Hugging Face's logo

Spaces:
Dvidia
/
robot-simulation-lab
Running

App Files Community
robot-simulation-lab / README.md
tarzelf's picture
tarzelf
Publish continuous shoe grip comparison and complete benchmark evidence
b9a3306 verified
preview code
|
Raw History Blame Contribute Delete
36.5 kB
metadata 
title: DVIDIA Robot Simulation Lab
emoji: 🤖
colorFrom: gray
colorTo: green
sdk: static
app_file: index.html
license: mit
pinned: false
tags:
  - robotics
  - mujoco
  - simulation
  - open-source

DVIDIA Robot Simulation Lab

Explore recorded native robot simulation, inspect every attempted task, and run the open-source tools on your own computer or server.

This Space contains the actual recorded UR5e + Robotiq 2F-85 rigid-block task: native meshes, full body poses, requested commands, grasp/carry/release attempts, matched open-jaw and hold controls, and the complete sealed source receipt. The browser replays recorded state; it does not rerun physics.

Frozen confirmation cohort: 18 planned attempts across 6 synthetic scenes. Counts are descriptive, with no population or physical reliability claim.

  • nominal / placement: 4/4 completed, 4 numerically valid
  • nominal / replay_open_jaw: 0/4 completed, 4 numerically valid
  • nominal / hold: 0/4 completed, 4 numerically valid
  • low_friction / placement: 2/2 completed, 2 numerically valid
  • low_friction / replay_open_jaw: 0/2 completed, 2 numerically valid
  • low_friction / hold: 0/2 completed, 2 numerically valid

These are small synthetic scenarios using an authored controller with privileged simulator state. They do not establish learned skills, human-video action recovery, shoe manipulation, hardware reliability, real-time camera reaction, or physical transfer. Replay selects an explicitly declared excerpt; the original full denominator and every attempt remain under replay/receipt/.

Install and reproduce locally 

python3 -m venv .venv
.venv/bin/python -m pip install "dvidia-training[arm] @ git+https://github.com/Dvidia-Inference/dvidia-training.git@0545ac47aedcd7c197341986533aa3e74c263ac6"
.venv/bin/dvidia-embodiment fetch-assets --output runs/reference-assets
.venv/bin/dvidia-reference-task benchmark runs/reference-assets --output runs/grasp-development
.venv/bin/dvidia-reference-task inspect runs/grasp-development
.venv/bin/dvidia-reference-task export-replay runs/reference-assets runs/grasp-development --output runs/grasp-replay

Use Python 3.12 or later and a platform supported by the pinned MuJoCo 3.15.0 and NumPy dependencies. The published native task ran on a CPU with no GPU. Native physics and learning are separate workloads; no minimum RAM or speed guarantee has been established for this task. WebGL2 is required only for the interactive browser viewer. Fetch dependencies and pinned assets once, then native simulation and exported replay work offline through a local HTTP server. The complete local task guide describes controls, scoring and measured timing boundaries.

Hardware profiles

UR5e + 2F-85 has the recorded native contact-task evidence above. Additional arms are listed with explicit implementation and evidence status in the gallery; an upcoming listing is not a verified simulator or a hardware adapter. No Space entry sends commands to a physical robot.

  • AgileX PiPER: 6 arm axes + 1 gripper axis; recorded unloaded native mechanics, not task-qualified. Complete report, traces, hashes and original MIT notice: profiles/piper/.
  • The Robot Studio SO101: 5 arm axes + 1 gripper axis; recorded unloaded native mechanics, not task-qualified. Complete report, traces, hashes and original Apache-2.0 notice: profiles/so101/.

PiPER contact pilot

Frozen confirmation cohort: 9 planned attempts across 3 synthetic scenes. Counts are descriptive, with no population or physical reliability claim.

  • nominal / placement: 2/2 completed, 2 numerically valid
  • nominal / replay_open_jaw: 0/2 completed, 2 numerically valid
  • nominal / hold: 0/2 completed, 2 numerically valid
  • low_friction / placement: 0/1 completed, 1 numerically valid
  • low_friction / replay_open_jaw: 0/1 completed, 1 numerically valid
  • low_friction / hold: 0/1 completed, 1 numerically valid

Authored controller, synthetic scenes and privileged native simulator state. Contact is uncalibrated; no learned skill or physical robot is qualified. This is separate from the unloaded PiPER mechanics bundle; no PiPER 3D viewer is included. Every attempted episode, including failed and invalid attempts, remains under tasks/piper/attempts/. Inspect the complete task report, frozen protocol, file hashes, and original MIT model license. Local PiPER contact guide.

PiPER guarded grip comparison

Placement completion: baseline 4/6 versus guarded 2/6.

This cohort does not show improved placement success.

Retries and qualification decisions do not imply successful recovery.

Frozen confirmation cohort: 30 planned attempts across six synthetic scenes. Two shared layouts at each friction; five variants per scene. Counts are descriptive, not physical reliability or detector accuracy.

  • μ=0.80 / baseline_placement: 2/2 completed, 2 numerically valid
  • μ=0.80 / baseline_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80 / guarded_placement: 2/2 completed, 2 numerically valid
  • μ=0.80 / guarded_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80 / hold: 0/2 completed, 2 numerically valid
  • μ=0.55 / baseline_placement: 2/2 completed, 2 numerically valid
  • μ=0.55 / baseline_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55 / guarded_placement: 0/2 completed, 2 numerically valid
  • μ=0.55 / guarded_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55 / hold: 0/2 completed, 2 numerically valid
  • μ=0.15 / baseline_placement: 0/2 completed, 2 numerically valid
  • μ=0.15 / baseline_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15 / guarded_placement: 0/2 completed, 2 numerically valid
  • μ=0.15 / guarded_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15 / hold: 0/2 completed, 2 numerically valid

Baseline and guarded authored controllers use privileged native state and their own matched open-jaw replays plus a common hold control. Full qualification decisions, unresolved outcomes, absent opportunities, alarms, retries and aborts remain in the source evidence. These synthetic trials are not detector accuracy, learned competence or physical qualification.

Every full attempt is stored as canonical deterministic gzip JSON under tasks/piper-grip/attempts/; the original seal records both compressed and decoded bytes and SHA-256. Download all files in tasks/piper-grip/ together and run the pinned dvidia-piper-grip inspect command from the local guide. No archive extraction is needed. Full comparison · Original protocol · Stored and decoded hashes.

PiPER grip recovery v0.4

Placement completion: v0.3 4/12, v0.4 carry-only 4/12, v0.4 full 4/12.

Retries, resumes and grip refusals do not imply successful placement or recovery. Shared successful scenes: 4; gained placements: 0; lost placements: 0. Mean completion-time change on shared successes: -1.270 simulated seconds (v0.4 minus v0.3). Hold-duration and post-resume motion remain separate in the report.

Frozen confirmation cohort: 84 planned attempts across twelve synthetic scenes. Three frictions, two masses and two shared layouts; seven variants per scene. Every failed or invalid attempt remains in the denominator. Counts are descriptive, not physical reliability or detector accuracy.

  • μ=0.80, mass=0.04 kg / history_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / history_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / recovery_carry_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / recovery_carry_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / recovery_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / history_placement: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / history_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / recovery_carry_placement: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / recovery_carry_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / recovery_placement: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / history_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / history_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / recovery_carry_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / recovery_carry_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / recovery_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / history_placement: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / history_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / recovery_carry_placement: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / recovery_carry_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / recovery_placement: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / history_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / history_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / recovery_carry_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / recovery_carry_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / recovery_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / history_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / history_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / recovery_carry_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / recovery_carry_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / recovery_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / hold: 0/2 completed, 2 numerically valid

Frozen v0.3, v0.4 carry-only and v0.4 full are authored controllers with privileged native state, each with its own matched open-jaw replay and a common hold control. Full diagnostics, contact-memory decisions, refusals, retries, resumes and aborts are retained. These synthetic trials establish no learned competence, physical reliability or detector accuracy.

Every attempt is canonical deterministic gzip JSON under tasks/piper-recovery-v04/attempts/; the original seal records both stored and decoded bytes and SHA-256. Download all files in tasks/piper-recovery-v04/ together for the pinned offline inspection workflow in the v0.4 guide. Full comparison · Every outcome · Original protocol · Stored and decoded hashes.

PiPER contact memory v0.3

Placement completion: baseline 4/12, probe-only 4/12, history 6/12.

Retries, resumes and release refusals do not imply successful placement or recovery.

Frozen confirmation cohort: 84 planned attempts across twelve synthetic scenes. Three frictions, two masses and two shared layouts; seven variants per scene. Every failed or invalid attempt remains in the denominator. Counts are descriptive, not physical reliability or detector accuracy.

  • μ=0.80, mass=0.04 kg / baseline_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / baseline_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / probe_only_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / probe_only_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / history_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / history_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / baseline_placement: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / baseline_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / probe_only_placement: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / probe_only_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / history_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / history_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / baseline_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / baseline_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / probe_only_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / probe_only_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / history_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / history_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / baseline_placement: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / baseline_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / probe_only_placement: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / probe_only_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / history_placement: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / history_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / baseline_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / baseline_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / probe_only_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / probe_only_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / history_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / history_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / baseline_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / baseline_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / probe_only_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / probe_only_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / history_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / history_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / hold: 0/2 completed, 2 numerically valid

Baseline, probe-only and history are authored controllers with privileged native state, each with its own matched open-jaw replay and a common hold control. Full diagnostics, contact-memory decisions, refusals, retries, resumes and aborts are retained. These synthetic trials establish no learned competence, physical reliability or detector accuracy.

Every attempt is canonical deterministic gzip JSON under tasks/piper-contact-v03/attempts/; the original seal records both stored and decoded bytes and SHA-256. Download all files in tasks/piper-contact-v03/ together for the pinned offline inspection workflow in the v0.3 guide. Full comparison · Every outcome · Original protocol · Stored and decoded hashes.

PiPER reliability v0.5 — confirmation-832847

Placement completion: v0.4 4/12, v0.5 second lift 6/12, v0.5 full 7/12.

v0.5 second lift versus v0.4: 2 gained, 0 lost, 4 shared successful scenes, 6 neither completed. Mean completion-time change on shared successes: +0.000 simulated seconds (candidate minus reference).

v0.5 full versus v0.4: 3 gained, 0 lost, 4 shared successful scenes, 5 neither completed. Mean completion-time change on shared successes: +0.847 simulated seconds (candidate minus reference).

v0.5 full versus v0.5 second lift: 1 gained, 0 lost, 6 shared successful scenes, 5 neither completed. Mean completion-time change on shared successes: +0.434 simulated seconds (candidate minus reference).

Protocol cohort: confirmation; seed 832847. 84 planned attempts across twelve synthetic scenes. Matched open-jaw and hold controls: 0/48 completed. Every failed or invalid attempt remains in the denominator. Cohorts are reported separately; a cohort label alone does not establish independence. Counts are descriptive synthetic evidence, not physical reliability, detector accuracy, learned competence or real-time sensor performance.

  • Pilot gate complete planned denominator: passed
  • Pilot gate all native attempts valid: passed
  • Pilot gate all controls fail: passed
  • Pilot gate native integrity clean: passed
  • Pilot gate gained without lost completions: passed
  • μ=0.80, mass=0.04 kg / recovery_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / reacquire_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / reacquire_grasp_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / recovery_placement: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / reacquire_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / reacquire_grasp_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / recovery_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / reacquire_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / reacquire_grasp_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / recovery_placement: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / reacquire_placement: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / reacquire_grasp_placement: 1/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / recovery_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / reacquire_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / reacquire_grasp_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / recovery_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / reacquire_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / reacquire_grasp_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / hold: 0/2 completed, 2 numerically valid

Complete comparison · Original seal. Download this entire cohort directory to retain all attempts and original notices.

PiPER reliability v0.5 — confirmation-227037

Placement completion: v0.4 6/12, v0.5 second lift 6/12, v0.5 full 8/12.

v0.5 second lift versus v0.4: 0 gained, 0 lost, 6 shared successful scenes, 6 neither completed. Mean completion-time change on shared successes: +0.000 simulated seconds (candidate minus reference).

v0.5 full versus v0.4: 2 gained, 0 lost, 6 shared successful scenes, 4 neither completed. Mean completion-time change on shared successes: +0.698 simulated seconds (candidate minus reference).

v0.5 full versus v0.5 second lift: 2 gained, 0 lost, 6 shared successful scenes, 4 neither completed. Mean completion-time change on shared successes: +0.698 simulated seconds (candidate minus reference).

Protocol cohort: confirmation; seed 227037. 84 planned attempts across twelve synthetic scenes. Matched open-jaw and hold controls: 0/48 completed. Every failed or invalid attempt remains in the denominator. Cohorts are reported separately; a cohort label alone does not establish independence. Counts are descriptive synthetic evidence, not physical reliability, detector accuracy, learned competence or real-time sensor performance.

  • Pilot gate complete planned denominator: passed
  • Pilot gate all native attempts valid: passed
  • Pilot gate all controls fail: passed
  • Pilot gate native integrity clean: passed
  • Pilot gate gained without lost completions: passed
  • μ=0.80, mass=0.04 kg / recovery_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / reacquire_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / reacquire_grasp_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / recovery_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / reacquire_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / reacquire_grasp_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / recovery_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / reacquire_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / reacquire_grasp_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / recovery_placement: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / reacquire_placement: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / reacquire_grasp_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / recovery_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / reacquire_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / reacquire_grasp_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / recovery_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / reacquire_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / reacquire_grasp_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / hold: 0/2 completed, 2 numerically valid

Complete comparison · Original seal. Download this entire cohort directory to retain all attempts and original notices.

PiPER reliability v0.5 — confirmation-862186

Placement completion: v0.4 5/12, v0.5 second lift 6/12, v0.5 full 8/12.

v0.5 second lift versus v0.4: 1 gained, 0 lost, 5 shared successful scenes, 6 neither completed. Mean completion-time change on shared successes: +0.000 simulated seconds (candidate minus reference).

v0.5 full versus v0.4: 3 gained, 0 lost, 5 shared successful scenes, 4 neither completed. Mean completion-time change on shared successes: +0.800 simulated seconds (candidate minus reference).

v0.5 full versus v0.5 second lift: 2 gained, 0 lost, 6 shared successful scenes, 4 neither completed. Mean completion-time change on shared successes: +0.602 simulated seconds (candidate minus reference).

Protocol cohort: confirmation; seed 862186. 84 planned attempts across twelve synthetic scenes. Matched open-jaw and hold controls: 0/48 completed. Every failed or invalid attempt remains in the denominator. Cohorts are reported separately; a cohort label alone does not establish independence. Counts are descriptive synthetic evidence, not physical reliability, detector accuracy, learned competence or real-time sensor performance.

  • Pilot gate complete planned denominator: passed
  • Pilot gate all native attempts valid: passed
  • Pilot gate all controls fail: passed
  • Pilot gate native integrity clean: passed
  • Pilot gate gained without lost completions: passed
  • μ=0.80, mass=0.04 kg / recovery_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / reacquire_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / reacquire_grasp_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.04 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / recovery_placement: 1/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / reacquire_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / reacquire_grasp_placement: 2/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.80, mass=0.06 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / recovery_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / reacquire_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / reacquire_grasp_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.04 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / recovery_placement: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / reacquire_placement: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / reacquire_grasp_placement: 2/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.55, mass=0.06 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / recovery_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / reacquire_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / reacquire_grasp_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.04 kg / hold: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / recovery_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / recovery_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / reacquire_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / reacquire_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / reacquire_grasp_placement: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / reacquire_grasp_open_jaw: 0/2 completed, 2 numerically valid
  • μ=0.15, mass=0.06 kg / hold: 0/2 completed, 2 numerically valid

Complete comparison · Original seal. Download this entire cohort directory to retain all attempts and original notices.

The reviewed publication now permits at most 512 MiB of stored files, to retain prior evidence alongside complete new comparisons and their failed attempts. The 32 MiB per-file limit and bounded single-member gzip decoding remain in place. This is a packaging limit, not a hardware or training memory requirement.

Provenance and licensing

The native export, its complete receipt, model license files, upstream notices, and original integrity bytes are retained without rewriting. See provenance.json, replay/replay-integrity.json, replay/receipt/integrity.json, replay/licenses/ and replay/reference_replay.third-party.txt. Hashes establish artifact consistency, not authenticity or physical qualification.

DVIDIA task/exporter/viewer software and bundled Three.js are MIT. The native UR5e model is BSD-3-Clause; the Robotiq 2F-85 model is BSD-2-Clause. Component license notices apply to their own files; the Space card's MIT label does not relicense upstream model assets. Additional profile assets retain their own notices. No neural policy weights, private footage, credentials or paid server compute are included.

DVIDIA · Research · GitHub · @iammrriver · hello@dvidia.org

Shoe asset library v0.1

Explore the asset library: six original rigid shoe proxies and UR5e/Robotiq, PiPER and SO-101 scenes. Complete benchmark. Native loading and support tests do not establish shoe grasping, deformation or physical transfer.

Live PiPER shoe workstation v0.1

Open the live workstation. MuJoCo 3.15.0 WASM runs in a browser worker. Drag targets, bounded joint commands, native gripper actuation and free-object contact are live. The shoe is a rigid synthetic proxy with uncalibrated properties. No learned skill or physical robot qualification. Complete evidence and failures, evidence file hashes, and research report. DVIDIA code and shoes: MIT; PiPER assets: MIT; Three.js: MIT; MuJoCo: Apache-2.0 with retained dependency notices.

The live workstation now loads one compressed native mesh bundle. Original meshes, physics parameters and original benchmark evidence are retained. Delivery validation.

The workstation download fix adds compressed WASM transport and bounded download recovery. Original engine, mesh and benchmark bytes are retained. Download-fix validation.

The Hugging Face workstation can load its three compressed simulator resources from DVIDIA’s research mirror, with exact byte bounds and SHA-256 verification. Original engine, mesh, geometry and benchmark bytes are retained. Mirror validation.

Skill prerequisites v0.1

Explore both simulations. Recorded replays of native MuJoCo runs; run the open-source Python environments locally for fresh physics. The rake is mounted, not autonomously acquired. Lid tests measure access and recovery, not shoe extraction. Failures and full benchmark evidence remain visible.

Pinned source · Research

Continuous shoe grip v0.2

Explore the grip comparison. Recorded replays of native MuJoCo runs compare the previous mounted-rake controller with continuous acquisition and grip-loss checks on matched scenes. Run the open-source Python benchmark locally for fresh physics. All planned attempts and independent outcome scores remain visible. Policies use privileged simulator observations; contact is synthetic and hardware transfer remains open.

Pinned source · Research