UR5e + Robotiq
6 arm axes + gripper. BSD model assets.
Explore shoes with UR5e + Robotiq →diff --git a/docs/downloads/reports/Shoe asset library v01.md b/docs/downloads/reports/Shoe asset library v01.md index b6526c1..5b2779e 100644 --- a/docs/downloads/reports/Shoe asset library v01.md +++ b/docs/downloads/reports/Shoe asset library v01.md @@ -77,7 +77,7 @@ worst-case latency bound. Rendering, physical sensing, transport and actuator response are outside these timings. A short passive scene does not establish training throughput or a real-time robot control guarantee. -The downloadable library contains **64 files totaling 21,197,720 bytes**, +The downloadable library contains **64 files totaling 21,197,840 bytes**, including all three recorded native previews, their original license notices, portable scene descriptors and the complete benchmark. Initial pinned robot downloads and installed dependencies require additional storage. @@ -93,18 +93,20 @@ The next shoe-specific step is a feasible grasp and placement task on one select The frozen benchmark runtime is [`441ee2871620104246ab9caeb89825f12a14b889`](https://github.com/Dvidia-Inference/dvidia-training/tree/441ee2871620104246ab9caeb89825f12a14b889). The publication and installation guide use the source revision -[`45bef6315c9632b33915e1ce993babd2926b8f7d`](https://github.com/Dvidia-Inference/dvidia-training/tree/45bef6315c9632b33915e1ce993babd2926b8f7d). -The corrected viewer was built from the preceding source revision +[`d291ee11f28b0d3c7a0b9ed062174e46fe764d15`](https://github.com/Dvidia-Inference/dvidia-training/tree/d291ee11f28b0d3c7a0b9ed062174e46fe764d15). +The browser JavaScript was built from the earlier source revision [`17a7c81784362697f09757712efbe00d98743501`](https://github.com/Dvidia-Inference/dvidia-training/tree/17a7c81784362697f09757712efbe00d98743501). -The later documentation correction makes the all-family export command explicit; -it changes neither the recorded previews nor the benchmark. -The publication preserves the original benchmark without rerunning its physics. +The publication updates HTML navigation to explicit `index.html` links for static +hosting while preserving that JavaScript and the native recordings. No physics +benchmark was rerun. The [viewer build receipt](https://research.dvidia.org/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/viewer-build.json) -records the corrected browser bundle and updated file seals. Native geometry, -motion, scene descriptors and license notices remain byte-identical to the frozen -exports. The original execution receipt retains its original preview byte counts; -the current library seal identifies the published viewer files. -The [installation and loading guide](https://github.com/Dvidia-Inference/dvidia-training/blob/45bef6315c9632b33915e1ce993babd2926b8f7d/docs/asset-library.md) +records the corrected browser bundle; the subsequent +[navigation build receipt](https://research.dvidia.org/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/navigation-build.json) +binds the published HTML and updated seals. Native geometry, motion, scene +descriptors and license notices remain byte-identical to the frozen exports. +The original execution receipt retains its original preview byte counts; the +current library seal identifies the published viewer files. +The [installation and loading guide](https://github.com/Dvidia-Inference/dvidia-training/blob/d291ee11f28b0d3c7a0b9ed062174e46fe764d15/docs/asset-library.md) explains provisioning the three robot profiles and opening a local preview. The [benchmark protocol](https://github.com/Dvidia-Inference/dvidia-training/blob/441ee2871620104246ab9caeb89825f12a14b889/docs/asset-library-benchmark-protocol.md) defines the exact support, gravity and unexpected-contact criteria. diff --git a/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/README.md b/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/README.md index 4f7528f..68e8508 100644 --- a/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/README.md +++ b/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/README.md @@ -2,11 +2,11 @@ Frozen benchmark source: [`441ee2871620104246ab9caeb89825f12a14b889`](https://github.com/Dvidia-Inference/dvidia-training/tree/441ee2871620104246ab9caeb89825f12a14b889). -Publication and installation-guide source: [`45bef6315c9632b33915e1ce993babd2926b8f7d`](https://github.com/Dvidia-Inference/dvidia-training/tree/45bef6315c9632b33915e1ce993babd2926b8f7d). +Publication and installation-guide source: [`d291ee11f28b0d3c7a0b9ed062174e46fe764d15`](https://github.com/Dvidia-Inference/dvidia-training/tree/d291ee11f28b0d3c7a0b9ed062174e46fe764d15). -Corrected viewer source: [`17a7c81784362697f09757712efbe00d98743501`](https://github.com/Dvidia-Inference/dvidia-training/tree/17a7c81784362697f09757712efbe00d98743501). The later documentation correction makes the all-family export command explicit; native previews and the original benchmark are retained without rerunning their physics. +Browser JavaScript source: [`17a7c81784362697f09757712efbe00d98743501`](https://github.com/Dvidia-Inference/dvidia-training/tree/17a7c81784362697f09757712efbe00d98743501). The publication updates HTML navigation to explicit `index.html` links for static hosting; the JavaScript, native recordings and original benchmark are retained without rerunning physics. -The [viewer build receipt](viewer-build.json) records the corrected browser bundle and updated seals. All native geometry, motion, descriptors and notices remain byte-identical. `execution.json` retains the original preview byte counts; `library/library-integrity.json` identifies the current 64 files totaling 21,197,720 bytes. +The [viewer build receipt](viewer-build.json) records the corrected browser bundle. The later [navigation build receipt](navigation-build.json) binds the published HTML and updated seals. All native geometry, motion, descriptors and notices remain byte-identical. `execution.json` retains the original preview byte counts; `library/library-integrity.json` identifies the current 64 files totaling 21,197,840 bytes. - Six original MIT rigid shoe proxies: trainer, ankle-boot and slipper left/right pairs. - UR5e/Robotiq, PiPER and SO-101 native scenes, portable descriptors and local 3D previews. @@ -36,7 +36,7 @@ Across 27,000 native `mj_step` calls, median **8.50 µs**, p95 **15.75 µs**, p9 These are synthetic rigid proxies with untested mass/contact assumptions. Native loading, gravity and passive support do not establish shoe grasping, deformation, learned competence or physical transfer. Geometry is intentionally outside the held robot's immediate contacts; reachability and graspability are unqualified. -See the [installation and scene-loading guide](https://github.com/Dvidia-Inference/dvidia-training/blob/45bef6315c9632b33915e1ce993babd2926b8f7d/docs/asset-library.md) and [benchmark protocol](https://github.com/Dvidia-Inference/dvidia-training/blob/441ee2871620104246ab9caeb89825f12a14b889/docs/asset-library-benchmark-protocol.md). Fetch UR5e into a fresh root first, then add PiPER and SO-101. After provisioning: +See the [installation and scene-loading guide](https://github.com/Dvidia-Inference/dvidia-training/blob/d291ee11f28b0d3c7a0b9ed062174e46fe764d15/docs/asset-library.md) and [benchmark protocol](https://github.com/Dvidia-Inference/dvidia-training/blob/441ee2871620104246ab9caeb89825f12a14b889/docs/asset-library-benchmark-protocol.md). Fetch UR5e into a fresh root first, then add PiPER and SO-101. After provisioning: ```sh dvidia-assets benchmark --assets ./robot-assets --seed 612047 --repeats 3 --source-commit 441ee2871620104246ab9caeb89825f12a14b889 --output ./asset-benchmark diff --git a/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/index.html b/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/index.html index b23f656..07f71d6 100644 --- a/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/index.html +++ b/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/index.html @@ -1,11 +1,11 @@
VERSION 0.1 · LOCAL CPU SIMULATION
Six original rigid shoe proxies across trainers, ankle boots and slippers, with left/right identities, three supported robot models and reusable scene files. Inspect the native 3D geometry, choose a pair, then download the scene and run it locally.
6 arm axes + gripper. BSD model assets.
Explore shoes with UR5e + Robotiq →6 arm axes + gripper. MIT model assets.
Explore shoes with PiPER →5 arm axes + gripper. Apache-2.0 model assets.
Explore shoes with SO-101 →These are synthetic geometry and contact-development assets. Dimensions, mass and friction are declared assumptions. Deformation, shoe grasping and hardware transfer have not been validated. The browser replays recorded settling; native physics runs locally.
git clone https://github.com/Dvidia-Inference/dvidia-training.git
cd dvidia-training
-git checkout 45bef6315c9632b33915e1ce993babd2926b8f7d
+git checkout d291ee11f28b0d3c7a0b9ed062174e46fe764d15
python3.12 -m venv .venv
.venv/bin/python -m pip install '.[arm]'
.venv/bin/dvidia-assets fetch piper --output runs/assets
diff --git a/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/library-integrity.json b/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/library-integrity.json
index a465727..846d8ff 100644
--- a/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/library-integrity.json
+++ b/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/library-integrity.json
@@ -35,7 +35,7 @@
},
"index.html": {
"bytes": 4328,
- "sha256": "56a8cc13b3a319e4a599011606e876c09b14ce458dc5f47331095a035c95a419"
+ "sha256": "1cd2d49fb548efe9652a635c84453c20a1c8d190ba58d153caf4e2beaa8f2161"
},
"previews/piper/LICENSE": {
"bytes": 1076,
@@ -58,8 +58,8 @@
"sha256": "51365a5a661d8062e28c66abc6bdb843ee410811f84b49d62c5878c265b3631f"
},
"previews/piper/index.html": {
- "bytes": 9636,
- "sha256": "41a97a426a2b72520e39a469eaa78f4b281c7fef9cf9607301af9c398594251f"
+ "bytes": 9676,
+ "sha256": "e8bb5c8fc00e454e3abaa4ffbdfbc48ab1c1e91114030028247d14cba92620a0"
},
"previews/piper/licenses/agilex_piper/LICENSE": {
"bytes": 1065,
@@ -83,7 +83,7 @@
},
"previews/piper/preview-integrity.json": {
"bytes": 2364,
- "sha256": "3b92ff77fea9983f6422b94bda96fa1376ef7379e7fe0743855303763ddf6668"
+ "sha256": "4d68dbb16f18feaf922226aef95593f86b22f50485fb97093a43641caec91f35"
},
"previews/piper/reference_replay.third-party.txt": {
"bytes": 1199,
@@ -126,8 +126,8 @@
"sha256": "494d7ce4ae7df05e51e0e9f7ed5c2dcb0a489817d72f8c83944e1d50e078396e"
},
"previews/so101/index.html": {
- "bytes": 9636,
- "sha256": "41a97a426a2b72520e39a469eaa78f4b281c7fef9cf9607301af9c398594251f"
+ "bytes": 9676,
+ "sha256": "e8bb5c8fc00e454e3abaa4ffbdfbc48ab1c1e91114030028247d14cba92620a0"
},
"previews/so101/licenses/robotstudio_so101/LICENSE": {
"bytes": 11357,
@@ -151,7 +151,7 @@
},
"previews/so101/preview-integrity.json": {
"bytes": 2375,
- "sha256": "fc1ec111e1b44b8102e9a2967f7c659ea33d0654d94b0742608159a8a3edc235"
+ "sha256": "533950bd420960b09d2d708fb8874fe206190a1e86c82129e559fcb224d92c1c"
},
"previews/so101/reference_replay.third-party.txt": {
"bytes": 1199,
@@ -194,8 +194,8 @@
"sha256": "cc162219dac340db4c663b34b3e54b0e5574679a7f737fa90853129d8a68978f"
},
"previews/ur5e/index.html": {
- "bytes": 9636,
- "sha256": "41a97a426a2b72520e39a469eaa78f4b281c7fef9cf9607301af9c398594251f"
+ "bytes": 9676,
+ "sha256": "e8bb5c8fc00e454e3abaa4ffbdfbc48ab1c1e91114030028247d14cba92620a0"
},
"previews/ur5e/licenses/robotiq_2f85/LICENSE": {
"bytes": 1297,
@@ -227,7 +227,7 @@
},
"previews/ur5e/preview-integrity.json": {
"bytes": 2685,
- "sha256": "9d4f72a84abda97a3748fdde396a71a07ffb5d30560454e29321bb61257fdfd9"
+ "sha256": "6507983a1ab46c4fc7d08a10c3618808c1feee21093c4580125f9ba02b202e9c"
},
"previews/ur5e/reference_replay.third-party.txt": {
"bytes": 1199,
@@ -251,7 +251,7 @@
},
"provenance.json": {
"bytes": 462,
- "sha256": "95b0c5c6302527658938dcc66f470ca1ac44c534b3cb2d7b1f94a96ab886594c"
+ "sha256": "df21d8aaaf4725f02f9e4ee11ebd8b0f7530b5d76d0896d831291c05c48d5851"
}
}
}
diff --git a/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/previews/piper/index.html b/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/previews/piper/index.html
index 843d8af..a4caf14 100644
--- a/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/previews/piper/index.html
+++ b/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/previews/piper/index.html
@@ -1,4 +1,4 @@
DVIDIA · Asset library DVIDIA RESEARCHResearch footprints ↗ A scene you can take with you.
Loading a reusable arm and shoe scene from the local asset library.
Drag to orbit · Scroll or pinch to zoom · Arrow keys orbit · + / − zoom · H homeOpening local geometry and recorded body poses…2.00 / 2.00 sBrowser physics: none. Native body poses were recorded locally. Geometry you can inspect. Assumptions you can see.
The arm uses its pinned upstream meshes and joint mechanism. Each shoe combines original boxes and ellipsoids for a lightweight collision model. The left and right instances have persistent identities.
Shoe mass, friction, size and shape are authored assumptions. These rigid proxies do not reproduce real shoe deformation, laces, fabric, insole compression or calibrated grasp forces.
What this preview measures.
A two-second native settling run under gravity, with arm commands held at initialization. Use the timeline to inspect initial and settled states, and collision mode to see the surfaces used by the engine.
It does not demonstrate a learned shoe-organizing policy, grasp success, physical transfer or a real-time guarantee. The browser replays Float32 body poses; the simulation used native precision.
+[hidden]{display:none!important}DVIDIA RESEARCHResearch footprints ↗ A scene you can take with you.
Loading a reusable arm and shoe scene from the local asset library.
Drag to orbit · Scroll or pinch to zoom · Arrow keys orbit · + / − zoom · H homeOpening local geometry and recorded body poses…2.00 / 2.00 sBrowser physics: none. Native body poses were recorded locally. Geometry you can inspect. Assumptions you can see.
The arm uses its pinned upstream meshes and joint mechanism. Each shoe combines original boxes and ellipsoids for a lightweight collision model. The left and right instances have persistent identities.
Shoe mass, friction, size and shape are authored assumptions. These rigid proxies do not reproduce real shoe deformation, laces, fabric, insole compression or calibrated grasp forces.
What this preview measures.
A two-second native settling run under gravity, with arm commands held at initialization. Use the timeline to inspect initial and settled states, and collision mode to see the surfaces used by the engine.
It does not demonstrate a learned shoe-organizing policy, grasp success, physical transfer or a real-time guarantee. The browser replays Float32 body poses; the simulation used native precision.
diff --git a/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/previews/piper/preview-integrity.json b/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/previews/piper/preview-integrity.json
index b7e5d7b..9a1ff38 100644
--- a/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/previews/piper/preview-integrity.json
+++ b/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/previews/piper/preview-integrity.json
@@ -27,8 +27,8 @@
"sha256": "7dac62907bf79874eceb0246e5c5ca25308549363cdad866daa7f0a97f480d18"
},
"index.html": {
- "bytes": 9636,
- "sha256": "41a97a426a2b72520e39a469eaa78f4b281c7fef9cf9607301af9c398594251f"
+ "bytes": 9676,
+ "sha256": "e8bb5c8fc00e454e3abaa4ffbdfbc48ab1c1e91114030028247d14cba92620a0"
},
"asset_library.bundle.js": {
"bytes": 529697,
diff --git a/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/previews/so101/index.html b/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/previews/so101/index.html
index 843d8af..a4caf14 100644
--- a/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/previews/so101/index.html
+++ b/docs/downloads/topics/physical-grounding/runs/2026-10-12-shoe-asset-library-v01/library/previews/so101/index.html
@@ -1,4 +1,4 @@
DVIDIA · Asset library DVIDIA RESEARCHResearch footprints ↗ A scene you can take with you.
Loading a reusable arm and shoe scene from the local asset library.
Drag to orbit · Scroll or pinch to zoom · Arrow keys orbit · + / − zoom · H homeOpening local geometry and recorded body poses…2.00 / 2.00 sBrowser physics: none. Native body poses were recorded locally. Geometry you can inspect. Assumptions you can see.
The arm uses its pinned upstream meshes and joint mechanism. Each shoe combines original boxes and ellipsoids for a lightweight collision model. The left and right instances have persistent identities.
Shoe mass, friction, size and shape are authored assumptions. These rigid proxies do not reproduce real shoe deformation, laces, fabric, insole compression or calibrated grasp forces.
What this preview measures.
A two-second native settling run under gravity, with arm commands held at initialization. Use the timeline to inspect initial and settled states, and collision mode to see the surfaces used by the engine.
It does not demonstrate a learned shoe-organizing policy, grasp success, physical transfer or a real-time guarantee. The browser replays Float32 body poses; the simulation used native precision.
+[hidden]{display:none!important}DVIDIA RESEARCHResearch footprints ↗ A scene you can take with you.
Loading a reusable arm and shoe scene from the local asset library.
Drag to orbit · Scroll or pinch to zoom · Arrow keys orbit · + / − zoom · H homeOpening local geometry and recorded body poses…2.00 / 2.00 sBrowser physics: none. Native body poses were recorded locally. Geometry you can inspect. Assumptions you can see.
The arm uses its pinned upstream meshes and joint mechanism. Each shoe combines original boxes and ellipsoids for a lightweight collision model. The left and right instances have persistent identities.
Shoe mass, friction, size and shape are authored assumptions. These rigid proxies do not reproduce real shoe deformation, laces, fabric, insole compression or calibrated grasp forces.
What this preview measures.
A two-second native settling run under gravity, with arm commands held at initialization. Use the timeline to inspect initial and settled states, and collision mode to see the surfaces used by the engine.
It does not demonstrate a learned shoe-organizing policy, grasp success, physical transfer or a real-time guarantee. The browser replays Float32 body poses; the simulation used native precision.