Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 10 additions & 0 deletions docs/downloads/reports/Skillspace footage training pipeline.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,16 @@

The public, standalone [DVIDIA training repository](https://github.com/Dvidia-Inference/dvidia-training) provides a local CPU pipeline that turns actual Skillspace media into a fitted visual model and, when aligned numeric actions are supplied, a learned movement head. The useful advance is a reproducible path from source files through provenance checks, grouped splits, measured fitting and portable artifacts. **Video prediction and robot movement remain separate claims.** Ten synthetic six-second clips demonstrate visual fitting; simulator telemetry supplies the movement pilot's action labels. Its exported capsule installs and passes native qualification in one exact scene, while identical joint targets with open jaws fail. Neither experiment learns shoe organization or reconstructs robot actions from human footage. This report documents the implemented path and its limited measured evidence, rather than establishing a sufficient demonstration count or hardware competence. ([Standalone source archive](https://research.dvidia.org/downloads/topics/physical-grounding/runs/2026-10-08-footage-pipeline-v1/release/dvidia-training-v0.1.zip), [Qualification receipt](https://research.dvidia.org/downloads/topics/physical-grounding/runs/2026-10-08-footage-pipeline-v1/qualification/qualification.json))

## Published artifacts on Hugging Face

The [DVIDIA organization on Hugging Face](https://huggingface.co/Dvidia) mirrors the measured pilot for inspection and local reproduction:

- [Example dataset](https://huggingface.co/datasets/Dvidia/dvidia-training-examples): explicitly synthetic moving-disc footage and native simulator recordings, with frozen prepared inputs and grouped train/development/test splits. No human footage is included.
- [Pilot model artifacts](https://huggingface.co/Dvidia/dvidia-training-pilot): visual weights, the learned joint-delta candidate, evaluation reports and native qualification/control evidence. The movement candidate learns from aligned numerical simulator labels and remains bound to its declared adapter.
- [Training evidence Space](https://huggingface.co/spaces/Dvidia/dvidia-training): a static overview of the measurements and the saved contact/lift/place replay from one exact held-out scene.

Training runs on the user's own computer or server through the [local installation guide](https://github.com/Dvidia-Inference/dvidia-training/blob/main/docs/installation.md). The Space presents fixed evidence; it provides no hosted training or uploads. These published artifacts remain research-alpha, simulation-only evidence, with no qualified hardware skill or established human-video action bridge.

## Actual media and connected source groups gate training

Start with the repository's [installation guide](https://github.com/Dvidia-Inference/dvidia-training/blob/main/docs/installation.md) and [hardware guide](https://github.com/Dvidia-Inference/dvidia-training/blob/main/docs/hardware.md). The pipeline accepts a local folder, ZIP or prepared dataset; explicitly enabled online mode also accepts a supported public DVIDIA manifest and its declared media. Visual training requires Python 3.12+, NumPy and FFmpeg/ffprobe; arm simulation adds MuJoCo. A Skillspace description with no accessible video does not become a dataset. Images alone are rejected. An audit-only preparation step checks media, copies the exact bytes, records hashes and freezes train, development and test assignments before fitting. Fresh output directories prevent silent replacement of earlier runs. The command-line interface and a loopback training studio expose this same path. ([Implementation](https://github.com/Dvidia-Inference/dvidia-training))
Expand Down
2 changes: 1 addition & 1 deletion docs/index.html

Large diffs are not rendered by default.

4 changes: 2 additions & 2 deletions docs/papers/clef-and-grounded-video-observations/index.html

Large diffs are not rendered by default.

Large diffs are not rendered by default.

4 changes: 2 additions & 2 deletions docs/papers/from-skillspace-to-robot-skill/index.html

Large diffs are not rendered by default.

4 changes: 2 additions & 2 deletions docs/papers/grok-robotics-research-on-x/index.html

Large diffs are not rendered by default.

4 changes: 2 additions & 2 deletions docs/papers/offline-cable-training-environment/index.html

Large diffs are not rendered by default.

4 changes: 2 additions & 2 deletions docs/papers/physical-skill-research-agenda/index.html

Large diffs are not rendered by default.

4 changes: 2 additions & 2 deletions docs/papers/skill-capsule-acquisition/index.html

Large diffs are not rendered by default.

4 changes: 2 additions & 2 deletions docs/papers/skillspace-arm-benchmarks/index.html

Large diffs are not rendered by default.

4 changes: 2 additions & 2 deletions docs/papers/skillspace-arm-installation/index.html

Large diffs are not rendered by default.

12 changes: 10 additions & 2 deletions docs/papers/skillspace-footage-training/index.html

Large diffs are not rendered by default.

Large diffs are not rendered by default.

4 changes: 2 additions & 2 deletions docs/topics/capture-quality/index.html
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
<!doctype html>
<html lang="en">
<head><meta charset="utf-8"><meta name="viewport" content="width=device-width, initial-scale=1"><title>Capture quality · DVIDIA Research</title><meta name="description" content="Test whether hands, relevant objects and state changes are actually observable in a POV demonstration."><meta name="theme-color" content="#faf9f6"><link rel="canonical" href="https://research.dvidia.org/topics/capture-quality/"><link rel="alternate" type="application/rss+xml" title="DVIDIA Research" href="../../feed.xml"><link rel="icon" href="../../assets/favicon.svg" type="image/svg+xml"><link rel="stylesheet" href="../../styles.css"><script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","headline":"Capture quality","description":"Test whether hands, relevant objects and state changes are actually observable in a POV demonstration.","url":"https://research.dvidia.org/topics/capture-quality/","publisher":{"@type":"Organization","name":"DVIDIA","url":"https://dvidia.org"}}</script><script src="../../app.js" defer></script></head>
<body><a class="skip-link" href="#main">Skip to content</a><aside class="site-sidebar" aria-label="Research navigation"><a class="brand" href="../../index.html"><span class="brand-name">DVIDIA</span><span class="brand-subtitle">Research notebook</span></a><nav class="site-nav" id="site-navigation" aria-label="Main"><div class="nav-section"><span class="nav-label">Notebook</span><a href="../../index.html">Research journal</a><a href="../../index.html#progress">Progress</a></div><div class="nav-section"><span class="nav-label">Publications</span><a href="../../papers/skillspace-footage-training/index.html">Skillspace footage training pipeline</a><a href="../../papers/skill-capsule-acquisition/index.html">Skill capsule acquisition and optimization</a><a href="../../papers/skillspace-arm-benchmarks/index.html">Skillspace arm benchmarks and evolution</a><a href="../../papers/skillspace-arm-installation/index.html">Installing a Skillspace on a simulated arm</a><a href="../../papers/offline-cable-training-environment/index.html">Offline cable training environment</a><a href="../../papers/fast-accurate-robot-training-simulation/index.html">Fast accurate robot training simulation</a><a href="../../papers/grok-robotics-research-on-x/index.html">Grok robotics research on X</a><a href="../../papers/spatially-grounded-downloadable-robot-skills/index.html">Spatially grounded downloadable robot skills</a><a href="../../papers/clef-and-grounded-video-observations/index.html">Clef and grounded video observations</a><a href="../../papers/from-skillspace-to-robot-skill/index.html">From Skillspace to robot skill</a><a href="../../papers/physical-skill-research-agenda/index.html">Turn physical footage into testable evidence</a></div><div class="nav-section"><span class="nav-label">Questions</span><a href="index.html" aria-current="page">Capture quality</a><a href="../skill-relations/index.html">Skill relations</a><a href="../evaluation-provenance/index.html">Evaluation and provenance</a><a href="../physical-grounding/index.html">Physical grounding</a></div><div class="nav-section"><span class="nav-label">Follow the work</span><a href="../../feed.xml">RSS feed</a><a href="https://github.com/Dvidia-Inference/dvidia-research">GitHub</a><a href="https://x.com/IAMMRRIVR">X · @IAMMRRIVR</a><a href="mailto:hello@dvidia.org">hello@dvidia.org</a></div></nav><details class="mobile-navigation"><summary>Contents</summary><nav aria-label="Mobile"><div class="nav-section"><span class="nav-label">Notebook</span><a href="../../index.html">Research journal</a><a href="../../index.html#progress">Progress</a></div><div class="nav-section"><span class="nav-label">Publications</span><a href="../../papers/skillspace-footage-training/index.html">Skillspace footage training pipeline</a><a href="../../papers/skill-capsule-acquisition/index.html">Skill capsule acquisition and optimization</a><a href="../../papers/skillspace-arm-benchmarks/index.html">Skillspace arm benchmarks and evolution</a><a href="../../papers/skillspace-arm-installation/index.html">Installing a Skillspace on a simulated arm</a><a href="../../papers/offline-cable-training-environment/index.html">Offline cable training environment</a><a href="../../papers/fast-accurate-robot-training-simulation/index.html">Fast accurate robot training simulation</a><a href="../../papers/grok-robotics-research-on-x/index.html">Grok robotics research on X</a><a href="../../papers/spatially-grounded-downloadable-robot-skills/index.html">Spatially grounded downloadable robot skills</a><a href="../../papers/clef-and-grounded-video-observations/index.html">Clef and grounded video observations</a><a href="../../papers/from-skillspace-to-robot-skill/index.html">From Skillspace to robot skill</a><a href="../../papers/physical-skill-research-agenda/index.html">Turn physical footage into testable evidence</a></div><div class="nav-section"><span class="nav-label">Questions</span><a href="index.html" aria-current="page">Capture quality</a><a href="../skill-relations/index.html">Skill relations</a><a href="../evaluation-provenance/index.html">Evaluation and provenance</a><a href="../physical-grounding/index.html">Physical grounding</a></div><div class="nav-section"><span class="nav-label">Follow the work</span><a href="../../feed.xml">RSS feed</a><a href="https://github.com/Dvidia-Inference/dvidia-research">GitHub</a><a href="https://x.com/IAMMRRIVR">X · @IAMMRRIVR</a><a href="mailto:hello@dvidia.org">hello@dvidia.org</a></div></nav></details></aside><main class="page-shell" id="main"><article class="paper topic-paper"><header class="paper-header"><p class="eyebrow">Research question</p><h1>Capture quality</h1><p class="paper-meta">DVIDIA · Visible evidence</p><p class="paper-status">Protocols and progress are scoped below</p></header><section class="paper-abstract"><h2>Question</h2><p>Test whether hands, relevant objects and state changes are actually observable in a POV demonstration.</p></section><section class="related-papers"><h2>Publications in this area</h2><ul><li><a href="../../papers/skillspace-footage-training/index.html">Skillspace footage training pipeline</a></li><li><a href="../../papers/clef-and-grounded-video-observations/index.html">Clef and grounded video observations</a></li><li><a href="../../papers/physical-skill-research-agenda/index.html">Turn physical footage into testable evidence</a></li></ul></section><div class="paper-body"><p><strong>Question:</strong> Test whether hands, relevant objects and state changes are actually observable in a POV demonstration.</p>
<body><a class="skip-link" href="#main">Skip to content</a><aside class="site-sidebar" aria-label="Research navigation"><a class="brand" href="../../index.html"><span class="brand-name">DVIDIA</span><span class="brand-subtitle">Research notebook</span></a><nav class="site-nav" id="site-navigation" aria-label="Main"><div class="nav-section"><span class="nav-label">Notebook</span><a href="../../index.html">Research journal</a><a href="../../index.html#progress">Progress</a></div><div class="nav-section"><span class="nav-label">Publications</span><a href="../../papers/skillspace-footage-training/index.html">Skillspace footage training pipeline</a><a href="../../papers/skill-capsule-acquisition/index.html">Skill capsule acquisition and optimization</a><a href="../../papers/skillspace-arm-benchmarks/index.html">Skillspace arm benchmarks and evolution</a><a href="../../papers/skillspace-arm-installation/index.html">Installing a Skillspace on a simulated arm</a><a href="../../papers/offline-cable-training-environment/index.html">Offline cable training environment</a><a href="../../papers/fast-accurate-robot-training-simulation/index.html">Fast accurate robot training simulation</a><a href="../../papers/grok-robotics-research-on-x/index.html">Grok robotics research on X</a><a href="../../papers/spatially-grounded-downloadable-robot-skills/index.html">Spatially grounded downloadable robot skills</a><a href="../../papers/clef-and-grounded-video-observations/index.html">Clef and grounded video observations</a><a href="../../papers/from-skillspace-to-robot-skill/index.html">From Skillspace to robot skill</a><a href="../../papers/physical-skill-research-agenda/index.html">Turn physical footage into testable evidence</a></div><div class="nav-section"><span class="nav-label">Questions</span><a href="index.html" aria-current="page">Capture quality</a><a href="../skill-relations/index.html">Skill relations</a><a href="../evaluation-provenance/index.html">Evaluation and provenance</a><a href="../physical-grounding/index.html">Physical grounding</a></div><div class="nav-section"><span class="nav-label">Follow the work</span><a href="../../feed.xml">RSS feed</a><a href="https://github.com/Dvidia-Inference/dvidia-research">GitHub</a><a href="https://huggingface.co/Dvidia">Hugging Face</a><a href="https://x.com/IAMMRRIVR">X · @IAMMRRIVR</a><a href="mailto:hello@dvidia.org">hello@dvidia.org</a></div></nav><details class="mobile-navigation"><summary>Contents</summary><nav aria-label="Mobile"><div class="nav-section"><span class="nav-label">Notebook</span><a href="../../index.html">Research journal</a><a href="../../index.html#progress">Progress</a></div><div class="nav-section"><span class="nav-label">Publications</span><a href="../../papers/skillspace-footage-training/index.html">Skillspace footage training pipeline</a><a href="../../papers/skill-capsule-acquisition/index.html">Skill capsule acquisition and optimization</a><a href="../../papers/skillspace-arm-benchmarks/index.html">Skillspace arm benchmarks and evolution</a><a href="../../papers/skillspace-arm-installation/index.html">Installing a Skillspace on a simulated arm</a><a href="../../papers/offline-cable-training-environment/index.html">Offline cable training environment</a><a href="../../papers/fast-accurate-robot-training-simulation/index.html">Fast accurate robot training simulation</a><a href="../../papers/grok-robotics-research-on-x/index.html">Grok robotics research on X</a><a href="../../papers/spatially-grounded-downloadable-robot-skills/index.html">Spatially grounded downloadable robot skills</a><a href="../../papers/clef-and-grounded-video-observations/index.html">Clef and grounded video observations</a><a href="../../papers/from-skillspace-to-robot-skill/index.html">From Skillspace to robot skill</a><a href="../../papers/physical-skill-research-agenda/index.html">Turn physical footage into testable evidence</a></div><div class="nav-section"><span class="nav-label">Questions</span><a href="index.html" aria-current="page">Capture quality</a><a href="../skill-relations/index.html">Skill relations</a><a href="../evaluation-provenance/index.html">Evaluation and provenance</a><a href="../physical-grounding/index.html">Physical grounding</a></div><div class="nav-section"><span class="nav-label">Follow the work</span><a href="../../feed.xml">RSS feed</a><a href="https://github.com/Dvidia-Inference/dvidia-research">GitHub</a><a href="https://huggingface.co/Dvidia">Hugging Face</a><a href="https://x.com/IAMMRRIVR">X · @IAMMRRIVR</a><a href="mailto:hello@dvidia.org">hello@dvidia.org</a></div></nav></details></aside><main class="page-shell" id="main"><article class="paper topic-paper"><header class="paper-header"><p class="eyebrow">Research question</p><h1>Capture quality</h1><p class="paper-meta">DVIDIA · Visible evidence</p><p class="paper-status">Protocols and progress are scoped below</p></header><section class="paper-abstract"><h2>Question</h2><p>Test whether hands, relevant objects and state changes are actually observable in a POV demonstration.</p></section><section class="related-papers"><h2>Publications in this area</h2><ul><li><a href="../../papers/skillspace-footage-training/index.html">Skillspace footage training pipeline</a></li><li><a href="../../papers/clef-and-grounded-video-observations/index.html">Clef and grounded video observations</a></li><li><a href="../../papers/physical-skill-research-agenda/index.html">Turn physical footage into testable evidence</a></li></ul></section><div class="paper-body"><p><strong>Question:</strong> Test whether hands, relevant objects and state changes are actually observable in a POV demonstration.</p>
<p>Status: <strong>scoping</strong> · Updated 6 October 2026 · Results: <strong>not run</strong></p>
<p><a href="https://github.com/Dvidia-Inference/dvidia-research/issues/new?template=research-question.yml">Open a research question</a> · <a href="../../downloads/topics/capture-quality/experiments/0001-baseline.md">Baseline protocol</a> · <a href="../../downloads/topics/capture-quality/references.bib">Bibliography</a></p>
<h2 id="primary-sources">Primary sources</h2>
Expand Down Expand Up @@ -36,4 +36,4 @@ <h2 id="milestones">Milestones</h2>
</ul>
<h2 id="results">Results</h2>
<p><strong>Not run.</strong> Source review is complete; no experiment, validation claim or independent reproduction is complete. Add measured results and their limitations only after execution.</p>
</div></article><footer class="site-footer"><p>Published by DVIDIA · <a href="mailto:hello@dvidia.org">hello@dvidia.org</a></p><p><a href="https://github.com/Dvidia-Inference/dvidia-research">Source and revisions</a> · <a href="../../feed.xml">RSS</a> · <a href="https://x.com/IAMMRRIVR">@IAMMRRIVR</a></p><p>Writing: <a href="https://github.com/Dvidia-Inference/dvidia-research/blob/main/LICENSE">CC BY 4.0</a>. Original code: <a href="https://github.com/Dvidia-Inference/dvidia-research/blob/main/LICENSE-CODE">MIT</a>. External work retains its own terms.</p></footer></main></body></html>
</div></article><footer class="site-footer"><p>Published by DVIDIA · <a href="mailto:hello@dvidia.org">hello@dvidia.org</a></p><p><a href="https://github.com/Dvidia-Inference/dvidia-research">Source and revisions</a> · <a href="https://huggingface.co/Dvidia">Hugging Face</a> · <a href="../../feed.xml">RSS</a> · <a href="https://x.com/IAMMRRIVR">@IAMMRRIVR</a></p><p>Writing: <a href="https://github.com/Dvidia-Inference/dvidia-research/blob/main/LICENSE">CC BY 4.0</a>. Original code: <a href="https://github.com/Dvidia-Inference/dvidia-research/blob/main/LICENSE-CODE">MIT</a>. External work retains its own terms.</p></footer></main></body></html>
Loading
Loading