1d04211666
* update from scratch configs * update gym pretraining configs - use fewer epochs * update robomimic pretraining configs - use fewer epochs * allow trajectory plotting in eval agent * add simple vit unet * update avoid pretraining configs - use fewer epochs * update furniture pretraining configs - use same amount of epochs as before * add robomimic diffusion unet pretraining configs * update robomimic finetuning configs - higher lr * add vit unet checkpoint urls * update pretraining and finetuning instructions as configs are updated
59 lines
1.6 KiB
YAML
59 lines
1.6 KiB
YAML
defaults:
|
|
- _self_
|
|
hydra:
|
|
run:
|
|
dir: ${logdir}
|
|
_target_: agent.eval.eval_gaussian_agent.EvalGaussianAgent
|
|
|
|
name: ${env_name}_eval_gaussian_mlp_ta${horizon_steps}
|
|
logdir: ${oc.env:DPPO_LOG_DIR}/robomimic-eval/${name}/${now:%Y-%m-%d}_${now:%H-%M-%S}_${seed}
|
|
base_policy_path:
|
|
robomimic_env_cfg_path: cfg/robomimic/env_meta/${env_name}.json
|
|
normalization_path: ${oc.env:DPPO_DATA_DIR}/robomimic/${env_name}/normalization.npz
|
|
|
|
seed: 42
|
|
device: cuda:0
|
|
env_name: can
|
|
obs_dim: 23
|
|
action_dim: 7
|
|
cond_steps: 1
|
|
horizon_steps: 4
|
|
act_steps: 4
|
|
|
|
n_steps: 300 # each episode takes max_episode_steps / act_steps steps
|
|
render_num: 0
|
|
|
|
env:
|
|
n_envs: 50
|
|
name: ${env_name}
|
|
best_reward_threshold_for_success: 1
|
|
max_episode_steps: 300
|
|
save_video: False
|
|
wrappers:
|
|
robomimic_lowdim:
|
|
normalization_path: ${normalization_path}
|
|
low_dim_keys: ['robot0_eef_pos',
|
|
'robot0_eef_quat',
|
|
'robot0_gripper_qpos',
|
|
'object'] # same order of preprocessed observations
|
|
multi_step:
|
|
n_obs_steps: ${cond_steps}
|
|
n_action_steps: ${act_steps}
|
|
max_episode_steps: ${env.max_episode_steps}
|
|
reset_within_step: True
|
|
|
|
model:
|
|
_target_: model.common.gaussian.GaussianModel
|
|
randn_clip_value: 3
|
|
#
|
|
network_path: ${base_policy_path}
|
|
network:
|
|
_target_: model.common.mlp_gaussian.Gaussian_MLP
|
|
mlp_dims: [512, 512, 512]
|
|
residual_style: True
|
|
fixed_std: 0.1
|
|
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
|
|
horizon_steps: ${horizon_steps}
|
|
action_dim: ${action_dim}
|
|
horizon_steps: ${horizon_steps}
|
|
device: ${device} |