dodox/dppo
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

v0.5 to main (#10)

Browse Source 
* v0.5 (#9)

* update idql configs

* update awr configs

* update dipo configs

* update qsm configs

* update dqm configs

* update project version to 0.5.0
This commit is contained in:
Allen Z. Ren
2024-10-07 16:35:13 -04:00
committed by GitHub
parent dd14c5887c
commit e0842e71dc
267 changed files with 6769 additions and 1645 deletions
Download Patch File Download Diff File Expand all files Collapse all files
cfg/furniture/finetune/lamp_low/ft_ppo_diffusion_mlp.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 44
action_dim: 10
transition_dim: ${action_dim}
denoising_steps: 100
ft_denoising_steps: 5
cond_steps: 1
@@ -105,7 +104,7 @@ model:
residual_style: True
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
horizon_steps: ${horizon_steps}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/lamp_low/ft_ppo_diffusion_unet.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 44
action_dim: 10
transition_dim: ${action_dim}
denoising_steps: 100
ft_denoising_steps: 5
cond_steps: 1
@@ -107,7 +106,7 @@ model:
cond_predict_scale: True
groupnorm_eps: 1e-4
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/lamp_low/ft_ppo_gaussian_mlp.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 44
action_dim: 10
transition_dim: ${action_dim}
cond_steps: 1
horizon_steps: 8
act_steps: 8
@@ -98,7 +97,7 @@ model:
std_max: 0.2
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
horizon_steps: ${horizon_steps}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/lamp_med/ft_ppo_diffusion_mlp.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 44
action_dim: 10
transition_dim: ${action_dim}
denoising_steps: 100
ft_denoising_steps: 5
cond_steps: 1
@@ -105,7 +104,7 @@ model:
residual_style: True
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
horizon_steps: ${horizon_steps}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/lamp_med/ft_ppo_diffusion_unet.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 44
action_dim: 10
transition_dim: ${action_dim}
denoising_steps: 100
ft_denoising_steps: 5
cond_steps: 1
@@ -106,7 +105,7 @@ model:
smaller_encoder: False
cond_predict_scale: True
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/lamp_med/ft_ppo_gaussian_mlp.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 44
action_dim: 10
transition_dim: ${action_dim}
cond_steps: 1
horizon_steps: 8
act_steps: 8
@@ -98,7 +97,7 @@ model:
std_max: 0.2
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
horizon_steps: ${horizon_steps}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/one_leg_low/ft_ppo_diffusion_mlp.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 58
action_dim: 10
transition_dim: ${action_dim}
denoising_steps: 100
ft_denoising_steps: 5
cond_steps: 1
@@ -105,7 +104,7 @@ model:
residual_style: True
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
horizon_steps: ${horizon_steps}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/one_leg_low/ft_ppo_diffusion_unet.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 58
action_dim: 10
transition_dim: ${action_dim}
denoising_steps: 100
ft_denoising_steps: 5
cond_steps: 1
@@ -107,7 +106,7 @@ model:
cond_predict_scale: True
groupnorm_eps: 1e-4
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/one_leg_low/ft_ppo_gaussian_mlp.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 58
action_dim: 10
transition_dim: ${action_dim}
cond_steps: 1
horizon_steps: 8
act_steps: 8
@@ -98,7 +97,7 @@ model:
std_max: 0.2
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
horizon_steps: ${horizon_steps}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/one_leg_med/ft_ppo_diffusion_mlp.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 58
action_dim: 10
transition_dim: ${action_dim}
denoising_steps: 100
ft_denoising_steps: 5
cond_steps: 1
@@ -105,7 +104,7 @@ model:
residual_style: True
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
horizon_steps: ${horizon_steps}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/one_leg_med/ft_ppo_diffusion_unet.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 58
action_dim: 10
transition_dim: ${action_dim}
denoising_steps: 100
ft_denoising_steps: 5
cond_steps: 1
@@ -107,7 +106,7 @@ model:
cond_predict_scale: True
groupnorm_eps: 1e-4
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/one_leg_med/ft_ppo_gaussian_mlp.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 58
action_dim: 10
transition_dim: ${action_dim}
cond_steps: 1
horizon_steps: 8
act_steps: 8
@@ -98,7 +97,7 @@ model:
std_max: 0.2
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
horizon_steps: ${horizon_steps}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/round_table_low/ft_ppo_diffusion_mlp.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 44
action_dim: 10
transition_dim: ${action_dim}
denoising_steps: 100
ft_denoising_steps: 5
cond_steps: 1
@@ -105,7 +104,7 @@ model:
residual_style: True
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
horizon_steps: ${horizon_steps}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/round_table_low/ft_ppo_diffusion_unet.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 44
action_dim: 10
transition_dim: ${action_dim}
denoising_steps: 100
ft_denoising_steps: 5
cond_steps: 1
@@ -107,7 +106,7 @@ model:
cond_predict_scale: True
groupnorm_eps: 1e-4
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/round_table_low/ft_ppo_gaussian_mlp.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 44
action_dim: 10
transition_dim: ${action_dim}
cond_steps: 1
horizon_steps: 8
act_steps: 8
@@ -98,7 +97,7 @@ model:
std_max: 0.2
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
horizon_steps: ${horizon_steps}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/round_table_med/ft_ppo_diffusion_mlp.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 44
action_dim: 10
transition_dim: ${action_dim}
denoising_steps: 100
ft_denoising_steps: 5
cond_steps: 1
@@ -105,7 +104,7 @@ model:
residual_style: True
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
horizon_steps: ${horizon_steps}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/round_table_med/ft_ppo_diffusion_unet.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 44
action_dim: 10
transition_dim: ${action_dim}
denoising_steps: 100
ft_denoising_steps: 5
cond_steps: 1
@@ -106,7 +105,7 @@ model:
smaller_encoder: False
cond_predict_scale: True
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
cfg/furniture/finetune/round_table_med/ft_ppo_gaussian_mlp.yaml
+1 -2
View File
@@ -15,7 +15,6 @@ device: cuda:0
env_name: ${env.specific.furniture}_${env.specific.randomness}_dim
obs_dim: 44
action_dim: 10
transition_dim: ${action_dim}
cond_steps: 1
horizon_steps: 8
act_steps: 8
@@ -98,7 +97,7 @@ model:
std_max: 0.2
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}
horizon_steps: ${horizon_steps}
transition_dim: ${transition_dim}
action_dim: ${action_dim}
critic:
_target_: model.common.critic.CriticObs
cond_dim: ${eval:'${obs_dim} * ${cond_steps}'}