Unverified Commit 1e3b6ec2 authored by Lingheng Meng's avatar Lingheng Meng Committed by GitHub

Fixes `net_arch` in `sb3_ppo_cfg.yaml` for Isaac-Lift-Cube-Franka-v0 task (#1249)

# Description
Two changes are made to fix _[Bug Report] sb3_ppo_cfg.yaml not work for
Isaac-Lift-Cube-Franka-v0 #1248_:

1. Add `sb3_cfg_entry_point` to
`source/extensions/omni.isaac.lab_tasks/omni/isaac/lab_tasks/manager_based/manipulation/lift/config/franka/init.py`
2. Change `net_arch=[32, 32, dict(pi=[256, 128, 64], vf=[256, 128,
64])]` to `net_arch=dict(pi=[256, 128, 64], vf=[256, 128, 64])` in line
19 of `sb3_ppo_cfg.yaml`.


Fixes #1248

## Type of change

- Bug fix (non-breaking change which fixes an issue)

- [x] I have run the [`pre-commit` checks](https://pre-commit.com/) with
`./isaaclab.sh --format`
- [x] My changes generate no new warnings
parent c8f77a78
...@@ -23,6 +23,7 @@ gym.register( ...@@ -23,6 +23,7 @@ gym.register(
"rsl_rl_cfg_entry_point": f"{agents.__name__}.rsl_rl_ppo_cfg:LiftCubePPORunnerCfg", "rsl_rl_cfg_entry_point": f"{agents.__name__}.rsl_rl_ppo_cfg:LiftCubePPORunnerCfg",
"skrl_cfg_entry_point": f"{agents.__name__}:skrl_ppo_cfg.yaml", "skrl_cfg_entry_point": f"{agents.__name__}:skrl_ppo_cfg.yaml",
"rl_games_cfg_entry_point": f"{agents.__name__}:rl_games_ppo_cfg.yaml", "rl_games_cfg_entry_point": f"{agents.__name__}:rl_games_ppo_cfg.yaml",
"sb3_cfg_entry_point": f"{agents.__name__}:sb3_ppo_cfg.yaml",
}, },
disable_env_checker=True, disable_env_checker=True,
) )
...@@ -35,6 +36,7 @@ gym.register( ...@@ -35,6 +36,7 @@ gym.register(
"rsl_rl_cfg_entry_point": f"{agents.__name__}.rsl_rl_ppo_cfg:LiftCubePPORunnerCfg", "rsl_rl_cfg_entry_point": f"{agents.__name__}.rsl_rl_ppo_cfg:LiftCubePPORunnerCfg",
"skrl_cfg_entry_point": f"{agents.__name__}:skrl_ppo_cfg.yaml", "skrl_cfg_entry_point": f"{agents.__name__}:skrl_ppo_cfg.yaml",
"rl_games_cfg_entry_point": f"{agents.__name__}:rl_games_ppo_cfg.yaml", "rl_games_cfg_entry_point": f"{agents.__name__}:rl_games_ppo_cfg.yaml",
"sb3_cfg_entry_point": f"{agents.__name__}:sb3_ppo_cfg.yaml",
}, },
disable_env_checker=True, disable_env_checker=True,
) )
......
...@@ -16,7 +16,7 @@ learning_rate: !!float 3e-4 ...@@ -16,7 +16,7 @@ learning_rate: !!float 3e-4
clip_range: 0.2 clip_range: 0.2
policy_kwargs: "dict( policy_kwargs: "dict(
activation_fn=nn.ELU, activation_fn=nn.ELU,
net_arch=[32, 32, dict(pi=[256, 128, 64], vf=[256, 128, 64])] net_arch=dict(pi=[256, 128, 64], vf=[256, 128, 64])
)" )"
target_kl: 0.01 target_kl: 0.01
max_grad_norm: 1.0 max_grad_norm: 1.0
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment