[Bug Report] Isaac-Stack-Cube-Franka-Blueprint-v0 env does not exist

[yanan1116]

hi,

there is env called Isaac-Stack-Cube-Franka-Blueprint-v0 on your website https://isaac-sim.github.io/IsaacLab/main/source/overview/environments.html#manipulation

however, when i try to list all envs via
./isaaclab.sh -p scripts/environments/list_envs.py

it does not exist in the list...

what exist in the env list is

|  132   | Isaac-Stack-Cube-Franka-IK-Rel-Blueprint-v0                   | isaaclab.envs:ManagerBasedRLEnv                                                                       | <class 'isaaclab_tasks.manager_based.manipulation.stack.config.franka.stack_ik_rel_blueprint_env_cfg.FrankaCubeStackBlueprintEnvCfg'>

which I think it is from script source/isaaclab_tasks/isaaclab_tasks/manager_based/manipulation/stack/config/franka/__init__.py

and as you can see in the script, th register id is Isaac-Stack-Cube-Franka-IK-Rel-Blueprint-v0, which may not be correct.

Could you please take a look at this issue ?

thanks.

[louislelay]

Hey @yanan1116, this looks like a docs mismatch.
The Teleop & Imitation page references Isaac-Stack-Cube-Franka-IK-Rel-Blueprint-v0, and as you noted, that’s also what’s registered in the code, so the environments list should mention this ID.

[yanan1116]

@louislelay
hi, Isaac-Stack-Cube-Franka-IK-Rel-Blueprint-v0 does not exist on this page: https://isaac-sim.github.io/IsaacLab/main/source/overview/teleop_imitation.html

what I want to do is like this

./isaaclab.sh -p scripts/reinforcement_learning/sb3/play.py --task Isaac-Stack-Cube-Franka-IK-Rel-Blueprint-v0  \
--num_envs 1  --enable_cameras --use_last_checkpoint --headless

I find that tasks of stack operation by franka, currently, are not available for RL training and inference.
some configurations are needed to make them work.

[louislelay]

Hi @yanan1116, you’re right, I misread the ID, but my previous point still applies.

The Blueprint environment (Isaac-Stack-Cube-Franka-IK-Rel-Blueprint-v0) isn’t listed in the RL-ready environments. It’s mainly provided as a base configuration / template that you can extend to build custom tasks, and it’s also referenced in some imitation learning examples. That’s why you won’t be able to run it directly with SB3.

If you’d like to train with reinforcement learning, I suggest checking the comprehensive list of environments at the end of the page to see which tasks are fully supported.