Spaces:
Running
Running
Publish reviewed DVIDIA native robot simulation evidence and local tools
Browse files- README.md +2 -2
- index.html +2 -2
- provenance.json +2 -2
- replay/reference_replay.bundle.js +0 -0
- replay/replay-integrity.json +1 -1
README.md
CHANGED
|
@@ -43,7 +43,7 @@ denominator and every attempt remain under `replay/receipt/`.
|
|
| 43 |
|
| 44 |
```sh
|
| 45 |
python3 -m venv .venv
|
| 46 |
-
.venv/bin/python -m pip install "dvidia-training[arm] @ git+https://github.com/Dvidia-Inference/dvidia-training.git@
|
| 47 |
.venv/bin/dvidia-embodiment fetch-assets --output runs/reference-assets
|
| 48 |
.venv/bin/dvidia-reference-task benchmark runs/reference-assets --output runs/grasp-development
|
| 49 |
.venv/bin/dvidia-reference-task inspect runs/grasp-development
|
|
@@ -56,7 +56,7 @@ physics and learning are separate workloads; no minimum RAM or speed guarantee
|
|
| 56 |
has been established for this task. WebGL2 is required only for the interactive
|
| 57 |
browser viewer. Fetch dependencies and pinned assets once, then native simulation
|
| 58 |
and exported replay work offline through a local HTTP server. The complete
|
| 59 |
-
[local task guide](https://github.com/Dvidia-Inference/dvidia-training/blob/
|
| 60 |
describes controls, scoring and measured timing boundaries.
|
| 61 |
|
| 62 |
## Hardware profiles
|
|
|
|
| 43 |
|
| 44 |
```sh
|
| 45 |
python3 -m venv .venv
|
| 46 |
+
.venv/bin/python -m pip install "dvidia-training[arm] @ git+https://github.com/Dvidia-Inference/dvidia-training.git@525571b739c2530f7cbf570fc6ade0d5d2764fb9"
|
| 47 |
.venv/bin/dvidia-embodiment fetch-assets --output runs/reference-assets
|
| 48 |
.venv/bin/dvidia-reference-task benchmark runs/reference-assets --output runs/grasp-development
|
| 49 |
.venv/bin/dvidia-reference-task inspect runs/grasp-development
|
|
|
|
| 56 |
has been established for this task. WebGL2 is required only for the interactive
|
| 57 |
browser viewer. Fetch dependencies and pinned assets once, then native simulation
|
| 58 |
and exported replay work offline through a local HTTP server. The complete
|
| 59 |
+
[local task guide](https://github.com/Dvidia-Inference/dvidia-training/blob/525571b739c2530f7cbf570fc6ade0d5d2764fb9/docs/native-task-benchmark.md)
|
| 60 |
describes controls, scoring and measured timing boundaries.
|
| 61 |
|
| 62 |
## Hardware profiles
|
index.html
CHANGED
|
@@ -5,6 +5,6 @@
|
|
| 5 |
<h1>See the motion. Inspect the evidence.</h1><p>A native six-joint arm, an articulated gripper and a free rigid object. Recorded task attempts make the simulator's behavior visible, and the complete receipts make its limits checkable.</p><p class="scope">This is recorded simulation from an authored controller with privileged simulator state. The browser does not run physics or connect to hardware. Learned skills, shoe manipulation and physical transfer remain future stages.</p>
|
| 6 |
<div class="summary"><strong>Frozen confirmation cohort: 18 planned attempts across 6 synthetic scenes. Counts are descriptive, with no population or physical reliability claim.</strong><br>nominal / placement: 4/4 completed, 4 numerically valid<br>nominal / replay_open_jaw: 0/4 completed, 4 numerically valid<br>nominal / hold: 0/4 completed, 4 numerically valid<br>low_friction / placement: 2/2 completed, 2 numerically valid<br>low_friction / replay_open_jaw: 0/2 completed, 2 numerically valid<br>low_friction / hold: 0/2 completed, 2 numerically valid</div>
|
| 7 |
<section aria-labelledby="replay-title"><h2 id="replay-title">Native grasp, carry and release.</h2><div class="links"><a href="replay/index.html">Open the full 3D replay ↗</a><a href="replay/receipt/report.html">Every attempted task ↗</a><a href="README.md">Install and reproduce ↗</a></div><div class="replay"><iframe src="replay/index.html" title="DVIDIA recorded native robot task replay" allow="fullscreen"></iframe></div><p class="tiny">Use orbit, zoom, recorded motion selection and the timeline inside the viewer. Replay is an excerpt; the full benchmark includes every matched control and invalid attempt. The native meshes, binary poses, complete receipts and license notices are served locally by this Space.</p></section>
|
| 8 |
-
<section aria-labelledby="profiles-title"><h2 id="profiles-title">Hardware profiles and the next experiments.</h2><div class="gallery"><article class="profile"><span class="status">Recorded native contact task</span><h2>UR5e + Robotiq 2F-85</h2><p>Six arm joints, eight articulated gripper coordinates and a free rigid block. Full source outcomes and matched controls are available.</p><a href="replay/index.html">Inspect the task replay ↗</a></article><article class="profile"><span class="status">Recorded native mechanics · unloaded</span><h2>AgileX PiPER</h2><p>6 arm axes + 1 gripper axis. Native model structure and authored unloaded motion are measured. No contact task, learned policy or physical robot is qualified. Simulation asset license: MIT.</p><div class="links"><a href="profiles/piper/report.html">Numerical report ↗</a><a href="profiles/piper/report.json">Complete recorded states ↗</a><a href="profiles/piper/integrity.json">File hashes ↗</a><a href="profiles/piper/licenses/piper/LICENSE">Model license ↗</a></div><a href="https://github.com/Dvidia-Inference/dvidia-training/blob/
|
| 9 |
-
<section><h2>Run it on your own machine.</h2><p>The native task is a local CPU workload. Fetch the pinned dependencies and assets once, then simulate offline. WebGL2 supports viewing; no hosted inference or paid GPU server is requested by this static Space. Minimum RAM and arbitrary-dataset speed guarantees remain unmeasured.</p><div class="links"><a href="https://github.com/Dvidia-Inference/dvidia-training/blob/
|
| 10 |
<footer><span>Open-source DVIDIA research</span><a href="https://x.com/iammrriver">@iammrriver ↗</a><a href="mailto:hello@dvidia.org">hello@dvidia.org</a><a href="https://huggingface.co/Dvidia">Hugging Face organization ↗</a></footer></main></body></html>
|
|
|
|
| 5 |
<h1>See the motion. Inspect the evidence.</h1><p>A native six-joint arm, an articulated gripper and a free rigid object. Recorded task attempts make the simulator's behavior visible, and the complete receipts make its limits checkable.</p><p class="scope">This is recorded simulation from an authored controller with privileged simulator state. The browser does not run physics or connect to hardware. Learned skills, shoe manipulation and physical transfer remain future stages.</p>
|
| 6 |
<div class="summary"><strong>Frozen confirmation cohort: 18 planned attempts across 6 synthetic scenes. Counts are descriptive, with no population or physical reliability claim.</strong><br>nominal / placement: 4/4 completed, 4 numerically valid<br>nominal / replay_open_jaw: 0/4 completed, 4 numerically valid<br>nominal / hold: 0/4 completed, 4 numerically valid<br>low_friction / placement: 2/2 completed, 2 numerically valid<br>low_friction / replay_open_jaw: 0/2 completed, 2 numerically valid<br>low_friction / hold: 0/2 completed, 2 numerically valid</div>
|
| 7 |
<section aria-labelledby="replay-title"><h2 id="replay-title">Native grasp, carry and release.</h2><div class="links"><a href="replay/index.html">Open the full 3D replay ↗</a><a href="replay/receipt/report.html">Every attempted task ↗</a><a href="README.md">Install and reproduce ↗</a></div><div class="replay"><iframe src="replay/index.html" title="DVIDIA recorded native robot task replay" allow="fullscreen"></iframe></div><p class="tiny">Use orbit, zoom, recorded motion selection and the timeline inside the viewer. Replay is an excerpt; the full benchmark includes every matched control and invalid attempt. The native meshes, binary poses, complete receipts and license notices are served locally by this Space.</p></section>
|
| 8 |
+
<section aria-labelledby="profiles-title"><h2 id="profiles-title">Hardware profiles and the next experiments.</h2><div class="gallery"><article class="profile"><span class="status">Recorded native contact task</span><h2>UR5e + Robotiq 2F-85</h2><p>Six arm joints, eight articulated gripper coordinates and a free rigid block. Full source outcomes and matched controls are available.</p><a href="replay/index.html">Inspect the task replay ↗</a></article><article class="profile"><span class="status">Recorded native mechanics · unloaded</span><h2>AgileX PiPER</h2><p>6 arm axes + 1 gripper axis. Native model structure and authored unloaded motion are measured. No contact task, learned policy or physical robot is qualified. Simulation asset license: MIT.</p><div class="links"><a href="profiles/piper/report.html">Numerical report ↗</a><a href="profiles/piper/report.json">Complete recorded states ↗</a><a href="profiles/piper/integrity.json">File hashes ↗</a><a href="profiles/piper/licenses/piper/LICENSE">Model license ↗</a></div><a href="https://github.com/Dvidia-Inference/dvidia-training/blob/525571b739c2530f7cbf570fc6ade0d5d2764fb9/docs/robot-simulation-library.md">Pinned local simulator guide ↗</a></article><article class="profile"><span class="status">Recorded native mechanics · unloaded</span><h2>The Robot Studio SO101</h2><p>5 arm axes + 1 gripper axis. Native model structure and authored unloaded motion are measured. No contact task, learned policy or physical robot is qualified. Simulation asset license: Apache-2.0.</p><div class="links"><a href="profiles/so101/report.html">Numerical report ↗</a><a href="profiles/so101/report.json">Complete recorded states ↗</a><a href="profiles/so101/integrity.json">File hashes ↗</a><a href="profiles/so101/licenses/so101/LICENSE">Model license ↗</a></div><a href="https://github.com/Dvidia-Inference/dvidia-training/blob/525571b739c2530f7cbf570fc6ade0d5d2764fb9/docs/robot-simulation-library.md">Pinned local simulator guide ↗</a></article><article class="profile"><span class="status">Upcoming · no task evidence</span><h2>Two arms and dexterous hands</h2><p>The roadmap extends to bilateral manipulation and five-finger hands. Contact tasks, sensing adapters and physical qualification remain to be built.</p><a href="https://research.dvidia.org/">Follow the research ↗</a></article></div><p class="tiny">A native mechanics benchmark establishes model structure and recorded motion. A native task benchmark adds measured contact outcomes. Neither establishes hardware reliability or a learned motor policy. Upcoming profiles have no verified implementation claim.</p></section>
|
| 9 |
+
<section><h2>Run it on your own machine.</h2><p>The native task is a local CPU workload. Fetch the pinned dependencies and assets once, then simulate offline. WebGL2 supports viewing; no hosted inference or paid GPU server is requested by this static Space. Minimum RAM and arbitrary-dataset speed guarantees remain unmeasured.</p><div class="links"><a href="https://github.com/Dvidia-Inference/dvidia-training/blob/525571b739c2530f7cbf570fc6ade0d5d2764fb9/docs/native-task-benchmark.md">Local installation and benchmark guide ↗</a><a href="provenance.json">Artifact provenance ↗</a><a href="replay/replay-integrity.json">Replay file hashes ↗</a><a href="replay/licenses/universal_robots_ur5e/LICENSE">UR5e model license ↗</a><a href="replay/licenses/robotiq_2f85/LICENSE">Gripper model license ↗</a></div></section>
|
| 10 |
<footer><span>Open-source DVIDIA research</span><a href="https://x.com/iammrriver">@iammrriver ↗</a><a href="mailto:hello@dvidia.org">hello@dvidia.org</a><a href="https://huggingface.co/Dvidia">Hugging Face organization ↗</a></footer></main></body></html>
|
provenance.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
"schema": "dvidia.robot-simulation-space-provenance/v1",
|
| 3 |
"scope": "Recorded native simulation evidence; authored controller, no learned policy or physical qualification",
|
| 4 |
-
"source_commit": "
|
| 5 |
"physical_robot_ready": false,
|
| 6 |
"learned_policy": false,
|
| 7 |
"live_browser_physics": false,
|
|
@@ -11,7 +11,7 @@
|
|
| 11 |
"replay_excerpt_note": "Selected attempts are not the full success-rate denominator",
|
| 12 |
"source_replay_integrity": {
|
| 13 |
"bytes": 4876,
|
| 14 |
-
"sha256": "
|
| 15 |
},
|
| 16 |
"source_report": {
|
| 17 |
"bytes": 107300,
|
|
|
|
| 1 |
{
|
| 2 |
"schema": "dvidia.robot-simulation-space-provenance/v1",
|
| 3 |
"scope": "Recorded native simulation evidence; authored controller, no learned policy or physical qualification",
|
| 4 |
+
"source_commit": "525571b739c2530f7cbf570fc6ade0d5d2764fb9",
|
| 5 |
"physical_robot_ready": false,
|
| 6 |
"learned_policy": false,
|
| 7 |
"live_browser_physics": false,
|
|
|
|
| 11 |
"replay_excerpt_note": "Selected attempts are not the full success-rate denominator",
|
| 12 |
"source_replay_integrity": {
|
| 13 |
"bytes": 4876,
|
| 14 |
+
"sha256": "f8f64bf5a09ea4aa20e779a1eb0103effde38fcf8e12fddb2772fb62e349d3f2"
|
| 15 |
},
|
| 16 |
"source_report": {
|
| 17 |
"bytes": 107300,
|
replay/reference_replay.bundle.js
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
replay/replay-integrity.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"schema":"dvidia.reference-replay-integrity/v1","scope":"Artifact consistency only; no authenticity or physical qualification","files":{"scene.json":{"bytes":30608,"sha256":"d0577fed664bfa1ad1d6b58d8489ce2d886a6126e43ad4c7829e7df309b82176"},"motion.json":{"bytes":8091934,"sha256":"a73f5e02e34205f1812e4286e69be56fa17277049727e9d8f5b4457c577c5487"},"geometry.bin":{"bytes":3374124,"sha256":"8349bbefd55572e69151d8d21a5e6ae1fd2f80153f2782db26d72729a9bdf9ec"},"motion.bin":{"bytes":6054048,"sha256":"59efb4d7728963542a386af7825b3efc0084b322ec218b9f4dca618d7133dd9e"},"geometry.bin.gz":{"bytes":1831265,"sha256":"cc162219dac340db4c663b34b3e54b0e5574679a7f737fa90853129d8a68978f"},"motion.bin.gz":{"bytes":1688294,"sha256":"c313220899fc40c0dec88df7a46925f4d036565241b251ba51162f22a1e2f529"},"LICENSE":{"bytes":1076,"sha256":"ab47d245880a661fdca3c6a4cb91671031302deee50d64f8d2d6df86832962e1"},"index.html":{"bytes":13937,"sha256":"1a321f181420e5c923943da6cb8eb3e27e73e26fd625822b8a21b63de8c27033"},"reference_replay.bundle.js":{"bytes":
|
|
|
|
| 1 |
+
{"schema":"dvidia.reference-replay-integrity/v1","scope":"Artifact consistency only; no authenticity or physical qualification","files":{"scene.json":{"bytes":30608,"sha256":"d0577fed664bfa1ad1d6b58d8489ce2d886a6126e43ad4c7829e7df309b82176"},"motion.json":{"bytes":8091934,"sha256":"a73f5e02e34205f1812e4286e69be56fa17277049727e9d8f5b4457c577c5487"},"geometry.bin":{"bytes":3374124,"sha256":"8349bbefd55572e69151d8d21a5e6ae1fd2f80153f2782db26d72729a9bdf9ec"},"motion.bin":{"bytes":6054048,"sha256":"59efb4d7728963542a386af7825b3efc0084b322ec218b9f4dca618d7133dd9e"},"geometry.bin.gz":{"bytes":1831265,"sha256":"cc162219dac340db4c663b34b3e54b0e5574679a7f737fa90853129d8a68978f"},"motion.bin.gz":{"bytes":1688294,"sha256":"c313220899fc40c0dec88df7a46925f4d036565241b251ba51162f22a1e2f529"},"LICENSE":{"bytes":1076,"sha256":"ab47d245880a661fdca3c6a4cb91671031302deee50d64f8d2d6df86832962e1"},"index.html":{"bytes":13937,"sha256":"1a321f181420e5c923943da6cb8eb3e27e73e26fd625822b8a21b63de8c27033"},"reference_replay.bundle.js":{"bytes":536928,"sha256":"72d3729db5ed72f0c73fcae1e105478bd55c3fac43f05355337a69403af90c21"},"reference_replay.third-party.txt":{"bytes":1199,"sha256":"ea4da6cf7a8a54649702f0eb8d057c8ced12f8fed2329c5e96d92a3316922813"},"licenses/universal_robots_ur5e/LICENSE":{"bytes":1470,"sha256":"5ec71ccf66c8d03261448f2441a586765b97f2248c860a1ae19689fb1c45cee6"},"licenses/universal_robots_ur5e/README.md":{"bytes":1689,"sha256":"435600d4f00cfdbd7edec4b60f0cd7fbb539f39abaefcd8bb3994beed3b1156a"},"licenses/robotiq_2f85/LICENSE":{"bytes":1297,"sha256":"3d5edd18d1e22a17666e9bb84270171ed5ac90b192d24e0177fd171708ea2316"},"licenses/robotiq_2f85/README.md":{"bytes":1367,"sha256":"a10f34027d417f8f7b9102b74b42976031f0246f3090444728ab322bc4632aaf"},"receipt/report.json":{"bytes":107300,"sha256":"025752735310f131977d6758ae12ae4251f68b723a45d65d87dc00167d11355f"},"receipt/report.html":{"bytes":7288,"sha256":"554e5fa164cdbb46fcb030bf30a82898f9926914ab2fe022cb4f5e20a9b02293"},"receipt/protocol.json":{"bytes":7606,"sha256":"313d0be81ad13a0e20c218f88983650f44a65ea590d955c8d30adda47231c0a6"},"receipt/attempts/nominal-00--placement.json":{"bytes":3401073,"sha256":"853e74a126f737bff945b051f55b3794640d1279a780bdff87adada230b4f7e2"},"receipt/attempts/nominal-00--replay_open_jaw.json":{"bytes":3422242,"sha256":"b475692ec41074d1f029fd3f26b9801849f44d497611dd1653084674ed2707b6"},"receipt/attempts/nominal-00--hold.json":{"bytes":3422417,"sha256":"55bfac500c29adec913ea095ccd82af38dadac417e696191fe016402e5d0fb70"},"receipt/attempts/nominal-01--placement.json":{"bytes":3402489,"sha256":"57db0c7e46facf5eef598e6810a2695034b102c03e80ac927298bec2ac36493b"},"receipt/attempts/nominal-01--replay_open_jaw.json":{"bytes":3424746,"sha256":"f4d1721c5c453c91e047f0e99ce30c6a2664fdb35bbce5841f23350c83f2511d"},"receipt/attempts/nominal-01--hold.json":{"bytes":3424354,"sha256":"f5472973a88acb1bc90e41d30ec9af7b543863def6e2438f25aadfa4a453513d"},"receipt/attempts/nominal-02--placement.json":{"bytes":3399433,"sha256":"819d93d64fbce0b8c5bdf48260e59a1d8b44931d96763024f5dac366eaea0f3c"},"receipt/attempts/nominal-02--replay_open_jaw.json":{"bytes":3419980,"sha256":"af39f63c750defb1035982d09310d309615499afb8090d941e443032f7c8e5f2"},"receipt/attempts/nominal-02--hold.json":{"bytes":3419656,"sha256":"d624787c798e0485f727737a7c69714b48a3c4751af905f736e8ba66532ff665"},"receipt/attempts/nominal-03--placement.json":{"bytes":3399100,"sha256":"6b99ef0f93e4b3fdf95704c56a4367391f96a3d4b93d97d464f8b9ec8ba74667"},"receipt/attempts/nominal-03--replay_open_jaw.json":{"bytes":3423116,"sha256":"a345d09eddeb5d2e81d87cf90dd84993534787f55703b7b7d3751c46664fe008"},"receipt/attempts/nominal-03--hold.json":{"bytes":3425502,"sha256":"384087f0020567ee52beab36774ae9a5fe5d3ad8453d986aee6a9518dd109e81"},"receipt/attempts/low_friction-00--placement.json":{"bytes":3403473,"sha256":"c604e8d0d1b372c63abf2b85068863cc33bd353dcd365a3bef973c266e6d206c"},"receipt/attempts/low_friction-00--replay_open_jaw.json":{"bytes":3431280,"sha256":"0b5c3d4895012272b601f7b658623e3c3f855b936c420860e5954313a80803e4"},"receipt/attempts/low_friction-00--hold.json":{"bytes":3431260,"sha256":"a522d6af3586a1289c2fe70ec66d6db4781aac060b76811261f1bd610e757f9d"},"receipt/attempts/low_friction-01--placement.json":{"bytes":3393258,"sha256":"39b00b3cc527d9aeafb5f48cdac21033332e79a92055c5760da69903360b97fe"},"receipt/attempts/low_friction-01--replay_open_jaw.json":{"bytes":3422169,"sha256":"c6b1c12a0aa59feb1afa8c73f892e1cf33d6304a96c724dd1135a0c2df3b56a3"},"receipt/attempts/low_friction-01--hold.json":{"bytes":3425168,"sha256":"916358cc535793f31ae5e8d3f72d5ff2f6df53d63dcf8a5921a1306da95a02be"},"receipt/integrity.json":{"bytes":2789,"sha256":"d72d3bbd1b443d735437dc18fed875e0f957886a7b05bd0aac3424ac195905a0"},"README.md":{"bytes":1279,"sha256":"fb5bb13accb7b58d23932184ec466728ee67b4ebb3ec53684bd546ea2150ab97"}}}
|