SimplerEnv

SimplerEnv is evaluated through a dedicated runtime. The caller's policy environment does not need to import the simulator stack.

Install

pip install "praxis-eval[simpler]"

This extra installs praxis-simpler and praxis-remote in the caller environment. The simulator itself usually runs in a dedicated conda or micromamba environment created by setup.

Setup

praxis-eval-setup simpler
praxis-eval-setup simpler --env-manager micromamba --env-name simpler-praxis
praxis-eval-setup simpler --ms-asset-dir /data/simpler/maniskill_assets
praxis-eval-setup simpler --skip-assets
praxis-eval-setup simpler --help

Setup reads the env spec from the installed praxis-simpler package resource simpler_env/praxis_conda_env.yaml, installs praxis-eval, praxis-remote, and praxis-simpler into that runtime, downloads ManiSkill assets, and writes Vulkan/EGL activation hooks.

Verify

Run this only on a machine with the dedicated SimplerEnv runtime:

praxis-eval-verify simpler --env-name simpler-praxis --num-episodes 1 --num-envs 1
praxis-eval-verify simpler --env-python-bin /path/to/simpler-praxis/bin/python
praxis-eval-verify simpler --help

Task Selection

Default evaluator task: bridge.

Selector	Environment id	Task id
widowx_carrot_on_plate	PutCarrotOnPlateInScene-v1	0
widowx_spoon_on_towel	PutSpoonOnTableClothInScene-v1	1
widowx_stack_cube	StackGreenCubeOnYellowCubeBakedTexInScene-v1	2
widowx_put_eggplant_in_basket	PutEggplantInBasketScene-v1	3
bridge / bridge_mt4	all four tasks	all

EvalConfig.task_ids filters the resolved task list by index.

Observation Format

Policy-facing keys emitted by the SimplerEnv remote wrapper:

Key	Shape / dtype	Notes
task	str	Bridge task instruction.
observation.images.image	(3, H, W), uint8 or float32	Primary image key, configurable with primary_image_key.
observation.state	implementation-specific, often (7,), float32	State key, configurable with state_key.

The wrapper preserves uint8 image transport when possible and converts HWC images to CHW.

Action Format

Field	Value
Shape	(7,)
Dtype	float32
Range	Not bounded by the public ActionSpec
Convention	simpler_bridge_widowx_action

The external runtime formats policy actions into the SimplerEnv Bridge action dictionary. The first three values are the world vector, the next three are rotation delta, and the last value controls the gripper. env_kwargs["action_scale"] scales the action before it reaches the runtime adapter.

Runtime Config

Pass the dedicated runtime interpreter through env_kwargs:

from praxis_eval import EvalConfig

config = EvalConfig(
    task="widowx_carrot_on_plate",
    num_eval_per_task=5,
    num_parallel_env=2,
    output_dir="eval/simpler_carrot",
    env_kwargs={
        "python_bin": "/path/to/simpler-praxis/bin/python",
        "ms_asset_dir": "/data/simpler/maniskill_assets",
        "shader": "default",
    },
)

Caveats

• Evaluation launches external subprocesses and communicates with the policy through praxis-remote.
• env.python_bin is required for normal evaluation unless the caller supplies an equivalent runtime path.
• Recording videos uses a dedicated single-env pass when enabled.