7918d91d890788e0567db", + "bytes": 79907592, + "config_sha256": "ffb7f534ba814f2ac11babbbf5c31e1ec388d4bec5ad43dd16875487c76eadcc", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4308/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4308, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4308/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4308/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4308.json", + "config": "configs/figure3/connectome/seed_4308.json", + "sha256": "bb4f32daa6af75d1a10a3700b6fe85a5c7c2adddf4c05ac3219955de7b20b32a", + "bytes": 79884013, + "config_sha256": "ffb7f534ba814f2ac11babbbf5c31e1ec388d4bec5ad43dd16875487c76eadcc", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4309/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4309, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4309/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4309/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4309.json", + "config": "configs/figure3/connectome/seed_4309.json", + "sha256": "df16481ac05fa9131f766d27f3759eb69fdc931098d03806d79fad05fe2e87a6", + "bytes": 79907528, + "config_sha256": "410d00afba13bc044bdc2af949a32917162870b1b11331f4258d23358fef6bd3", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4309/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4309, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4309/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4309/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4309.json", + "config": "configs/figure3/connectome/seed_4309.json", + "sha256": "12da7634dde0413466b35e9209bd97ee77d867d30df73037b2a2a30e4a320e59", + "bytes": 79884013, + "config_sha256": "410d00afba13bc044bdc2af949a32917162870b1b11331f4258d23358fef6bd3", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4310/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4310, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4310/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4310/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4310.json", + "config": "configs/figure3/connectome/seed_4310.json", + "sha256": "ce34d9bb3281622e80c6bdd0510dda8d2f4751d9c6e292e12f484488a06dc9c0", + "bytes": 79907656, + "config_sha256": "85a3913cf7aa4b6f0f3cb089bb718dac0b01825ba63a980c1ecf3173e37ddc4d", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4310/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4310, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4310/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4310/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4310.json", + "config": "configs/figure3/connectome/seed_4310.json", + "sha256": "7f70915e78bc789aff7ef4a763de6573c9cc463ceabe28dab13bd4b5a3add679", + "bytes": 79884013, + "config_sha256": "85a3913cf7aa4b6f0f3cb089bb718dac0b01825ba63a980c1ecf3173e37ddc4d", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4311/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4311, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4311/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4311/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4311.json", + "config": "configs/figure3/connectome/seed_4311.json", + "sha256": "d2390d6c3956bf82aa771b36aca1381cd27312cf59a2b412b3d87d7e3bd873ad", + "bytes": 79907592, + "config_sha256": "34a318941dc787af445d649588473e54bdaaad055314c7e9e960bbe429f82ad1", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4311/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4311, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4311/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4311/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4311.json", + "config": "configs/figure3/connectome/seed_4311.json", + "sha256": "c06e1b771b0d80554e051a7bd4d49a148f3317089af085ce48a09ab94d0af803", + "bytes": 79884013, + "config_sha256": "34a318941dc787af445d649588473e54bdaaad055314c7e9e960bbe429f82ad1", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4312/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4312, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4312/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4312/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4312.json", + "config": "configs/figure3/connectome/seed_4312.json", + "sha256": "fd10a23fa3845dc0e24363180a63a1a622cbdedaa640cd14c47bf2a139f1bc0a", + "bytes": 79907592, + "config_sha256": "96ddaaf9e8fbd11408562f628ef0d4daa484fb8abadf1a3bac83e3a3a156400e", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4312/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4312, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4312/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4312/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4312.json", + "config": "configs/figure3/connectome/seed_4312.json", + "sha256": "7b3b9c09f42e29af2cce1e65cb25359389ffe9dab21892abd152aa6e1ba67e46", + "bytes": 79884013, + "config_sha256": "96ddaaf9e8fbd11408562f628ef0d4daa484fb8abadf1a3bac83e3a3a156400e", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4313/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4313, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4313/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4313/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4313.json", + "config": "configs/figure3/connectome/seed_4313.json", + "sha256": "b8f45d44a65edf792371802f7167c93fb80b55c43150c7a370aa190a44ecfb3c", + "bytes": 79907656, + "config_sha256": "34e3ed35645cd6eb081a99cf0a6b88fdbea003c8a5a5d038734b8b045f60ac88", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4313/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4313, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4313/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4313/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4313.json", + "config": "configs/figure3/connectome/seed_4313.json", + "sha256": "95346274f3b60c22329916709e03ae20e50bf8c454cf5cfc7d7fe2b3bfcc189a", + "bytes": 79884013, + "config_sha256": "34e3ed35645cd6eb081a99cf0a6b88fdbea003c8a5a5d038734b8b045f60ac88", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4314/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4314, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4314/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4314/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4314.json", + "config": "configs/figure3/connectome/seed_4314.json", + "sha256": "7caad30d912db893573fbc0354b5e8976ba3d025bf2cea53e54f0bda0ba8d0c6", + "bytes": 79907592, + "config_sha256": "207aad08fe93f395a8d55a86ff5ee76882fae89dfa541dfef57d6b3f5273e1d7", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4314/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4314, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4314/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4314/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4314.json", + "config": "configs/figure3/connectome/seed_4314.json", + "sha256": "81c79eeb5340b7db649ba57df65843f5e7c3ae1d45c0aea95d6817d200d364ab", + "bytes": 79884013, + "config_sha256": "207aad08fe93f395a8d55a86ff5ee76882fae89dfa541dfef57d6b3f5273e1d7", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4315/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4315, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4315/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4315/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4315.json", + "config": "configs/figure3/connectome/seed_4315.json", + "sha256": "19dc2e9bed1e40e1185e67b3b6b97f5002b2384d908c98df33005e78b38f006b", + "bytes": 79907656, + "config_sha256": "e4fbc4b11bef39a9ad3fc89f3b6626b2099a24de6f085dbbccf3326043bfaa2d", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4315/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4315, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4315/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4315/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4315.json", + "config": "configs/figure3/connectome/seed_4315.json", + "sha256": "0717f7c8e4c0286decf0da145987e4313cf6978c440c885f475579dcaa4a9d8f", + "bytes": 79884013, + "config_sha256": "e4fbc4b11bef39a9ad3fc89f3b6626b2099a24de6f085dbbccf3326043bfaa2d", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4316/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4316, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4316/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4316/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4316.json", + "config": "configs/figure3/connectome/seed_4316.json", + "sha256": "4503bea34342bd56ff935b9632ad6b6d188428a4625cf8e7b369e3629b6f327c", + "bytes": 79907656, + "config_sha256": "2bcc86a75c5db99a62759a0550a0d5cf23c025f4966aad2f0431280dbf916afc", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4316/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4316, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4316/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4316/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4316.json", + "config": "configs/figure3/connectome/seed_4316.json", + "sha256": "ae3694301c86030ea1d228c22e04a0e5bb3b3605cbfe2ab45c3701ab994fb004", + "bytes": 79884013, + "config_sha256": "2bcc86a75c5db99a62759a0550a0d5cf23c025f4966aad2f0431280dbf916afc", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4317/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4317, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4317/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4317/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4317.json", + "config": "configs/figure3/connectome/seed_4317.json", + "sha256": "6277923d343b95dd2220dad6a090349f7c5c2993926a36dcd2b27b819a1ec115", + "bytes": 79907592, + "config_sha256": "b53bf73a38a440708a7b210c02633cb959344c4e55e1d9ef908f78947700b051", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4317/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4317, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4317/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4317/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4317.json", + "config": "configs/figure3/connectome/seed_4317.json", + "sha256": "58906a10479c440eb5d62ee870010d158f1cf7026dd56f7ca050987a88c6f240", + "bytes": 79884013, + "config_sha256": "b53bf73a38a440708a7b210c02633cb959344c4e55e1d9ef908f78947700b051", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4318/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4318, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4318/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4318/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4318.json", + "config": "configs/figure3/connectome/seed_4318.json", + "sha256": "ea656ba617ac5f3cb55b835f77df0682a5d96441dcf0ac218ea3840ab58e987d", + "bytes": 79907592, + "config_sha256": "ca462e3b79aa9dbe17287698d3459e2b1fa14b681f5edb353c2f2c4cdcc0f092", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4318/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4318, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4318/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4318/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4318.json", + "config": "configs/figure3/connectome/seed_4318.json", + "sha256": "2e89e4074d170bdc81d70c4cdf6a9c4d3e66e08106b2e6ee90217b34bb963cf8", + "bytes": 79884013, + "config_sha256": "ca462e3b79aa9dbe17287698d3459e2b1fa14b681f5edb353c2f2c4cdcc0f092", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4319/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4319, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4319/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4319/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4319.json", + "config": "configs/figure3/connectome/seed_4319.json", + "sha256": "1b86aedc26976b2f4506ac701c52020309ab9678f0d237313cfa276ef298b786", + "bytes": 79907656, + "config_sha256": "b6fd1c920f0a771036eb28ab7a8cfc30238e599f8e6101e3e0d609cc438ca0e6", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4319/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4319, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4319/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4319/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4319.json", + "config": "configs/figure3/connectome/seed_4319.json", + "sha256": "84a159f71df871f8e10070a6521f689cae2faddc74625e72a7e422dba05dca52", + "bytes": 79884013, + "config_sha256": "b6fd1c920f0a771036eb28ab7a8cfc30238e599f8e6101e3e0d609cc438ca0e6", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4320/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4320, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4320/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4320/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4320.json", + "config": "configs/figure3/connectome/seed_4320.json", + "sha256": "b5f21041e749fc49565c024f059bd03267c547b077c518cd61199718dfd62a8c", + "bytes": 79907592, + "config_sha256": "76ba43b274fba6658efa811acb58cc87c17fc848b62ae70ea6e9937ab498ea74", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4320/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4320, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4320/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4320/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4320.json", + "config": "configs/figure3/connectome/seed_4320.json", + "sha256": "34b40e50eeae02271d5009ce59a7d199bcccea74cd860ffdbd2659d5666d0209", + "bytes": 79884013, + "config_sha256": "76ba43b274fba6658efa811acb58cc87c17fc848b62ae70ea6e9937ab498ea74", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4321/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4321, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4321/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4321/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4321.json", + "config": "configs/figure3/connectome/seed_4321.json", + "sha256": "f59ee3c92d139fe2a534f5f16962fd1ab0cc410db0233317483b286eec1c233f", + "bytes": 79907656, + "config_sha256": "c618ee12c64fb6f2cff49154ac71628b8a671cd3b55083ad061422b18d6d8b68", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4321/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4321, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4321/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4321/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4321.json", + "config": "configs/figure3/connectome/seed_4321.json", + "sha256": "47e674c59d9ccff9ea2300b56654c11eec0bd9a0b4c7dd7a1c8699213077eb0e", + "bytes": 79884013, + "config_sha256": "c618ee12c64fb6f2cff49154ac71628b8a671cd3b55083ad061422b18d6d8b68", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4322/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4322, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4322/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4322/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4322.json", + "config": "configs/figure3/connectome/seed_4322.json", + "sha256": "d4f0ecf4634d2bed4ee046d2a6fdf6a7cf80e2e0b0de86a4f25b8dd4a02e94e1", + "bytes": 79907592, + "config_sha256": "f8381586f5835be82c6ad8246b4415f909fa5aea34079e5a6b9f07042333e31d", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4322/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4322, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4322/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4322/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4322.json", + "config": "configs/figure3/connectome/seed_4322.json", + "sha256": "eae2ad318f7996ef3c67836a52ee65edb5c2b2bdbd0707069354cdba42dd5a1f", + "bytes": 79884013, + "config_sha256": "f8381586f5835be82c6ad8246b4415f909fa5aea34079e5a6b9f07042333e31d", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4323/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4323, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4323/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4323/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4323.json", + "config": "configs/figure3/connectome/seed_4323.json", + "sha256": "bdc594c79db01c4f611ee8e429ad107562b1d8439b1f69a710966a82115cb714", + "bytes": 79907656, + "config_sha256": "750cfe311a587a6bd38a3794c8d4ee4cfa7ab035207e9121ae7165af6884c7d1", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4323/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4323, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4323/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4323/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4323.json", + "config": "configs/figure3/connectome/seed_4323.json", + "sha256": "d1f354519af15413c26d6a4041c75066f1976543d1dee0e4c0b1eead7e15a976", + "bytes": 79884013, + "config_sha256": "750cfe311a587a6bd38a3794c8d4ee4cfa7ab035207e9121ae7165af6884c7d1", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4324/initial", + "paper_item": "figure3", + "method": "connectome", + "seed": 4324, + "stage": "initial", + "path": "checkpoints/figure3/connectome/seed_4324/initial.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4324/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4324.json", + "config": "configs/figure3/connectome/seed_4324.json", + "sha256": "098a869f814e12d3cb3e04e66b758aeabd05c79f971696baf9ab4ef399457901", + "bytes": 79907592, + "config_sha256": "d4aa286fe909a02a199ac4d62f5b36cea083c8bd0b7d62460f4c81297a5f3517", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/connectome/seed_4324/final", + "paper_item": "figure3", + "method": "connectome", + "seed": 4324, + "stage": "final", + "path": "checkpoints/figure3/connectome/seed_4324/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/connectome_seed4324/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/connectome_seed4324.json", + "config": "configs/figure3/connectome/seed_4324.json", + "sha256": "cc68c26826723b6377c2b6b715fd492b1ab4b040c0b3c9192b68a04864c13852", + "bytes": 79884013, + "config_sha256": "d4aa286fe909a02a199ac4d62f5b36cea083c8bd0b7d62460f4c81297a5f3517", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4301/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4301, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4301/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4301/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4301.json", + "config": "configs/figure3/mlp/seed_4301.json", + "sha256": "960e645fdd4b13a1784f5847f6ef9208ddddcb41b508780a2e6cb393f6db929d", + "bytes": 5233549, + "config_sha256": "bf3b91e0248421b605b5756d9f18201b94c8e0aa1eaee6c564170e487fbead42", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4302/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4302, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4302/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4302/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4302.json", + "config": "configs/figure3/mlp/seed_4302.json", + "sha256": "52375f6a37c93c97aeb886fdf2b66e53dbd9c9e1529e7a8309d7ea9fadb1262c", + "bytes": 5233549, + "config_sha256": "4f6294524a4588d55c3c64302184adfbae1ea7251a902e1b8f9bcd9e403a4ca3", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4303/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4303, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4303/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4303/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4303.json", + "config": "configs/figure3/mlp/seed_4303.json", + "sha256": "fa64cccf9122f0bd92048d15ec2bba4296413246a76561089d5ae050d7ac150c", + "bytes": 5233549, + "config_sha256": "cf6fac0a65dc483e048357a85b8da066d5c9b7d0eef4abaae88d0ed5e674f9f5", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4304/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4304, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4304/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4304/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4304.json", + "config": "configs/figure3/mlp/seed_4304.json", + "sha256": "c208ac2428857e43f3114ab2bca1e7888df4d5ad7cf1c58c53481c8759029d61", + "bytes": 5233549, + "config_sha256": "9504ed3e5edcb9c7437d9abe31b6fdb6a282e483faf4105f4f37487e60b86519", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4305/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4305, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4305/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4305/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4305.json", + "config": "configs/figure3/mlp/seed_4305.json", + "sha256": "14f84ccff8b6b4cdaa4a099526858141ec1d913f0074c8d129bac8e82da0b8f6", + "bytes": 5233549, + "config_sha256": "526514e37bd2980d20646ac52576f8ffbe6dc78fc18be4084135e3a3495343e1", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4306/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4306, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4306/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4306/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4306.json", + "config": "configs/figure3/mlp/seed_4306.json", + "sha256": "3eeb61be636d34898eb9bff7b71699b3ec9110cd1d03f5be49dd8e5cee62ad00", + "bytes": 5233549, + "config_sha256": "328fd395cd7ecab085bd2d52f95c266e6923af4ca197f791daba8992bb776324", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4307/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4307, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4307/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4307/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4307.json", + "config": "configs/figure3/mlp/seed_4307.json", + "sha256": "2cb343611bffd2ceba494488fd73c3e9367601ee0905e3dc2772a8e2dfba9266", + "bytes": 5233549, + "config_sha256": "197683f3273e0009eeca77a9d77a0312b3c1ea9c973381cf0cf96a1b932ad1d9", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4308/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4308, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4308/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4308/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4308.json", + "config": "configs/figure3/mlp/seed_4308.json", + "sha256": "13cd6d7a9f50a2da641640235ad97d2f50b54ab237ac2e521375b65b41c483cb", + "bytes": 5233549, + "config_sha256": "684ce8d475be7d49e1b361f98a0d4bf9b0bb0a8f24a1fc77bf2f2327b825e81e", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4309/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4309, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4309/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4309/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4309.json", + "config": "configs/figure3/mlp/seed_4309.json", + "sha256": "84abdc9dc538076839d97a16d92b9fd5c56af42b9791c625d1eb034287bea53d", + "bytes": 5233549, + "config_sha256": "69774e76f669891e84d909f66dfc3c19749c6a8c0b47325d7ba3752a9827347c", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4310/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4310, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4310/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4310/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4310.json", + "config": "configs/figure3/mlp/seed_4310.json", + "sha256": "cfed25a157829ff25efa3a9a9575fae44c44df9d8e99f0d05eaa10f32ff74601", + "bytes": 5233549, + "config_sha256": "4a0a0e18f43662abea99b7c385304a4fa23a6fc02ccecab72649af5bab48c5fb", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4311/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4311, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4311/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4311/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4311.json", + "config": "configs/figure3/mlp/seed_4311.json", + "sha256": "f63581d9f97711e50152b50a5cb29b03a46dd0f5588421c1619f512512dfcafb", + "bytes": 5233549, + "config_sha256": "751f5c043779e80da3dde903e50eb0515560036231b7da0e94eefe0329e80bbb", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4312/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4312, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4312/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4312/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4312.json", + "config": "configs/figure3/mlp/seed_4312.json", + "sha256": "d5b06228d0f216c0887e2374d8eafee9fac9a6a4534afd0115b788d2a6e81d3e", + "bytes": 5233549, + "config_sha256": "36c54dc2689f0f7e244890571a9e86d565bc512301d8c1fe5e13363574afc754", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4313/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4313, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4313/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4313/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4313.json", + "config": "configs/figure3/mlp/seed_4313.json", + "sha256": "996a1fdc89acffb940931848abce708eebe970ac8cb92c5aeb182255a790f215", + "bytes": 5233549, + "config_sha256": "5b9a111c7718a3d6aa66672171787cc7a385e37b27023f81f02c0ce7a9740896", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4314/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4314, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4314/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4314/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4314.json", + "config": "configs/figure3/mlp/seed_4314.json", + "sha256": "f27e0b919310ef5f75adf7d792155a93f009c1b975d6ce671fd619a89455e63f", + "bytes": 5233549, + "config_sha256": "175aff3badf916193833c6a87841b642f5b52090756a48252f19162f89a09eca", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4315/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4315, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4315/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4315/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4315.json", + "config": "configs/figure3/mlp/seed_4315.json", + "sha256": "aecf183bb4c37a7050fabf45eeb4cca795b0bd0ade5c9b2275d4915345a255b5", + "bytes": 5233549, + "config_sha256": "b6e9bb8fef39952eeef4f595a179b8e5a26cf02f3f866b75f9ee59ba9c2c7210", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4316/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4316, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4316/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4316/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4316.json", + "config": "configs/figure3/mlp/seed_4316.json", + "sha256": "64d42950257ba0355861b8c79f5855adc5824ae0bdeffdbd826873d286edacc6", + "bytes": 5233549, + "config_sha256": "c7fe43b75bd5684fdacec45e875ac5f5d3feb04ac1fff60975a797e49afc9512", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4317/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4317, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4317/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4317/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4317.json", + "config": "configs/figure3/mlp/seed_4317.json", + "sha256": "522dcd224ee93025147f8cc34d2a7c1eb70131c545f8d18ea83aece63218df6b", + "bytes": 5233549, + "config_sha256": "f576be61c5a9a9d1916913628cedbb921568b9ff7ddb3ebc893336efd84ea48e", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4318/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4318, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4318/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4318/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4318.json", + "config": "configs/figure3/mlp/seed_4318.json", + "sha256": "a26780f84a455b09125c800a40ac23a9021a697f0e0fa49a9ce82a6dd94d33e9", + "bytes": 5233549, + "config_sha256": "bbe0e1cea1f74b4d1ed3fbccfd561731a2a5c293e4e95f65fdef8491b08cfcad", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4319/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4319, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4319/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4319/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4319.json", + "config": "configs/figure3/mlp/seed_4319.json", + "sha256": "48a6497642f3d054d6cfad22179c758f23cd3e17ac1d97d1355baec011999f90", + "bytes": 5233549, + "config_sha256": "f088add84a877f47b716398d707a45a29c5f7b250562c5a4cb24878a63d94af5", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4320/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4320, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4320/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4320/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4320.json", + "config": "configs/figure3/mlp/seed_4320.json", + "sha256": "8e16aeb0e06c6ba5c86d0de989fea574e5d505bcc1bf53c9069a47aabc09a5bc", + "bytes": 5233549, + "config_sha256": "cb35311d74f0fe5b93c75db7c8c7f3bf2a16710096dbaf7dfc0113fb07ee0db5", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4321/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4321, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4321/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4321/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4321.json", + "config": "configs/figure3/mlp/seed_4321.json", + "sha256": "fe7e10c3367d131b0a0df60aa5f8b11332270ab7bae859af1b2fecd82b6908d7", + "bytes": 5233549, + "config_sha256": "f48988687f97b7b1287d71dbfd97b29c7c4f59003a4e25c0b1da6dbec8078f3f", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4322/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4322, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4322/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4322/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4322.json", + "config": "configs/figure3/mlp/seed_4322.json", + "sha256": "6741102bd65d5660fc143ce9d2380a39c327caa4c41ccdaee55eacd5e051f6e3", + "bytes": 5233549, + "config_sha256": "c9f86ff80dbe5270f23899b55b590a8e56f4e031e7b534ef4b7a685c9b70a425", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4323/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4323, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4323/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4323/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4323.json", + "config": "configs/figure3/mlp/seed_4323.json", + "sha256": "6931e28e02702d9a13952d89a8b37f69297e4fc77ac4db5a18470165e4ec8e89", + "bytes": 5233549, + "config_sha256": "ce3b6cd14dbffa3e9c66d9270f6fc15f6ddb0a437307b635a3522275a58d2ad2", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure3/mlp/seed_4324/final", + "paper_item": "figure3", + "method": "mlp", + "seed": 4324, + "stage": "final", + "path": "checkpoints/figure3/mlp/seed_4324/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure3/mlp_seed4324/policy_00008192.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure3/experiments/mlp_seed4324.json", + "config": "configs/figure3/mlp/seed_4324.json", + "sha256": "8227a315473b52601a3e605ea0c6c8e466635839ed3ecd64d9fa4439ede8a47c", + "bytes": 5233549, + "config_sha256": "0d83422b0c9bfc1114dcc71689e26426d0afe3a30c5c3f8df64f86f6f813ff12", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 64, + "action_dim": 2, + "timesteps_in_checkpoint": 8192, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure3/graph.npz" + }, + { + "id": "figure4/connectome/seed_4401/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4401, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4401/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4401/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4401.json", + "config": "configs/figure4/connectome/seed_4401.json", + "sha256": "39d9720144a7c9b500b04ec748952fee1a7b571cb3e42c000924538660cb4c11", + "bytes": 78895341, + "config_sha256": "858a2aa88150d71dd8410f6d49479b5484e6f5660ca13a7fd95df896b13e5977", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4401/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4401, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4401/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4401/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4401.json", + "config": "configs/figure4/mlp/seed_4401.json", + "sha256": "29c05b3d50a8bed8e0ea012da9a42e2c85c951c5efadbcbb8395af439e4e8292", + "bytes": 4244429, + "config_sha256": "21220c182bda80ca6dccd74a6ba21ba0e78fa86848735afb5e62a4a6ff7adf3c", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4402/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4402, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4402/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4402/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4402.json", + "config": "configs/figure4/connectome/seed_4402.json", + "sha256": "0ff66710eb46ce0c1db0d21bb51df8e1bba72210c421f37972eb6554ea55cda1", + "bytes": 78895341, + "config_sha256": "ac641f05951b8441a0095d478d289b5c6af44157f06d662ae57ef4f78a63e567", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4402/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4402, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4402/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4402/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4402.json", + "config": "configs/figure4/mlp/seed_4402.json", + "sha256": "46998cee996ea68a2b25653f3b2a33be674a27228498cf29d099b634fbcaea8d", + "bytes": 4244429, + "config_sha256": "9da92c63b3312eb84333da69bdb2d6936a080e6f273c925f392a74050b88b00d", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4403/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4403, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4403/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4403/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4403.json", + "config": "configs/figure4/connectome/seed_4403.json", + "sha256": "a4f267cec7b896111856f24d6f624c192e1ebe3ca8cf126b2612268aaf489ea9", + "bytes": 78895341, + "config_sha256": "01462d07faebb36561025ca4e96b3ee1e2a61c824e71b6aa3d1fb5d58c4aa09a", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4403/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4403, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4403/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4403/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4403.json", + "config": "configs/figure4/mlp/seed_4403.json", + "sha256": "0367384f05da9e8b39905363cf829f8216a36e5f35a8bdfc3c59dc328316e99f", + "bytes": 4244429, + "config_sha256": "b0238fe0353e3998a8e8df813b0d3fd4a3bcb98cfffc4b3f4a63c8d013aec4b2", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4404/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4404, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4404/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4404/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4404.json", + "config": "configs/figure4/connectome/seed_4404.json", + "sha256": "e01fdbfb62575053c0b460c1a8e3054cfca96e0cf4439bbe62fc6d5875c2816b", + "bytes": 78895341, + "config_sha256": "322209f8b6369527f6cdbe029ffb2f1217930ce5a1e45969bf0f40c10014d29c", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4404/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4404, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4404/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4404/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4404.json", + "config": "configs/figure4/mlp/seed_4404.json", + "sha256": "05d98049f8d4787084cc4e75319ac8603836d6ad7352950992d3d88815bd11b0", + "bytes": 4244429, + "config_sha256": "8b800a19eacf1dafb569c476aef28f7676b511260960604746c5d915f558fa6a", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4405/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4405, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4405/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4405/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4405.json", + "config": "configs/figure4/connectome/seed_4405.json", + "sha256": "49c78af66bca9d9c78bdcd0381a53b5e059ab16dc25030fbce7c25ec3f917a91", + "bytes": 78895341, + "config_sha256": "4218ae95ce2877c290aef335c3287bff0956d3d3b30ff66d7190fc9aa87a6c21", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4405/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4405, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4405/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4405/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4405.json", + "config": "configs/figure4/mlp/seed_4405.json", + "sha256": "622e1beeb0d67b6d539e6c2fc2d7511fb0017058a612bfb2f8c313cb7f964339", + "bytes": 4244429, + "config_sha256": "452b52911de0a57a5703d8dc49e38011e8baf51cb9ffd1139e7a0527d3894d80", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4406/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4406, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4406/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4406/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4406.json", + "config": "configs/figure4/connectome/seed_4406.json", + "sha256": "2d9b83ffea5b50807c550047021d534d388120113a898091264bf2c04f40921c", + "bytes": 78895341, + "config_sha256": "ed45c7ddddac946c910510ce690ccb0f96616d5233533024176ef6982b97e2f1", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4406/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4406, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4406/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4406/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4406.json", + "config": "configs/figure4/mlp/seed_4406.json", + "sha256": "eb36657148fbdc3238dea975774fe8c4bb8013f42d2b108965080c58e6ec7451", + "bytes": 4244429, + "config_sha256": "5260e60a9d1dd02565fa4a0004e958c1101030c82a075380f6b577f913fa4d36", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4407/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4407, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4407/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4407/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4407.json", + "config": "configs/figure4/connectome/seed_4407.json", + "sha256": "fdef391d53f579a02261bb7b4a19bd5d4f13bffd4b6bca5763d3749b7ad3a8d1", + "bytes": 78895341, + "config_sha256": "b8662b51dce56ac0220723dffa57508dc6d51bbe612da692e0813111d18c43ba", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4407/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4407, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4407/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4407/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4407.json", + "config": "configs/figure4/mlp/seed_4407.json", + "sha256": "d50d968ae560dc232f53bab2869c0e3de66e04c5ed5a5ba8393a5ff77ad14235", + "bytes": 4244429, + "config_sha256": "06a463e77b8af9782d4afed29b6ea796aeb514b9d93c8782ee30513cc81e2c78", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4408/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4408, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4408/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4408/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4408.json", + "config": "configs/figure4/connectome/seed_4408.json", + "sha256": "8d5efaed4632ca3bf5656ef71251bc05676e819bed82a115bab5b6f6afba442b", + "bytes": 78895341, + "config_sha256": "b82eaddbdd058852cc784175869fb2762e8b8afd34f7bf3a24f81820b4310804", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4408/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4408, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4408/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4408/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4408.json", + "config": "configs/figure4/mlp/seed_4408.json", + "sha256": "082042bbb7b237339468f31cb4ff99afc91f8e24d935dbf2a4a9d406dbcc919a", + "bytes": 4244429, + "config_sha256": "7e5568fc175b0b3bb3b275fb5436c75bfc4fd1845bf2f1b0fd1fd78338939587", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4409/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4409, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4409/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4409/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4409.json", + "config": "configs/figure4/connectome/seed_4409.json", + "sha256": "bc28bdd4a1edf877578f0ab379d6f293e497e95af27afccfc3e810026672cb50", + "bytes": 78895341, + "config_sha256": "926f7ec6644653f73589110e4b36f42dce5b86005a698b7c7b075d0c5a2d29af", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4409/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4409, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4409/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4409/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4409.json", + "config": "configs/figure4/mlp/seed_4409.json", + "sha256": "a64950ff6366cedacc277fdf30773d692d2b69ba3e3bc6b721bd2de05e7d50ef", + "bytes": 4244429, + "config_sha256": "d3620c3e8323ebb4afaa26da44c79fbd91d1b3e77b21b17fd6aec02dcd7ce768", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4410/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4410, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4410/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4410/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4410.json", + "config": "configs/figure4/connectome/seed_4410.json", + "sha256": "109687b33d6226472fde59f63f0ba5a40ebcac5903dc7e437963f768fd023794", + "bytes": 78895341, + "config_sha256": "17a57d9bca047f1d7b271c052dde90d504db1f37142bdcd3b26a73e9f82a18ee", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4410/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4410, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4410/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4410/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4410.json", + "config": "configs/figure4/mlp/seed_4410.json", + "sha256": "58c8eeda071bf444911d4b6346d4d162cc5a14605369b02254df9bd49f906d78", + "bytes": 4244429, + "config_sha256": "c7bf90a1d8a1d6f876d5605422018076164c4e2235ec58ab247a45283841644b", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4411/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4411, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4411/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4411/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4411.json", + "config": "configs/figure4/connectome/seed_4411.json", + "sha256": "1d91d231cce3a2e102d50d871564a6e5f6e2d58bc3f2c58d01515ee87a8a82dd", + "bytes": 78895341, + "config_sha256": "432fc3d3b7535287e3056690ed85a92dfb9353bfeebb308ffb6d36c76e0e2f06", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4411/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4411, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4411/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4411/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4411.json", + "config": "configs/figure4/mlp/seed_4411.json", + "sha256": "3f4691f5c424fe08c17602c9b9624200326efcefd0161513d51791a8f69a1c40", + "bytes": 4244429, + "config_sha256": "b4b9638084a5b51783b1bf0fcc66be92d0746d1df9958af9bf169dd4cdcee0a2", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4412/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4412, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4412/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4412/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4412.json", + "config": "configs/figure4/connectome/seed_4412.json", + "sha256": "d79b5df2364eb8bc950ccee2abbeff4c4be165b37d9293edd8192c9e8d848cf0", + "bytes": 78895341, + "config_sha256": "aad84b909cb5f601edd062fa8946a009b086d6edfa60e5891f4706d15217ea9a", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4412/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4412, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4412/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4412/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4412.json", + "config": "configs/figure4/mlp/seed_4412.json", + "sha256": "b14e3a865007a409f8a4459f575027455e186e21d1506adbda5ca93f02e022a8", + "bytes": 4244429, + "config_sha256": "695a9896b6d033b1c72c4c63a85ba836d096df1d7cb33896a4d517a43326bda5", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4413/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4413, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4413/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4413/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4413.json", + "config": "configs/figure4/connectome/seed_4413.json", + "sha256": "cf8ed84aa4a4e2063c92bd454cf7e3e1387474ffa6e12dc45e6287dfc5933637", + "bytes": 78895341, + "config_sha256": "89c4bf812bbb4861bd429471482f6c8d3280b6cfb8c3fc32c6c7a30667419982", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4413/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4413, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4413/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4413/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4413.json", + "config": "configs/figure4/mlp/seed_4413.json", + "sha256": "03f730d46c1bc73fd3e2dc6022cc977f6c6ac3d567b4bf8bd73476a3a9cc5196", + "bytes": 4244429, + "config_sha256": "978805ab527a5b336f01fa76b7b68b3bcdf85def70455ab2242b1e8dcb2028fd", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4414/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4414, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4414/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4414/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4414.json", + "config": "configs/figure4/connectome/seed_4414.json", + "sha256": "36c74b78b41fbed806a4430b8cd303c402339f6a35ec9636de86b8bca76e13b8", + "bytes": 78895341, + "config_sha256": "5af8e99223233abe80c9f7e642821ff3add83d81a2edd92ad8d2fc724d5c3708", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4414/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4414, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4414/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4414/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4414.json", + "config": "configs/figure4/mlp/seed_4414.json", + "sha256": "3f91baa3e6cb8571af578bb2072f03880a207a3ee41668ee9ac305a91a810304", + "bytes": 4244429, + "config_sha256": "5e323e8293c7c1a5cee08a011774b60142689b25f1c614e64afb77b511cb5eea", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4415/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4415, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4415/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4415/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4415.json", + "config": "configs/figure4/connectome/seed_4415.json", + "sha256": "b7deb1d90d6aa81e2289c940d472da9b0088028f56931d0d311fdc85afdf28d9", + "bytes": 78895341, + "config_sha256": "802ce0491d834a1c9a79aa83b97ce542a8db981c24642bf0272dcf1d41c2fb4c", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4415/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4415, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4415/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4415/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4415.json", + "config": "configs/figure4/mlp/seed_4415.json", + "sha256": "5adfe202c532c3fa4b5641aec506fd5a6353612d20894ca3b8723cf1ad4f902f", + "bytes": 4244429, + "config_sha256": "fb75f5621ae5f746c0f14948a4b00a633a97b3326fa0528103f95b5f2e6261d1", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4416/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4416, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4416/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4416/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4416.json", + "config": "configs/figure4/connectome/seed_4416.json", + "sha256": "271b182e9fdc3c4599014d368e3dc29166ea80b17d9afac5ed02371973128eef", + "bytes": 78895341, + "config_sha256": "a5cce2c54f7e56c8fed2e69f868e17da06cef51f893eedc89d983e3152d0b265", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4416/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4416, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4416/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4416/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4416.json", + "config": "configs/figure4/mlp/seed_4416.json", + "sha256": "42ed6556da9933ca46e5b98f253c9144a7739a352002a5f8d460323d74346361", + "bytes": 4244429, + "config_sha256": "eb7ebbca9c5d25a325fb4f22b33da42dbe11655eee71e98dbac8e7b27b6cafdc", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4417/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4417, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4417/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4417/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4417.json", + "config": "configs/figure4/connectome/seed_4417.json", + "sha256": "08089e09b1879a42e30fcea5939162e73793f63bf02a3c3d74c862ad86e064e8", + "bytes": 78895341, + "config_sha256": "01585e6c01648f8437fa5613db1af5d390d0781cb20dd08fc8bbc850c8e19b24", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4417/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4417, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4417/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4417/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4417.json", + "config": "configs/figure4/mlp/seed_4417.json", + "sha256": "15265d7bcee872eebe646b5206220d0cf032d0be13dec6720066a963c52e064c", + "bytes": 4244429, + "config_sha256": "a0be6d8e2c38771d6efa3e0304c26988faf94929279a0218446862acf2f4ff1b", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4418/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4418, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4418/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4418/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4418.json", + "config": "configs/figure4/connectome/seed_4418.json", + "sha256": "9a5c24bb5dc712fe41f47c92599403a06b087e83efc512b6f32debfb7e4a1553", + "bytes": 78895341, + "config_sha256": "2e23dd71470c06a9e6b839cc9b33464e0b38534ead26534b72c0a24cbb536c81", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4418/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4418, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4418/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4418/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4418.json", + "config": "configs/figure4/mlp/seed_4418.json", + "sha256": "44022cc036b31017b1dacb965d058bc48eb3197e57b97a3e16ab2448a479e019", + "bytes": 4244429, + "config_sha256": "8a63871ab5653b025ba00c9fcdc38d7de8be2b5cbb29b754bfa7a0a355882079", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4419/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4419, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4419/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4419/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4419.json", + "config": "configs/figure4/connectome/seed_4419.json", + "sha256": "bce9259a9d1e1d4234fec47802b4999c2a272d350fa3355b6e921d4b9f31878c", + "bytes": 78895341, + "config_sha256": "b66961f2155c66c609393de6908b11bf7fc5f1e1ca5da81a8ac4f06ae1e61d5e", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4419/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4419, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4419/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4419/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4419.json", + "config": "configs/figure4/mlp/seed_4419.json", + "sha256": "4dceda7559759b5bf676bab960e30051852aaa8579ab18440227a5c8d1676c9c", + "bytes": 4244429, + "config_sha256": "6d516a35d842ffc2b9014fd933f9eda349ec081a0380ca6f9088627d43365fd6", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/connectome/seed_4420/final", + "paper_item": "figure4", + "method": "connectome", + "seed": 4420, + "stage": "final", + "path": "checkpoints/figure4/connectome/seed_4420/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/connectome_seed4420/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/connectome_seed4420.json", + "config": "configs/figure4/connectome/seed_4420.json", + "sha256": "c6e5f1ed5bcd3b539da714150b386aada18d2081d61398b1e58b3d8945c21210", + "bytes": 78895341, + "config_sha256": "d046a8b21356ec4270fd2e469f459c821915a64a2677f3c90d0fbb94eca17a82", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "figure4/mlp/seed_4420/final", + "paper_item": "figure4", + "method": "mlp", + "seed": 4420, + "stage": "final", + "path": "checkpoints/figure4/mlp/seed_4420/final.pt", + "source": "/data2/ysy/github/fly-locomotion/models/figure4/mlp_seed4420/policy.pt", + "source_config": "/data2/ysy/github/fly-locomotion/configs/figure4/experiments/mlp_seed4420.json", + "config": "configs/figure4/mlp/seed_4420.json", + "sha256": "97b62ab2fde2e5a760ca22766d68f4c070af2d9760f4f7681de1cb7e640ab4f4", + "bytes": 4244429, + "config_sha256": "134167fa572eb6c70a141c487009798aae07bafa112b88945d17270d37ed4823", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 15872, + "observation_normalization_present": false, + "policy_engine": "common_with_40ms_substeps", + "selection": "recorded_final_policy", + "graph_dependency": "data/figure4/graph.npz" + }, + { + "id": "S3/adult/seed_3/final", + "paper_item": "S3", + "method": "adult", + "seed": 3, + "stage": "final", + "path": "checkpoints/S3/adult/seed_3/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP14_baseline_completion6_20260928/training/direction_adult_seed3/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP14_baseline_completion6_20260928/training/direction_adult_seed3/config.yaml", + "config": "configs/S3/adult/seed_3.yaml", + "sha256": "ab82a9e81118c6c47367a9018db7f0b6482081d38d7d0491b991e11ff70bc09d", + "bytes": 78662049, + "config_sha256": "5561e7d41c2183d9958391e8629e6a77904aac1245fe79e4e814f055f253b4e0", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S3/larva/seed_3/final", + "paper_item": "S3", + "method": "larva", + "seed": 3, + "stage": "final", + "path": "checkpoints/S3/larva/seed_3/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP14_baseline_completion6_20260928/training/direction_larva_seed3/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP14_baseline_completion6_20260928/training/direction_larva_seed3/config.yaml", + "config": "configs/S3/larva/seed_3.yaml", + "sha256": "6a3f7d46f778c2e5a602f813ab960630ab048b7f19a4b7ed7837dd89fb8b7719", + "bytes": 69905893, + "config_sha256": "801eeaf7a66f487c319c420c4920724fbc307722ebcb6144c2dce121e8d4d7f6", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S3/adult/seed_4/final", + "paper_item": "S3", + "method": "adult", + "seed": 4, + "stage": "final", + "path": "checkpoints/S3/adult/seed_4/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP14_baseline_completion6_20260928/training/direction_adult_seed4/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP14_baseline_completion6_20260928/training/direction_adult_seed4/config.yaml", + "config": "configs/S3/adult/seed_4.yaml", + "sha256": "12985639471cfb153d3de3cdf6a072738608214e809d34faa27b755c5e604ed8", + "bytes": 78662049, + "config_sha256": "f39d77ef60286221a0951b5c5617ad6b227dcc26af2201c5a398f39622f702e2", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S3/larva/seed_4/final", + "paper_item": "S3", + "method": "larva", + "seed": 4, + "stage": "final", + "path": "checkpoints/S3/larva/seed_4/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP14_baseline_completion6_20260928/training/direction_larva_seed4/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP14_baseline_completion6_20260928/training/direction_larva_seed4/config.yaml", + "config": "configs/S3/larva/seed_4.yaml", + "sha256": "bbc338a20871f199c40212429a135ff88426245c6041be7a5004da233f2493f3", + "bytes": 69905893, + "config_sha256": "e9c53817c98d89b5481facd3f3c96f9c6aaf4d410cfbbb06ea6688da25b21255", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S3/adult/seed_5/final", + "paper_item": "S3", + "method": "adult", + "seed": 5, + "stage": "final", + "path": "checkpoints/S3/adult/seed_5/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP14_baseline_completion6_20260928/training/direction_adult_seed5/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP14_baseline_completion6_20260928/training/direction_adult_seed5/config.yaml", + "config": "configs/S3/adult/seed_5.yaml", + "sha256": "1569c8c9140994ea649dce52a6ceaa9b2f871c4bfe5b8190639d733f1ef132c2", + "bytes": 78662049, + "config_sha256": "524d295a96fd383c5c430d05747a4a145aa5d0337aac6afc3cf5f484fc8cf30a", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S3/larva/seed_5/final", + "paper_item": "S3", + "method": "larva", + "seed": 5, + "stage": "final", + "path": "checkpoints/S3/larva/seed_5/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP14_baseline_completion6_20260928/training/direction_larva_seed5/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP14_baseline_completion6_20260928/training/direction_larva_seed5/config.yaml", + "config": "configs/S3/larva/seed_5.yaml", + "sha256": "d332d0d7ce774147bdc993219045520db99576b58d2f5f9751012c79f8829d90", + "bytes": 69905893, + "config_sha256": "91ad0cb2fe2c11ebafca55c689d8b815ee7523483066937ab36dc97167dad0f3", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S3/adult/seed_0/final", + "paper_item": "S3", + "method": "adult", + "seed": 0, + "stage": "final", + "path": "checkpoints/S3/adult/seed_0/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP01_steering_target/training/steering_target_adult_original_command2_seed0/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP01_steering_target/training/steering_target_adult_original_command2_seed0/config.yaml", + "config": "configs/S3/adult/seed_0.yaml", + "sha256": "9078699c3f2eb698b018fc5310e46857bac1f5fd49b777c98b1418db32dd8915", + "bytes": 78661281, + "config_sha256": "7dbe2d0a99b052fba4227a2b47c09097a886d60bb5866b12910afa844509fbf0", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S3/adult/seed_1/final", + "paper_item": "S3", + "method": "adult", + "seed": 1, + "stage": "final", + "path": "checkpoints/S3/adult/seed_1/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP01_steering_target/training/steering_target_adult_original_command2_seed1/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP01_steering_target/training/steering_target_adult_original_command2_seed1/config.yaml", + "config": "configs/S3/adult/seed_1.yaml", + "sha256": "f1d0f6837c37773ccbe202759951faff3f0b7d613ed69dd628a9164e28c5d849", + "bytes": 78661281, + "config_sha256": "bbfe825dab7a06d635f553fe43cc588397dd4913982428ebe773320bdad11fb2", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S3/adult/seed_2/final", + "paper_item": "S3", + "method": "adult", + "seed": 2, + "stage": "final", + "path": "checkpoints/S3/adult/seed_2/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP01_steering_target/training/steering_target_adult_original_command2_seed2/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP01_steering_target/training/steering_target_adult_original_command2_seed2/config.yaml", + "config": "configs/S3/adult/seed_2.yaml", + "sha256": "e0a377ec66fec122b8b8be20d2ee956890407347372ef52e6bf5bbfe53cc0383", + "bytes": 78661281, + "config_sha256": "43f31cff624fab115213d61b2dcc24b860f77b7ca0dadd6b2b52766992690406", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S3/larva/seed_0/final", + "paper_item": "S3", + "method": "larva", + "seed": 0, + "stage": "final", + "path": "checkpoints/S3/larva/seed_0/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP01_steering_target/training/steering_target_larva_original_command2_seed0/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP01_steering_target/training/steering_target_larva_original_command2_seed0/config.yaml", + "config": "configs/S3/larva/seed_0.yaml", + "sha256": "fc5f86ab7296c05e428337a28e1e665add3acd687373ff77c60f2031d6a96108", + "bytes": 69905189, + "config_sha256": "c19eb7fbd5b5c53e5760da428dcf81f7c63e3d03a24f95634e715ba680184030", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S3/larva/seed_1/final", + "paper_item": "S3", + "method": "larva", + "seed": 1, + "stage": "final", + "path": "checkpoints/S3/larva/seed_1/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP01_steering_target/training/steering_target_larva_original_command2_seed1/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP01_steering_target/training/steering_target_larva_original_command2_seed1/config.yaml", + "config": "configs/S3/larva/seed_1.yaml", + "sha256": "90a9085bca43eaf098da2da4ae299d057da95e5f064f2ebd98eada1ec08f943e", + "bytes": 69905189, + "config_sha256": "736368bba8c1b38817c3071a81983b6c976b30bb8e7c7f22342f22422625b091", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S3/larva/seed_2/final", + "paper_item": "S3", + "method": "larva", + "seed": 2, + "stage": "final", + "path": "checkpoints/S3/larva/seed_2/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP01_steering_target/training/steering_target_larva_original_command2_seed2/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP01_steering_target/training/steering_target_larva_original_command2_seed2/config.yaml", + "config": "configs/S3/larva/seed_2.yaml", + "sha256": "d514a52e7e13bfc14d6e5a598d8b8ad495bd192f47a1403a9526b1c0ae26573b", + "bytes": 69905189, + "config_sha256": "e89fbf32664a8cd946ef610b4e91676eddf1f27d015d6341194aa0c090ff01c5", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S4/adult/seed_0/final", + "paper_item": "S4", + "method": "adult", + "seed": 0, + "stage": "final", + "path": "checkpoints/S4/adult/seed_0/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/adult_seed0/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/adult_seed0/config.yaml", + "config": "configs/S4/adult/seed_0.yaml", + "sha256": "071bef04672ad1911ea0d8fd07e6a43c0ea3e8a54e965e913668a927914711d8", + "bytes": 78663393, + "config_sha256": "73aa02ed75ff681ce08909dc55543b19879d60f6a2d74a63b339ff64a5e99317", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S4/adult/seed_1/final", + "paper_item": "S4", + "method": "adult", + "seed": 1, + "stage": "final", + "path": "checkpoints/S4/adult/seed_1/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/adult_seed1/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/adult_seed1/config.yaml", + "config": "configs/S4/adult/seed_1.yaml", + "sha256": "09b743d0f48fc91927e59a0addc804017766eb13ca0cde019d4f68ad9944b888", + "bytes": 78663393, + "config_sha256": "f1934025c87516615eb21684163086fb720b411ea8a8f86a299b19fd70428bf6", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S4/adult/seed_2/final", + "paper_item": "S4", + "method": "adult", + "seed": 2, + "stage": "final", + "path": "checkpoints/S4/adult/seed_2/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/adult_seed2/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/adult_seed2/config.yaml", + "config": "configs/S4/adult/seed_2.yaml", + "sha256": "e98ef7e66b2839ab3f03fd3734deb1d83f120901915d66cd67d397569575b2c9", + "bytes": 78663393, + "config_sha256": "168378ba3b635ff464e77af77adda10c68ac9704a6aac88cb1a9c5e1fc57b45c", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S4/adult/seed_3/final", + "paper_item": "S4", + "method": "adult", + "seed": 3, + "stage": "final", + "path": "checkpoints/S4/adult/seed_3/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/adult_seed3_attempt1/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/adult_seed3_attempt1/config.yaml", + "config": "configs/S4/adult/seed_3.yaml", + "sha256": "e10ab041a9714d30936d296bc49316b107d676f3323f247ef50516718bd93a66", + "bytes": 78663393, + "config_sha256": "0a9f59ff619256de84d4bc0b23419d1b6e58998ad3b9c475a468375a4f2ed8ea", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S4/adult/seed_4/final", + "paper_item": "S4", + "method": "adult", + "seed": 4, + "stage": "final", + "path": "checkpoints/S4/adult/seed_4/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/adult_seed4/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/adult_seed4/config.yaml", + "config": "configs/S4/adult/seed_4.yaml", + "sha256": "1e7a62574ca4f336f2fc1a72a00c0ab9d45b64f2037965144ad2ed05a224f84c", + "bytes": 78663393, + "config_sha256": "5145ebfab140cff8ba87004732732b6a0848c1c886f07ebc702389317d52eb5c", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S4/adult/seed_5/final", + "paper_item": "S4", + "method": "adult", + "seed": 5, + "stage": "final", + "path": "checkpoints/S4/adult/seed_5/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/adult_seed5/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/adult_seed5/config.yaml", + "config": "configs/S4/adult/seed_5.yaml", + "sha256": "ec700a609c0c86de53bf0f8a6be1577a083000888ab70f8ae5eb56101cb41037", + "bytes": 78663393, + "config_sha256": "4a50c3410ac706633e73d8317a325d50d96c2d053fe5e4269fdcbbed8625ae23", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S4/larva/seed_0/final", + "paper_item": "S4", + "method": "larva", + "seed": 0, + "stage": "final", + "path": "checkpoints/S4/larva/seed_0/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/training/larva_seed0/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/training/larva_seed0/config.yaml", + "config": "configs/S4/larva/seed_0.yaml", + "sha256": "a3c3fce5eb6ca27f9a811d4a166492ca98e2673ed7185602d8d315d8e76c3bba", + "bytes": 1464097, + "config_sha256": "89b2473ac5d487df2a70c8b6f17c0543bce7d6736c2dd18bfa386f780f4d9b4c", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S4/larva/seed_1/final", + "paper_item": "S4", + "method": "larva", + "seed": 1, + "stage": "final", + "path": "checkpoints/S4/larva/seed_1/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/training/larva_seed1/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/training/larva_seed1/config.yaml", + "config": "configs/S4/larva/seed_1.yaml", + "sha256": "6d53d908bec89b0b99e1a1d5d305f9e4f4ba90219d3ff199a6a93c111a8c9419", + "bytes": 1464097, + "config_sha256": "641d30e427a1f064870b13ea1aa86a2019d2731fb3e4a02e24451179cef2b37d", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S4/larva/seed_2/final", + "paper_item": "S4", + "method": "larva", + "seed": 2, + "stage": "final", + "path": "checkpoints/S4/larva/seed_2/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/training/larva_seed2/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/training/larva_seed2/config.yaml", + "config": "configs/S4/larva/seed_2.yaml", + "sha256": "8ae44a81887c09a73b5ec500a031fe1b62eb7a52925ce6c94c0ea44d4d614f44", + "bytes": 1464097, + "config_sha256": "bb8bbfb87b4ca336dd39892479290e3e0af2a24726befd482284b67301df1beb", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S4/larva/seed_3/final", + "paper_item": "S4", + "method": "larva", + "seed": 3, + "stage": "final", + "path": "checkpoints/S4/larva/seed_3/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/training/larva_seed3/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/training/larva_seed3/config.yaml", + "config": "configs/S4/larva/seed_3.yaml", + "sha256": "f10672ec0de034023b724d5a9fbc82210836dbd528c971cd9deb32012927f670", + "bytes": 1464097, + "config_sha256": "c79acb2e0677f962eb7c775ec0d2bf49283d258a5b3a566569b0e07a5d11e819", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S4/larva/seed_4/final", + "paper_item": "S4", + "method": "larva", + "seed": 4, + "stage": "final", + "path": "checkpoints/S4/larva/seed_4/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/training/larva_seed4/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/training/larva_seed4/config.yaml", + "config": "configs/S4/larva/seed_4.yaml", + "sha256": "9704803a93c1facfd12fa8f32dcf1536632f52b573146d5182529ee903448f1d", + "bytes": 1464097, + "config_sha256": "4d2852c044c0a0986bce3899d616b525b409e7139667ba2141321cb1f7c20ff1", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S4/larva/seed_5/final", + "paper_item": "S4", + "method": "larva", + "seed": 5, + "stage": "final", + "path": "checkpoints/S4/larva/seed_5/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/training/larva_seed5/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_larva6_20260928/training/larva_seed5/config.yaml", + "config": "configs/S4/larva/seed_5.yaml", + "sha256": "1073662d57189b90e91894ab3d557d4a28d28f259202605149a1e7c9ca792875", + "bytes": 1464097, + "config_sha256": "e9eb21efd425b5e5ab6d1a8e7373ae0fcd487e282dba5b1a99b76b0f5db26eff", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_descending/seed_0/initial", + "paper_item": "S5", + "method": "banc_descending", + "seed": 0, + "stage": "initial", + "path": "checkpoints/S5/banc_descending/seed_0/initial.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_descending_seed0/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_descending_seed0.yaml", + "config": "configs/S5/banc_descending/seed_0.yaml", + "sha256": "bba2a1b7e2d6a6d11ffcad7fa1a0ec32de9b77b567684a07711bc9d4b734ba12", + "bytes": 56128370, + "config_sha256": "39a1a7bb183d97251cea581f2e734111979825daa00e308c447c8ffb65d794fb", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_descending/seed_0/final", + "paper_item": "S5", + "method": "banc_descending", + "seed": 0, + "stage": "final", + "path": "checkpoints/S5/banc_descending/seed_0/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_descending_seed0/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_descending_seed0.yaml", + "config": "configs/S5/banc_descending/seed_0.yaml", + "sha256": "7d2223f1e45b81a4d43878ace9a12a21065ca97e6891be8e7a9903d7951925fb", + "bytes": 56115041, + "config_sha256": "39a1a7bb183d97251cea581f2e734111979825daa00e308c447c8ffb65d794fb", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_motor/seed_0/initial", + "paper_item": "S5", + "method": "banc_motor", + "seed": 0, + "stage": "initial", + "path": "checkpoints/S5/banc_motor/seed_0/initial.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_motor_seed0/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_motor_seed0.yaml", + "config": "configs/S5/banc_motor/seed_0.yaml", + "sha256": "8d48f98b2b4b6c78bdbad3b957efd7ae6d349f6a26f442b13efce1788424518e", + "bytes": 56124530, + "config_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_motor/seed_0/final", + "paper_item": "S5", + "method": "banc_motor", + "seed": 0, + "stage": "final", + "path": "checkpoints/S5/banc_motor/seed_0/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_motor_seed0/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_motor_seed0.yaml", + "config": "configs/S5/banc_motor/seed_0.yaml", + "sha256": "97209e6d4bf7f7b7fd6f1be28a7eb1294d7fb72f3b4f829159819accf723088e", + "bytes": 56111201, + "config_sha256": "8509647c2c499dba02096caf6df2233db3cb9fe762abd6188d4f4828da0deda8", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_descending/seed_1/initial", + "paper_item": "S5", + "method": "banc_descending", + "seed": 1, + "stage": "initial", + "path": "checkpoints/S5/banc_descending/seed_1/initial.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_descending_seed1/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_descending_seed1.yaml", + "config": "configs/S5/banc_descending/seed_1.yaml", + "sha256": "8ddc9a148c0fabbab0ffc0062beff4d91ecbb278b0a3985182d2e9b8243330c9", + "bytes": 56128434, + "config_sha256": "d31bbfd26e9b995dcf389ad435b73b8121fae807582bfac5abcd5458c86f45fb", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_descending/seed_1/final", + "paper_item": "S5", + "method": "banc_descending", + "seed": 1, + "stage": "final", + "path": "checkpoints/S5/banc_descending/seed_1/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_descending_seed1/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_descending_seed1.yaml", + "config": "configs/S5/banc_descending/seed_1.yaml", + "sha256": "9470ecf0efb16ee7b0ef0a4b8719bd09b1412a82a4ba6d972942fe75fa8a4e7d", + "bytes": 56115041, + "config_sha256": "d31bbfd26e9b995dcf389ad435b73b8121fae807582bfac5abcd5458c86f45fb", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_motor/seed_1/initial", + "paper_item": "S5", + "method": "banc_motor", + "seed": 1, + "stage": "initial", + "path": "checkpoints/S5/banc_motor/seed_1/initial.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_motor_seed1/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_motor_seed1.yaml", + "config": "configs/S5/banc_motor/seed_1.yaml", + "sha256": "efd70b823bea49259fff00b4740cff9975cb7a2e690145a4dc56c52ef169be6a", + "bytes": 56124594, + "config_sha256": "9f3aa8c3a7e0535009c4d8879f48bba9b8fe4155d3ae52c237405337a47e59f5", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_motor/seed_1/final", + "paper_item": "S5", + "method": "banc_motor", + "seed": 1, + "stage": "final", + "path": "checkpoints/S5/banc_motor/seed_1/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_motor_seed1/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_motor_seed1.yaml", + "config": "configs/S5/banc_motor/seed_1.yaml", + "sha256": "6db763052030e8049c76a903cd4b8c57d7267a352ce1efba07469c2b8a8eacff", + "bytes": 56111201, + "config_sha256": "9f3aa8c3a7e0535009c4d8879f48bba9b8fe4155d3ae52c237405337a47e59f5", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_descending/seed_2/initial", + "paper_item": "S5", + "method": "banc_descending", + "seed": 2, + "stage": "initial", + "path": "checkpoints/S5/banc_descending/seed_2/initial.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_descending_seed2/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_descending_seed2.yaml", + "config": "configs/S5/banc_descending/seed_2.yaml", + "sha256": "d4486d35150a3bf77bf29441d197b9198b09cc3a8b5c4023df79b658f5d5083e", + "bytes": 56128370, + "config_sha256": "1145b63ce05abc4771399527289f001cc3deeba8cb0a63d22929acae5d294b36", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_descending/seed_2/final", + "paper_item": "S5", + "method": "banc_descending", + "seed": 2, + "stage": "final", + "path": "checkpoints/S5/banc_descending/seed_2/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_descending_seed2/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_descending_seed2.yaml", + "config": "configs/S5/banc_descending/seed_2.yaml", + "sha256": "396b1fab8941f42eb0b97ec463e829a2936c23681704121178c5969f585c9b0a", + "bytes": 56115041, + "config_sha256": "1145b63ce05abc4771399527289f001cc3deeba8cb0a63d22929acae5d294b36", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_motor/seed_2/initial", + "paper_item": "S5", + "method": "banc_motor", + "seed": 2, + "stage": "initial", + "path": "checkpoints/S5/banc_motor/seed_2/initial.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_motor_seed2/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_motor_seed2.yaml", + "config": "configs/S5/banc_motor/seed_2.yaml", + "sha256": "59fa66f5dd543e9a85100a9eb73294be5754983ed173cd364d9cbd3cdcd9d48e", + "bytes": 56124530, + "config_sha256": "e1631e62f76cd494adbfb37692271314ea18d22e752e6fc4402833651d40206d", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_motor/seed_2/final", + "paper_item": "S5", + "method": "banc_motor", + "seed": 2, + "stage": "final", + "path": "checkpoints/S5/banc_motor/seed_2/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_motor_seed2/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_motor_seed2.yaml", + "config": "configs/S5/banc_motor/seed_2.yaml", + "sha256": "3e47116e765aee97092cb096308e8f41b3d6bf1f9aa757ecc3fb01118a107ad2", + "bytes": 56111201, + "config_sha256": "e1631e62f76cd494adbfb37692271314ea18d22e752e6fc4402833651d40206d", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_descending/seed_3/initial", + "paper_item": "S5", + "method": "banc_descending", + "seed": 3, + "stage": "initial", + "path": "checkpoints/S5/banc_descending/seed_3/initial.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_descending_seed3/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_descending_seed3.yaml", + "config": "configs/S5/banc_descending/seed_3.yaml", + "sha256": "0801ef2e50477ca0883077fc1fe7d638e2d78d66f6662ff4c0c227a5f0cc67af", + "bytes": 56128370, + "config_sha256": "0f108fc6433a7e5cc41776360f900ed5ad9914e3b31d7e83e47b17d71a250bb0", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_descending/seed_3/final", + "paper_item": "S5", + "method": "banc_descending", + "seed": 3, + "stage": "final", + "path": "checkpoints/S5/banc_descending/seed_3/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_descending_seed3/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_descending_seed3.yaml", + "config": "configs/S5/banc_descending/seed_3.yaml", + "sha256": "dc43fa51276e377a84ebdc52d77d9efd6da183d1a3543571068499c02652369d", + "bytes": 56115041, + "config_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_motor/seed_3/initial", + "paper_item": "S5", + "method": "banc_motor", + "seed": 3, + "stage": "initial", + "path": "checkpoints/S5/banc_motor/seed_3/initial.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_motor_seed3/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_motor_seed3.yaml", + "config": "configs/S5/banc_motor/seed_3.yaml", + "sha256": "39bb0a9926f8bbe0c6667257795f91e2a4b1bc71957dc48401e7916c8f6524cd", + "bytes": 56124530, + "config_sha256": "8b3ddf4b32ee4a2c18d1ab54e4683b8919b8d70e2d3a8106d98485c2e71680cf", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_motor/seed_3/final", + "paper_item": "S5", + "method": "banc_motor", + "seed": 3, + "stage": "final", + "path": "checkpoints/S5/banc_motor/seed_3/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_motor_seed3/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_motor_seed3.yaml", + "config": "configs/S5/banc_motor/seed_3.yaml", + "sha256": "e56f95100440df10f5590b881e534948f5f73dd65e922070c8c8859f2a21b062", + "bytes": 56111201, + "config_sha256": "8b3ddf4b32ee4a2c18d1ab54e4683b8919b8d70e2d3a8106d98485c2e71680cf", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_descending/seed_4/initial", + "paper_item": "S5", + "method": "banc_descending", + "seed": 4, + "stage": "initial", + "path": "checkpoints/S5/banc_descending/seed_4/initial.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_descending_seed4/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_descending_seed4.yaml", + "config": "configs/S5/banc_descending/seed_4.yaml", + "sha256": "982d18368529ec1cf92bebd808890cd9b29a1a07b2ec8aaba099e84d5375eb52", + "bytes": 56128370, + "config_sha256": "fe883bfde2c239b789bc1517ff5c42b34e26a777e528d08db9502253b0b97e64", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_descending/seed_4/final", + "paper_item": "S5", + "method": "banc_descending", + "seed": 4, + "stage": "final", + "path": "checkpoints/S5/banc_descending/seed_4/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_descending_seed4/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_descending_seed4.yaml", + "config": "configs/S5/banc_descending/seed_4.yaml", + "sha256": "2f381f4a49aa1f5e3be2859e09ac4a623224cde6425d2a7e2a440b7529838a96", + "bytes": 56115041, + "config_sha256": "fe883bfde2c239b789bc1517ff5c42b34e26a777e528d08db9502253b0b97e64", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_motor/seed_4/initial", + "paper_item": "S5", + "method": "banc_motor", + "seed": 4, + "stage": "initial", + "path": "checkpoints/S5/banc_motor/seed_4/initial.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_motor_seed4/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_motor_seed4.yaml", + "config": "configs/S5/banc_motor/seed_4.yaml", + "sha256": "75ab260b51f49ea70bfbe4e34afcc3b0c01f0064625b74ff0d6c49974772c9fc", + "bytes": 56124530, + "config_sha256": "ed99e754716385aaf92e2399132ac4709fa3e815e21c01e2e31486e98a8c9689", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_motor/seed_4/final", + "paper_item": "S5", + "method": "banc_motor", + "seed": 4, + "stage": "final", + "path": "checkpoints/S5/banc_motor/seed_4/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_motor_seed4/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_motor_seed4.yaml", + "config": "configs/S5/banc_motor/seed_4.yaml", + "sha256": "01ed00d8d367b8ec5f01a9330f91454a6c025381d71358704b0fcf383cb965c9", + "bytes": 56111201, + "config_sha256": "ed99e754716385aaf92e2399132ac4709fa3e815e21c01e2e31486e98a8c9689", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_descending/seed_5/initial", + "paper_item": "S5", + "method": "banc_descending", + "seed": 5, + "stage": "initial", + "path": "checkpoints/S5/banc_descending/seed_5/initial.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_descending_seed5/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_descending_seed5.yaml", + "config": "configs/S5/banc_descending/seed_5.yaml", + "sha256": "1a1cf4710e4fa70f0e1b0ad0bc53fb11a9dc11d550cbfe6e2a516e7d7e18acb9", + "bytes": 56128434, + "config_sha256": "e54dca184969e4ddf47e19ebccd954c836a4e8b049c2f2b9eb6930259ee7ab7e", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_descending/seed_5/final", + "paper_item": "S5", + "method": "banc_descending", + "seed": 5, + "stage": "final", + "path": "checkpoints/S5/banc_descending/seed_5/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_descending_seed5/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_descending_seed5.yaml", + "config": "configs/S5/banc_descending/seed_5.yaml", + "sha256": "52ee8cd6baed1039c3b40487296b4a76480a863debca12d331be06cb849e2320", + "bytes": 56115041, + "config_sha256": "e54dca184969e4ddf47e19ebccd954c836a4e8b049c2f2b9eb6930259ee7ab7e", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_motor/seed_5/initial", + "paper_item": "S5", + "method": "banc_motor", + "seed": 5, + "stage": "initial", + "path": "checkpoints/S5/banc_motor/seed_5/initial.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_motor_seed5/initial_checkpoint.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_motor_seed5.yaml", + "config": "configs/S5/banc_motor/seed_5.yaml", + "sha256": "12063fe668fc245a0fafac1d1f227c52a3f1e54db2a71c60ec9f6ff26b5dfa77", + "bytes": 56124594, + "config_sha256": "9200b0c1de73aad964f38c2ca3907ddac727d7753c0e584c8ac549ef04699088", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": 0, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "initial_state_for_before_after_assay", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S5/banc_motor/seed_5/final", + "paper_item": "S5", + "method": "banc_motor", + "seed": 5, + "stage": "final", + "path": "checkpoints/S5/banc_motor/seed_5/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/training/banc_motor_seed5/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP04_continuous_speed_matched6_20260921/configs/banc_motor_seed5.yaml", + "config": "configs/S5/banc_motor/seed_5.yaml", + "sha256": "720cfc120a6c8b20a9dbb79e8ec7f90bcbd1a9c7c1d8296c2be9827c28b2c0b8", + "bytes": 56111201, + "config_sha256": "9200b0c1de73aad964f38c2ca3907ddac727d7753c0e584c8ac549ef04699088", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S6/banc_descending/seed_0/final", + "paper_item": "S6", + "method": "banc_descending", + "seed": 0, + "stage": "final", + "path": "checkpoints/S6/banc_descending/seed_0/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP03_heading_hold/training/heading_hold_banc_descending_original_command2_seed0/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP03_heading_hold/training/heading_hold_banc_descending_original_command2_seed0/config.yaml", + "config": "configs/S6/banc_descending/seed_0.yaml", + "sha256": "a07afe251f2b29b197ab68dc220fc5301c35f42d15591d1e6f67da50b24cb0c1", + "bytes": 56113121, + "config_sha256": "42bb395afd3423fb754fccf750e9d9bea26ee600def2e0dfe79f47eb3fbc4941", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S6/banc_motor/seed_0/final", + "paper_item": "S6", + "method": "banc_motor", + "seed": 0, + "stage": "final", + "path": "checkpoints/S6/banc_motor/seed_0/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP03_heading_hold/training/heading_hold_banc_original_command2_seed0/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP03_heading_hold/training/heading_hold_banc_original_command2_seed0/config.yaml", + "config": "configs/S6/banc_motor/seed_0.yaml", + "sha256": "ddec87ea3fc6c6cfccdd912636a1084675eb1bad2b2be0e1220aa72deb8ecb15", + "bytes": 56109217, + "config_sha256": "cdefe00f98c126c4fec05c7b1b1e3d4472bc88597cb906a7d87e741e34ddfa50", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S6/banc_descending/seed_1/final", + "paper_item": "S6", + "method": "banc_descending", + "seed": 1, + "stage": "final", + "path": "checkpoints/S6/banc_descending/seed_1/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP03_heading_hold/training/heading_hold_banc_descending_original_command2_seed1/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP03_heading_hold/training/heading_hold_banc_descending_original_command2_seed1/config.yaml", + "config": "configs/S6/banc_descending/seed_1.yaml", + "sha256": "e239917a235c87d20c37fe28bbd5362a44951b0104ede7d9f75eb41f2b2c8064", + "bytes": 56113121, + "config_sha256": "5cfb5d2b674e98f9741c413037480b124461bf3d0b709b84f0cd042b8e435c66", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S6/banc_motor/seed_1/final", + "paper_item": "S6", + "method": "banc_motor", + "seed": 1, + "stage": "final", + "path": "checkpoints/S6/banc_motor/seed_1/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP03_heading_hold/training/heading_hold_banc_original_command2_seed1/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP03_heading_hold/training/heading_hold_banc_original_command2_seed1/config.yaml", + "config": "configs/S6/banc_motor/seed_1.yaml", + "sha256": "c2d42e56853e34841d12667af2e06bd45db714f71ce71e6d4eacbd88b8b16ad2", + "bytes": 56109217, + "config_sha256": "5869c9325975291456d89e0a705e868f17c04e39cdc3f85c9b096b59f735bb97", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S6/banc_descending/seed_2/final", + "paper_item": "S6", + "method": "banc_descending", + "seed": 2, + "stage": "final", + "path": "checkpoints/S6/banc_descending/seed_2/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP03_heading_hold/training/heading_hold_banc_descending_original_command2_seed2/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP03_heading_hold/training/heading_hold_banc_descending_original_command2_seed2/config.yaml", + "config": "configs/S6/banc_descending/seed_2.yaml", + "sha256": "d829c9520e666ef49a149664bcaae9254b4cd1cafff951a8ff24a4d782229b70", + "bytes": 56113121, + "config_sha256": "927f4cc56e9e4d84ef618e72081d88b109c2398b20a8ba3d0b7779753e90f0c6", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + }, + { + "id": "S6/banc_motor/seed_2/final", + "paper_item": "S6", + "method": "banc_motor", + "seed": 2, + "stage": "final", + "path": "checkpoints/S6/banc_motor/seed_2/final.pt", + "source": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP03_heading_hold/training/heading_hold_banc_original_command2_seed2/policy.pt", + "source_config": "/data2/ysy/github/flygym/outputs/nmi_unified/experiments/EXP03_heading_hold/training/heading_hold_banc_original_command2_seed2/config.yaml", + "config": "configs/S6/banc_motor/seed_2.yaml", + "sha256": "adb7d1a566e2d916cbd3b81f7c1d3f47d4682c4d389c96c829a1e2979ad2ba95", + "bytes": 56109217, + "config_sha256": "75e76bafbb25853db52b5eb7902195333f032baa83b89f56ee7e9996e8860bd9", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "observation_dim": 51, + "action_dim": 2, + "timesteps_in_checkpoint": null, + "observation_normalization_present": false, + "policy_engine": "supplemental", + "selection": "recorded_final_policy", + "graph_dependency": "embedded_in_checkpoint" + } + ], + "code": { + "repository": "https://github.com/THU-NeuroML/FlyControl", + "revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4" + }, + "private_staging_date": "2026-11-02" +} diff --git a/validation.json b/validation.json new file mode 100644 index 0000000000000000000000000000000000000000..8f9e2b94ca90c3324080fe277b6bd6852830d385 --- /dev/null +++ b/validation.json @@ -0,0 +1,1870 @@ +{ + "all_checkpoints": 226, + "finite_tensor_checks": true, + "representative_strict_load_forward_checks": 18, + "full_simulator_replay_run": false, + "records": [ + { + "id": "figure2/connectome/seed_2011/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 688, + "action_dim": 6 + }, + { + "id": "figure2/mlp/seed_2011/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 688, + "action_dim": 6 + }, + { + "id": "figure2/connectome/seed_2012/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2012/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2013/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2013/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2014/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2014/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2015/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2015/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2016/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2016/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2017/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2017/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2018/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2018/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2019/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2019/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2020/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2020/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2021/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2021/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2022/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2022/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2023/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2023/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2024/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2024/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2025/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2025/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2026/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2026/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2027/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2027/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2028/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2028/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2029/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2029/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/connectome/seed_2030/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/mlp/seed_2030/final", + "tensor_count": 13, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2011/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 688, + "action_dim": 6 + }, + { + "id": "figure2/er/seed_2012/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2013/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2014/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2015/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2016/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2017/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2018/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2019/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2020/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2021/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2022/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2023/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2024/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2025/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2026/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2027/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2028/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2029/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure2/er/seed_2030/final", + "tensor_count": 17, + "finite": true, + "timesteps": 32768, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4301/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 64, + "action_dim": 2 + }, + { + "id": "figure3/connectome/seed_4301/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 64, + "action_dim": 2 + }, + { + "id": "figure3/connectome/seed_4302/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4302/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4303/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4303/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4304/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4304/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4305/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4305/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4306/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4306/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4307/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4307/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4308/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4308/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4309/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4309/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4310/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4310/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4311/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4311/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4312/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4312/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4313/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4313/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4314/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4314/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4315/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4315/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4316/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4316/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4317/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4317/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4318/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4318/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4319/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4319/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4320/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4320/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4321/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4321/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4322/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4322/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4323/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4323/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4324/initial", + "tensor_count": 17, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "figure3/connectome/seed_4324/final", + "tensor_count": 17, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4301/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 64, + "action_dim": 2 + }, + { + "id": "figure3/mlp/seed_4302/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4303/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4304/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4305/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4306/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4307/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4308/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4309/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4310/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4311/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4312/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4313/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4314/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4315/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4316/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4317/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4318/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4319/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4320/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4321/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4322/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4323/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure3/mlp/seed_4324/final", + "tensor_count": 13, + "finite": true, + "timesteps": 8192, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4401/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 51, + "action_dim": 2 + }, + { + "id": "figure4/mlp/seed_4401/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 51, + "action_dim": 2 + }, + { + "id": "figure4/connectome/seed_4402/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4402/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4403/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4403/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4404/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4404/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4405/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4405/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4406/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4406/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4407/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4407/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4408/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4408/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4409/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4409/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4410/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4410/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4411/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4411/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4412/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4412/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4413/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4413/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4414/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4414/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4415/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4415/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4416/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4416/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4417/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4417/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4418/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4418/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4419/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4419/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/connectome/seed_4420/final", + "tensor_count": 17, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "figure4/mlp/seed_4420/final", + "tensor_count": 13, + "finite": true, + "timesteps": 15872, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "S3/adult/seed_3/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 51, + "action_dim": 2 + }, + { + "id": "S3/larva/seed_3/final", + "tensor_count": 14, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 51, + "action_dim": 2 + }, + { + "id": "S3/adult/seed_4/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S3/larva/seed_4/final", + "tensor_count": 14, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S3/adult/seed_5/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S3/larva/seed_5/final", + "tensor_count": 14, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S3/adult/seed_0/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "S3/adult/seed_1/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "S3/adult/seed_2/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "S3/larva/seed_0/final", + "tensor_count": 14, + "finite": true, + "timesteps": null, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "S3/larva/seed_1/final", + "tensor_count": 14, + "finite": true, + "timesteps": null, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "S3/larva/seed_2/final", + "tensor_count": 14, + "finite": true, + "timesteps": null, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "S4/adult/seed_0/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 51, + "action_dim": 2 + }, + { + "id": "S4/adult/seed_1/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S4/adult/seed_2/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S4/adult/seed_3/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S4/adult/seed_4/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S4/adult/seed_5/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S4/larva/seed_0/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 51, + "action_dim": 2 + }, + { + "id": "S4/larva/seed_1/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S4/larva/seed_2/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S4/larva/seed_3/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S4/larva/seed_4/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S4/larva/seed_5/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S5/banc_descending/seed_0/initial", + "tensor_count": 15, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 51, + "action_dim": 2 + }, + { + "id": "S5/banc_descending/seed_0/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 51, + "action_dim": 2 + }, + { + "id": "S5/banc_motor/seed_0/initial", + "tensor_count": 15, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 51, + "action_dim": 2 + }, + { + "id": "S5/banc_motor/seed_0/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 51, + "action_dim": 2 + }, + { + "id": "S5/banc_descending/seed_1/initial", + "tensor_count": 15, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "S5/banc_descending/seed_1/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S5/banc_motor/seed_1/initial", + "tensor_count": 15, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "S5/banc_motor/seed_1/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S5/banc_descending/seed_2/initial", + "tensor_count": 15, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "S5/banc_descending/seed_2/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S5/banc_motor/seed_2/initial", + "tensor_count": 15, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "S5/banc_motor/seed_2/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S5/banc_descending/seed_3/initial", + "tensor_count": 15, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "S5/banc_descending/seed_3/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S5/banc_motor/seed_3/initial", + "tensor_count": 15, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "S5/banc_motor/seed_3/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S5/banc_descending/seed_4/initial", + "tensor_count": 15, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "S5/banc_descending/seed_4/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S5/banc_motor/seed_4/initial", + "tensor_count": 15, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "S5/banc_motor/seed_4/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S5/banc_descending/seed_5/initial", + "tensor_count": 15, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "S5/banc_descending/seed_5/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S5/banc_motor/seed_5/initial", + "tensor_count": 15, + "finite": true, + "timesteps": 0, + "checkpoint_kind": "initial", + "observation_normalization_present": false + }, + { + "id": "S5/banc_motor/seed_5/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": "final_policy", + "observation_normalization_present": false + }, + { + "id": "S6/banc_descending/seed_0/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": null, + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 51, + "action_dim": 2 + }, + { + "id": "S6/banc_motor/seed_0/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": null, + "observation_normalization_present": false, + "strict_load_and_three_forward_steps": true, + "observation_dim": 51, + "action_dim": 2 + }, + { + "id": "S6/banc_descending/seed_1/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "S6/banc_motor/seed_1/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "S6/banc_descending/seed_2/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": null, + "observation_normalization_present": false + }, + { + "id": "S6/banc_motor/seed_2/final", + "tensor_count": 15, + "finite": true, + "timesteps": null, + "checkpoint_kind": null, + "observation_normalization_present": false + } + ] +} diff --git a/validation_environment.json b/validation_environment.json new file mode 100644 index 0000000000000000000000000000000000000000..4f1df0733786b5be2e266ea7f415134f188c3f46 --- /dev/null +++ b/validation_environment.json @@ -0,0 +1,7 @@ +{ + "python": "3.10.20", + "torch": "2.11.0+cu128", + "numpy": "2.0.2", + "code_revision": "1a49c33bf1e6eab17cb2953123e02ecc2549a5f4", + "device": "cpu" +}