Skip to content

Update the template generator for RSL-RL #6806

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
kellyguo11 merged 3 commits into from
Aug 3, 2026
Merged

Update the template generator for RSL-RL #6806

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
0ffd84f
fix template for new rsl_rl
ClemensSchwarke Jul 30, 2026
69cacbe
Update RSL-RL template generator tests
kellyguo11 Aug 3, 2026
c98e6f2
Merge branch 'develop' into fix/rsl_rl_template
kellyguo11 Aug 3, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions source/isaaclab_rl/changelog.d/fix-rsl-rl-template-tests.skip
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
Internal: updated template generator tests for RSL-RL Distillation support.
10 changes: 7 additions & 3 deletions source/isaaclab_rl/test/test_template_generator.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,7 +25,7 @@

_SINGLE_AGENT_RL_LIBRARIES = [
{"name": "rl_games", "algorithms": ["ppo"]},
{"name": "rsl_rl", "algorithms": ["ppo"]},
{"name": "rsl_rl", "algorithms": ["distillation", "ppo"]},
{"name": "skrl", "algorithms": ["amp", "ppo"]},
{"name": "sb3", "algorithms": ["ppo"]},
]
Expand Down Expand Up @@ -106,7 +106,7 @@ def _unregister(task_id: str) -> None:
False,
{
"rl_games": ["PPO"],
"rsl_rl": ["PPO"],
"rsl_rl": ["DISTILLATION", "PPO"],
"skrl": ["AMP", "PPO"],
"sb3": ["PPO"],
},
Expand All @@ -126,7 +126,7 @@ def _unregister(task_id: str) -> None:
True,
{
"rl_games": ["PPO"],
"rsl_rl": ["PPO"],
"rsl_rl": ["DISTILLATION", "PPO"],
"skrl": ["AMP", "IPPO", "MAPPO", "PPO"],
"sb3": ["PPO"],
},
Expand Down Expand Up @@ -181,6 +181,10 @@ def test_generator_registers_single_agent_rl_config_entry_points_for_all_librari

assert spec.kwargs["env_cfg_entry_point"] == f"{module_name}.{task_folder}_env_cfg:{task_class}EnvCfg"
assert spec.kwargs["rl_games_cfg_entry_point"] == f"{agents_module}:rl_games_ppo_cfg.yaml"
assert (
spec.kwargs["rsl_rl_distillation_cfg_entry_point"]
== f"{agents_module}.rsl_rl_distillation_cfg:DistillationRunnerCfg"
)
assert spec.kwargs["rsl_rl_cfg_entry_point"] == f"{agents_module}.rsl_rl_ppo_cfg:PPORunnerCfg"
assert spec.kwargs["skrl_amp_cfg_entry_point"] == f"{agents_module}:skrl_amp_cfg.yaml"
assert spec.kwargs["skrl_cfg_entry_point"] == f"{agents_module}:skrl_ppo_cfg.yaml"
Expand Down
19 changes: 10 additions & 9 deletions tools/template/cli.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,7 @@
import importlib
import os
from collections.abc import Callable
from textwrap import fill

import rich.console
import rich.table
Expand Down Expand Up @@ -213,21 +214,21 @@ def main() -> None:
# - show supported RL libraries and features
rl_library_table = rich.table.Table(title="Supported RL libraries")
rl_library_table.add_column("RL/training feature", no_wrap=True)
rl_library_table.add_column("rl_games")
rl_library_table.add_column("rsl_rl")
rl_library_table.add_column("skrl")
rl_library_table.add_column("sb3")
rl_library_table.add_column("rl_games", overflow="fold")
rl_library_table.add_column("rsl_rl", overflow="fold")
rl_library_table.add_column("skrl", overflow="fold")
rl_library_table.add_column("sb3", overflow="fold")
rl_library_table.add_row("ML frameworks", "PyTorch", "PyTorch", "PyTorch, JAX", "PyTorch")
rl_library_table.add_row("Relative performance", "~1X", "~1X", "~1X", "~0.03X")
rl_library_table.add_row(
"Algorithms",
", ".join(algorithms_per_rl_library.get("rl_games", [])),
", ".join(algorithms_per_rl_library.get("rsl_rl", [])),
", ".join(algorithms_per_rl_library.get("skrl", [])),
", ".join(algorithms_per_rl_library.get("sb3", [])),
fill(", ".join(algorithms_per_rl_library.get("rl_games", [])), width=12, break_long_words=False),
fill(", ".join(algorithms_per_rl_library.get("rsl_rl", [])), width=12, break_long_words=False),
fill(", ".join(algorithms_per_rl_library.get("skrl", [])), width=12, break_long_words=False),
fill(", ".join(algorithms_per_rl_library.get("sb3", [])), width=12, break_long_words=False),
)
rl_library_table.add_row("Multi-agent support", State.No, State.No, State.Yes, State.No)
rl_library_table.add_row("Distributed training", State.Yes, State.No, State.Yes, State.No)
rl_library_table.add_row("Distributed training", State.Yes, State.Yes, State.Yes, State.No)
rl_library_table.add_row("Vectorized training", State.Yes, State.Yes, State.Yes, State.No)
rl_library_table.add_row("Fundamental/composite spaces", State.No, State.No, State.Yes, State.No)
cli_handler.output_table(rl_library_table)
Expand Down
2 changes: 1 addition & 1 deletion tools/template/common.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,5 +11,5 @@
TEMPLATE_DIR = os.path.join(ROOT_DIR, "tools", "template", "templates")

# RL algorithms
SINGLE_AGENT_ALGORITHMS = ["AMP", "PPO"]
SINGLE_AGENT_ALGORITHMS = ["AMP", "PPO", "DISTILLATION"]

@greptile-apps greptile-apps Bot Jul 30, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Algorithm expectations omit Distillation

Adding DISTILLATION changes the discovered RSL-RL single-agent algorithms to ["DISTILLATION", "PPO"], but the checked-in discovery tests still expect only ["PPO"], causing test_get_algorithms_per_rl_library_filters_by_workflow_type to fail.

MULTI_AGENT_ALGORITHMS = ["IPPO", "MAPPO"]
34 changes: 34 additions & 0 deletions tools/template/templates/agents/rsl_rl_distillation_cfg
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,34 @@
# Copyright (c) 2022-2026, The Isaac Lab Project Developers (https://github.com/isaac-sim/IsaacLab/blob/main/CONTRIBUTORS.md).
# All rights reserved.
#
# SPDX-License-Identifier: BSD-3-Clause

from isaaclab.utils.configclass import configclass

from isaaclab_rl.rsl_rl import RslRlDistillationAlgorithmCfg, RslRlDistillationRunnerCfg, RslRlMLPModelCfg


@configclass
class DistillationRunnerCfg(RslRlDistillationRunnerCfg):
num_steps_per_env = 60
max_iterations = 150
save_interval = 50
experiment_name = "cartpole_direct"
obs_groups = {"student": ["policy"], "teacher": ["policy"]}
student = RslRlMLPModelCfg(
hidden_dims=[32, 32],
activation="elu",
obs_normalization=False,
distribution_cfg=RslRlMLPModelCfg.GaussianDistributionCfg(init_std=1.0),
)
teacher = RslRlMLPModelCfg(
hidden_dims=[32, 32],
activation="elu",
obs_normalization=False,
distribution_cfg=RslRlMLPModelCfg.GaussianDistributionCfg(init_std=0.0),
)
algorithm = RslRlDistillationAlgorithmCfg(
num_learning_epochs=2,
learning_rate=1.0e-3,
gradient_length=15,
)
1 change: 1 addition & 0 deletions tools/template/templates/agents/rsl_rl_ppo_cfg
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@ class PPORunnerCfg(RslRlOnPolicyRunnerCfg):
max_iterations = 150
save_interval = 50
experiment_name = "cartpole_direct"
obs_groups = {"actor": ["policy"], "critic": ["policy"]}
actor = RslRlMLPModelCfg(
hidden_dims=[32, 32],
activation="elu",
Expand Down
3 changes: 2 additions & 1 deletion tools/template/templates/tasks/__init__task
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,7 +27,8 @@ gym.register(
{% for algorithm in rl_library.algorithms %}
{# configuration file #}
{% if rl_library.name == "rsl_rl" %}
{% set agent_config = "." ~ rl_library.name ~ "_" ~ algorithm ~ "_cfg:" ~ algorithm|upper ~ "RunnerCfg" %}
{% set runner_prefix = {"distillation": "Distillation"}.get(algorithm, algorithm|upper) %}
{% set agent_config = "." ~ rl_library.name ~ "_" ~ algorithm ~ "_cfg:" ~ runner_prefix ~ "RunnerCfg" %}
Comment on lines +30 to +31

@greptile-apps greptile-apps Bot Jul 30, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Distillation default entry point missing

When a project is generated with RSL-RL Distillation but without PPO, registration exposes only rsl_rl_distillation_cfg_entry_point, while the standard train and play commands request rsl_rl_cfg_entry_point, causing configuration loading to fail with Could not find configuration for the environment.

{% else %}
{% set agent_config = ":" ~ rl_library.name ~ "_" ~ algorithm ~ "_cfg.yaml" %}
{% endif %}
Expand Down
Loading