Fixes skrl train/play script configurations when using the --agent argument and rename agent configuration variable (isaac-sim#3643)

# Description

This PR address the following points:
* Fix skrl train/play script configuration when using the `--agent`
argument

    Example:

    ```bash
python scripts/reinforcement_learning/skrl/train.py --task
Isaac-Cart-Double-Pendulum-Direct-v0 --headless --agent
skrl_mappo_cfg_entry_point
    ```

    Error:

    ```
[INFO]: Parsing configuration from:
isaaclab_tasks.direct.cart_double_pendulum.cart_double_pendulum_env:CartDoublePendulumEnvCfg
[INFO]: Parsing configuration from:
/home/toni/Documents/RL/toni_IsaacLab/source/isaaclab_tasks/isaaclab_tasks/direct/cart_double_pendulum/agents/skrl_mappo_cfg.yaml
[INFO] Logging experiment in directory:
/home/toni/Documents/RL/toni_IsaacLab/logs/skrl/cart_double_pendulum_direct
    Error executing job with overrides: []
    Traceback (most recent call last):
File
"/home/toni/Documents/RL/toni_IsaacLab/source/isaaclab_tasks/isaaclab_tasks/utils/hydra.py",
line 101, in hydra_main
        func(env_cfg, agent_cfg, *args, **kwargs)
File
"/home/toni/Documents/RL/toni_IsaacLab/scripts/reinforcement_learning/skrl/train.py",
line 156, in main
log_dir = datetime.now().strftime("%Y-%m-%d_%H-%M-%S") +
f"_{algorithm}_{args_cli.ml_framework}"
^^^^^^^^^
    NameError: name 'algorithm' is not defined
    ```
 
* Replace `STATES` by `OBSERVATIONS` when defining skrl's agent
configuration model inputs to ensure a smooth and error-free transition
when the new mayor version of **skrl** gets released. In such mayor
version `OBSERVATIONS` and `STATES` have different value/usage.

## Type of change

<!-- As you go through the list, delete the ones that are not
applicable. -->

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

Author: Toni-SM

scripts/reinforcement_learning/skrl/play.py

@@ -121,6 +121,7 @@
     agent_cfg_entry_point = "skrl_cfg_entry_point" if algorithm in ["ppo"] else f"skrl_{algorithm}_cfg_entry_point"
 else:
     agent_cfg_entry_point = args_cli.agent
+    algorithm = agent_cfg_entry_point.split("_cfg")[0].split("skrl_")[-1].lower()
 
 
 @hydra_task_config(args_cli.task, agent_cfg_entry_point)

scripts/reinforcement_learning/skrl/train.py

@@ -119,6 +119,7 @@
     agent_cfg_entry_point = "skrl_cfg_entry_point" if algorithm in ["ppo"] else f"skrl_{algorithm}_cfg_entry_point"
 else:
     agent_cfg_entry_point = args_cli.agent
+    algorithm = agent_cfg_entry_point.split("_cfg")[0].split("skrl_")[-1].lower()
 
 
 @hydra_task_config(args_cli.task, agent_cfg_entry_point)

source/isaaclab_tasks/isaaclab_tasks/direct/allegro_hand/agents/skrl_ppo_cfg.yaml

@@ -19,7 +19,7 @@ models:
     initial_log_std: 0.0
     network:
       - name: net
-        input: STATES
+        input: OBSERVATIONS
         layers: [1024, 512, 256, 128]
         activations: elu
     output: ACTIONS
@@ -28,7 +28,7 @@ models:
     clip_actions: False
     network:
       - name: net
-        input: STATES
+        input: OBSERVATIONS
         layers: [1024, 512, 256, 128]
         activations: elu
     output: ONE

source/isaaclab_tasks/isaaclab_tasks/direct/ant/agents/skrl_ppo_cfg.yaml

@@ -19,7 +19,7 @@ models:
     initial_log_std: 0.0
     network:
       - name: net
-        input: STATES
+        input: OBSERVATIONS
         layers: [256, 128, 64]
         activations: elu
     output: ACTIONS
@@ -28,7 +28,7 @@ models:
     clip_actions: False
     network:
       - name: net
-        input: STATES
+        input: OBSERVATIONS
         layers: [256, 128, 64]
         activations: elu
     output: ONE

source/isaaclab_tasks/isaaclab_tasks/direct/anymal_c/agents/skrl_flat_ppo_cfg.yaml

@@ -19,7 +19,7 @@ models:
     initial_log_std: 0.0
     network:
       - name: net
-        input: STATES
+        input: OBSERVATIONS
         layers: [128, 128, 128]
         activations: elu
     output: ACTIONS
@@ -28,7 +28,7 @@ models:
     clip_actions: False
     network:
       - name: net
-        input: STATES
+        input: OBSERVATIONS
         layers: [128, 128, 128]
         activations: elu
     output: ONE

source/isaaclab_tasks/isaaclab_tasks/direct/anymal_c/agents/skrl_rough_ppo_cfg.yaml

@@ -19,7 +19,7 @@ models:
     initial_log_std: 0.0
     network:
       - name: net
-        input: STATES
+        input: OBSERVATIONS
         layers: [512, 256, 128]
         activations: elu
     output: ACTIONS
@@ -28,7 +28,7 @@ models:
     clip_actions: False
     network:
       - name: net
-        input: STATES
+        input: OBSERVATIONS
         layers: [512, 256, 128]
         activations: elu
     output: ONE

source/isaaclab_tasks/isaaclab_tasks/direct/cart_double_pendulum/agents/skrl_ippo_cfg.yaml

@@ -19,7 +19,7 @@ models:
     initial_log_std: 0.0
     network:
       - name: net
-        input: STATES
+        input: OBSERVATIONS
         layers: [32, 32]
         activations: elu
     output: ACTIONS
@@ -28,7 +28,7 @@ models:
     clip_actions: False
     network:
       - name: net
-        input: STATES
+        input: OBSERVATIONS
         layers: [32, 32]
         activations: elu
     output: ONE

source/isaaclab_tasks/isaaclab_tasks/direct/cart_double_pendulum/agents/skrl_mappo_cfg.yaml

@@ -19,7 +19,7 @@ models:
     initial_log_std: 0.0
     network:
       - name: net
-        input: STATES
+        input: OBSERVATIONS
         layers: [32, 32]
         activations: elu
     output: ACTIONS
@@ -28,7 +28,7 @@ models:
     clip_actions: False
     network:
       - name: net
-        input: STATES
+        input: OBSERVATIONS
         layers: [32, 32]
         activations: elu
     output: ONE

source/isaaclab_tasks/isaaclab_tasks/direct/cart_double_pendulum/agents/skrl_ppo_cfg.yaml

@@ -19,7 +19,7 @@ models:
     initial_log_std: 0.0
     network:
       - name: net
-        input: STATES
+        input: OBSERVATIONS
         layers: [32, 32]
         activations: elu
     output: ACTIONS
@@ -28,7 +28,7 @@ models:
     clip_actions: False
     network:
       - name: net
-        input: STATES
+        input: OBSERVATIONS
         layers: [32, 32]
         activations: elu
     output: ONE

source/isaaclab_tasks/isaaclab_tasks/direct/cartpole/agents/skrl_camera_ppo_cfg.yaml

@@ -19,7 +19,7 @@ models:
     initial_log_std: 0.0
     network:
       - name: features_extractor
-        input: permute(STATES, (0, 3, 1, 2))  # PyTorch NHWC -> NCHW. Warning: don't permute for JAX since it expects NHWC
+        input: permute(OBSERVATIONS, (0, 3, 1, 2))  # PyTorch NHWC -> NCHW. Warning: don't permute for JAX since it expects NHWC
         layers:
           - conv2d: {out_channels: 32, kernel_size: 8, stride: 4, padding: 0}
           - conv2d: {out_channels: 64, kernel_size: 4, stride: 2, padding: 0}
@@ -36,7 +36,7 @@ models:
     clip_actions: False
     network:
       - name: features_extractor
-        input: permute(STATES, (0, 3, 1, 2))  # PyTorch NHWC -> NCHW. Warning: don't permute for JAX since it expects NHWC
+        input: permute(OBSERVATIONS, (0, 3, 1, 2))  # PyTorch NHWC -> NCHW. Warning: don't permute for JAX since it expects NHWC
         layers:
           - conv2d: {out_channels: 32, kernel_size: 8, stride: 4, padding: 0}
           - conv2d: {out_channels: 64, kernel_size: 4, stride: 2, padding: 0}