Publish Ropedia Xperience-10M derived artifacts

ARTIFACT_GUIDE.md

@@ -156,7 +156,7 @@ The project separates these reading layers:
 | Artifact | Current status |
 | --- | --- |
 | [`results/omni_finetune/DATA_ACCESS_STATUS.md`](results/omni_finetune/DATA_ACCESS_STATUS.md) | Summarizes the staging requirement before the held-out Qwen3-Omni pilot can report metrics. |
-| [`results/omni_finetune/MULTI_EPISODE_ACCESS_STATUS.md`](results/omni_finetune/MULTI_EPISODE_ACCESS_STATUS.md) | Documents the public multi-episode access path, selected relay plan, and data requirements. |
+| [`results/omni_finetune/MULTI_EPISODE_ACCESS_STATUS.md`](results/omni_finetune/MULTI_EPISODE_ACCESS_STATUS.md) | Documents the public multi-episode access path, selected 128-episode pilot plan, and data requirements. |
 | [`scripts/omni/discover_xperience10m_sources.py`](scripts/omni/discover_xperience10m_sources.py) | Discovery gate for valid multi-episode Xperience-10M sources. |
 | [`scripts/omni/train_qwen3_omni_lora.py`](scripts/omni/train_qwen3_omni_lora.py) | Training entrypoint for the Qwen3-Omni LoRA pilot after the data gate passes. |
 | [`FOUNDATION_MODEL_PLAN.md`](FOUNDATION_MODEL_PLAN.md) | Adds the post-data-gate backbone selection plan: Qwen3-Omni first, Cosmos 3 for world modeling, and OpenVLA/openpi/GR00T for policy/action branches. |

EVIDENCE_CONTRACT.md

@@ -7,7 +7,7 @@ the dashboard as a basis for further work.
 | Project statement | Current evidence | Status | Current scope |
 | --- | --- | --- | --- |
 | A first-pass reader has a compact current-state summary. | `PROJECT_STATUS.md`, `docs/data/project_status.json` | Verified guide | Summarizes existing evidence and current limitations |
-| The staged research path is explicit. | `RESEARCH_ROADMAP.md`, `docs/data/research_roadmap.json` | Current roadmap | Connects public-sample task development to multi-episode staging, Qwen3-Omni LoRA, robustness runs, and larger omni-model extensions |
+| The research roadmap is explicit. | `RESEARCH_ROADMAP.md`, `docs/data/research_roadmap.json` | Current roadmap | Connects public-sample task development to multi-episode data preparation, Qwen3-Omni LoRA, robustness runs, and larger omni-model extensions |
 | The public dataset description is aligned with the official gated Xperience-10M dataset card and public sample card. | `XPERIENCE10M_DATASET_CARD_ALIGNMENT.md`, `docs/data/xperience10m_dataset_card_alignment.json` | Verified description alignment | Summarizes upstream public metadata, API listing facts, sample license/tooling, and card facts; does not grant access or mirror raw data |
 | Source facts, sample details, API-listing notes, and project coverage are aligned across repo, website, and HF cards. | `SOURCE_ALIGNMENT_AUDIT.md`, `docs/data/source_alignment_audit.json`, `scripts/validate_source_alignment.py` | Source alignment recorded | Offline committed-fact report; does not fetch private gated data |
 | Public figures are indexed as project evidence. | `FIGURE_INDEX.md`, `docs/data/figure_index.json`, `scripts/build_figure_index.py` | Verified visual evidence | Derived figures and thumbnails only; does not include raw MP4/HDF5/RRD data |
@@ -21,8 +21,8 @@ the dashboard as a basis for further work.
 | Minimal and neural heads use the same task contracts. | `scripts/neural_task_models.py`, `results/episode_task_suite/neural_mlp/`, `docs/assets/task_architectures.png` | Verified for 12 minimal heads and 12 neural MLP heads | Small heads only; not a foundation model |
 | Four Ropedia research directions are mapped honestly as direct, proxy, or diagnostic evidence. | `results/episode_task_suite/research_directions/research_direction_taxonomy.json`, `docs/data/research_directions.json` | Verified taxonomy | Some directions remain proxy-only |
 | Four extra direction probes are coded and evaluated. | `results/episode_task_suite/research_direction_extensions/research_direction_extension_results.json`, `docs/data/research_direction_extensions.json` | Verified single-episode probes | Not full human modeling, neural rendering, intent modeling, or world modeling solutions |
-| Qwen3-Omni infrastructure has passed setup checks. | `results/omni_finetune/RUN_REPORT.md`, `results/omni_finetune/dataset_manifest.json`, `results/omni_finetune/metrics_eval.json` | Setup-stage evidence | One episode, 128 train windows; full metrics require completed multi-episode staging and held-out evaluation |
-| The Qwen3-Omni LoRA pilot is in multi-episode staging. | `results/omni_finetune/DATA_ACCESS_STATUS.md`, `results/omni_finetune/MULTI_EPISODE_ACCESS_STATUS.md`, `results/omni_finetune/source_discovery.json` | Data staging | Full-dataset access is granted; held-out metrics come after selected relay, manifest construction, training, and test evaluation |
+| Qwen3-Omni infrastructure has passed setup checks. | `results/omni_finetune/RUN_REPORT.md`, `results/omni_finetune/dataset_manifest.json`, `results/omni_finetune/metrics_eval.json` | Setup-stage evidence | One episode, 128 train windows; full metrics require completed multi-episode data preparation and held-out evaluation |
+| The Qwen3-Omni LoRA pilot is in selected multi-episode preparation. | `results/omni_finetune/DATA_ACCESS_STATUS.md`, `results/omni_finetune/MULTI_EPISODE_ACCESS_STATUS.md`, `results/omni_finetune/source_discovery.json` | Data preparation | The gated Xperience-10M dataset is available; held-out metrics come after manifest construction, training, and test evaluation |
 | Older pilot path strings are tracked as setup-file provenance. | `scripts/validate_scope_claims.py`, `docs/data/scope_claims_audit.json` | Multi-episode pilot status | Run/path identifiers stay separate from completed held-out-episode results |
 | Prepared GitHub/Hugging Face mirrors carry matching critical files. | `scripts/validate_mirror_parity.py`, `docs/data/mirror_parity.json` | Mirror parity report | Compares prepared data files, visual assets, website HTML, and validator scripts before upload; live URLs are checked after publishing |
 | The public GitHub and Hugging Face bundles are ready to share. | `scripts/validate_publication_package.py`, `docs/data/publication_audit.json` | Public bundle contents | Covers public files, HF bundles, and current public-card assets; temporary local outputs are excluded |

PROJECT_BRIEF.md

@@ -32,7 +32,7 @@ results, and see what remains before multi-episode model-quality claims.
 | Task suite | 12 embodied-AI task contracts with inputs, targets, metrics, predictions, and case-study walkthroughs |
 | Models | Minimal linear/ridge/logistic baselines plus compact PyTorch MLP heads for the same 12 tasks |
 | Research map | Four Ropedia research directions with direct, proxy, diagnostic, and extension-task coverage |
-| Scale-up path | Qwen3-Omni LoRA code path prepared; full-dataset access is granted and a 128-episode selected relay is being staged |
+| Scale-up path | Qwen3-Omni LoRA code path prepared; the gated Xperience-10M dataset is available for a selected 128-episode pilot |
 
 ## How To Read It
 

PROJECT_README.md

@@ -71,7 +71,7 @@ before the multi-episode omni-model stage becomes a real held-out evaluation.
 | Task suite | 12 human-readable embodied-AI task contracts with input, process, output, metrics, predictions, and case-study walkthroughs |
 | Baselines | Minimal linear/ridge/logistic heads plus compact PyTorch MLP task heads over the same chronological split |
 | Research directions | Task mapping and extension probes for human modeling, 3D/4D reconstruction, egocentric interaction, and world modeling |
-| Scale-up path | Full-dataset access granted; a selected 128-episode pilot is being prepared before Qwen3-Omni LoRA, followed by Cosmos 3/world-model and VLA/policy branches |
+| Scale-up path | The gated Xperience-10M dataset is available for a selected 128-episode pilot before Qwen3-Omni LoRA, followed by Cosmos 3/world-model and VLA/policy branches |
 | Public surfaces | GitHub repo, GitHub Pages dashboard, HF Space, HF artifact dataset, HF baseline-model repo, and HF collection |
 
 For the fastest interpretation of the current metrics, start with
@@ -204,7 +204,7 @@ They give the current research state in one compact table:
 | Source alignment | Source facts, sample details, API-listing notes, and project coverage are consistent across repo, website, and HF cards |
 | Evaluation protocol | Verified generated protocol for windowing, split policy, leakage controls, and per-task metrics |
 | Website and HF mirrors | Verified by website reference reports, public presentation reports, mirror parity, and live-publication checks; the public dashboard uses six top-level tabs, including an explicit Directions tab, plus subsection tabs for dataset, task-suite, method, result, direction, and resource views |
-| Qwen3-Omni multi-episode pilot | Full-dataset access granted; selected 128-episode preparation is underway, with full metrics pending completed preprocessing, training, and held-out evaluation |
+| Qwen3-Omni multi-episode pilot | The gated Xperience-10M dataset is available for selected 128-episode preparation, with full metrics pending completed preprocessing, training, and held-out evaluation |
 | Raw Xperience-10M data / full Qwen weights | Not redistributed |
 
 ## 90-Second Research Project Path
@@ -218,7 +218,7 @@ If you are reading the project cold, open these in order:
 | 3 | Are source facts consistently presented? | [`SOURCE_ALIGNMENT_AUDIT.md`](SOURCE_ALIGNMENT_AUDIT.md), [`docs/data/source_alignment_audit.json`](docs/data/source_alignment_audit.json), [`scripts/validate_source_alignment.py`](scripts/validate_source_alignment.py) | Repo, website, and HF cards use the same full-dataset facts, sample-card facts, API-listing notes, and project coverage. |
 | 4 | How exactly are tasks evaluated? | [`EVALUATION_PROTOCOL.md`](EVALUATION_PROTOCOL.md), [`docs/data/evaluation_protocol.json`](docs/data/evaluation_protocol.json), [`scripts/build_evaluation_protocol.py`](scripts/build_evaluation_protocol.py) | The window unit, chronological split, leakage controls, task metrics, and current limitations are explicit. |
 | 5 | What do the current results mean? | [`RESEARCH_TAKEAWAYS.md`](RESEARCH_TAKEAWAYS.md), [`docs/data/research_takeaways.json`](docs/data/research_takeaways.json), [`docs/data/summary_metrics.json`](docs/data/summary_metrics.json) | The takeaways are generated from committed metrics and identify which signals are ready for larger held-out experiments. |
-| 6 | What is the staged roadmap? | [`RESEARCH_ROADMAP.md`](RESEARCH_ROADMAP.md), [`docs/data/research_roadmap.json`](docs/data/research_roadmap.json), [`DATA_ACCESS_STATUS.md`](results/omni_finetune/DATA_ACCESS_STATUS.md) | The roadmap connects public-sample task development to multi-episode staging, Qwen3-Omni LoRA, foundation-model selection, robustness runs, and larger omni/world-model extensions. |
+| 6 | What is the research roadmap? | [`RESEARCH_ROADMAP.md`](RESEARCH_ROADMAP.md), [`docs/data/research_roadmap.json`](docs/data/research_roadmap.json), [`DATA_ACCESS_STATUS.md`](results/omni_finetune/DATA_ACCESS_STATUS.md) | The roadmap connects public-sample task development to multi-episode data preparation, Qwen3-Omni LoRA, foundation-model selection, robustness runs, and larger omni/world-model extensions. |
 | 7 | Which foundation model comes next? | [`FOUNDATION_MODEL_PLAN.md`](FOUNDATION_MODEL_PLAN.md), [`docs/data/foundation_model_plan.json`](docs/data/foundation_model_plan.json) | Qwen3-Omni remains the first held-out LoRA baseline; Cosmos 3 is the first world-model branch; OpenVLA/openpi/GR00T wait for explicit action targets. |
 | 8 | How do I reproduce it? | [`REPRODUCIBILITY.md`](REPRODUCIBILITY.md), [`docs/data/reproducibility_matrix.json`](docs/data/reproducibility_matrix.json), [`notes/reproducibility_audit.md`](notes/reproducibility_audit.md) | Public commands, expected outputs, and the latest exact-match reproduction record are explicit. |
 | 9 | What is one model input? | [`windows.csv`](results/episode_task_suite/windows.csv), [`feature_manifest.json`](results/episode_task_suite/feature_manifest.json), [`available_modalities.json`](results/episode_task_suite/available_modalities.json) | The input is an aligned 8,546-dimensional multimodal window with synchronized video, audio, sensor, and language signals. |
@@ -632,7 +632,7 @@ Current status in this repo:
 - selected_episode_plan: 128 metadata-balanced episodes, 96/16/16 train/val/test
 - selected_download_size: 277.71 GiB excluding `visualization.rrd`
 - ready_for_held_out_pilot: false until the selected episodes are fully staged and checked
-- full-dataset access: granted; selected multi-episode data preparation is in progress
+- gated dataset: available for selected multi-episode data preparation
 - source_discovery: `results/omni_finetune/source_discovery.json`
 - data_status: `results/omni_finetune/DATA_ACCESS_STATUS.md`
 - access_status: `results/omni_finetune/MULTI_EPISODE_ACCESS_STATUS.md`

RESEARCH_ROADMAP.md

@@ -10,7 +10,7 @@ should exist before the stage is treated as complete.
 | Stage | Status | Entry condition | Research deliverables | Completion evidence |
 | --- | --- | --- | --- | --- |
 | Public-Sample Task Lab | Implemented | One public Xperience-10M sample episode is available. | 1,161 aligned windows, 12 task contracts, minimal heads, neural MLP heads, modality atlas, task walkthroughs, and derived figures. | `PROJECT_STATUS.md`, `EVALUATION_PROTOCOL.md`, `RESEARCH_TAKEAWAYS.md`, `docs/data/summary_metrics.json`, `results/episode_task_suite/summary_report.json` |
-| Multi-Episode Data Preparation | Active | Full-dataset access and enough storage for selected episodes. | 128 selected episodes, episode manifest, missing-view manifest, held-out episode split, and source-discovery report. | `results/omni_finetune/DATA_ACCESS_STATUS.md`, `results/omni_finetune/MULTI_EPISODE_ACCESS_STATUS.md`, `results/omni_finetune/source_discovery.json` |
+| Multi-Episode Data Preparation | Active | Gated dataset availability and enough storage for selected episodes. | 128 selected episodes, episode manifest, missing-view manifest, held-out episode split, and source-discovery report. | `results/omni_finetune/DATA_ACCESS_STATUS.md`, `results/omni_finetune/MULTI_EPISODE_ACCESS_STATUS.md`, `results/omni_finetune/source_discovery.json` |
 | Qwen3-Omni LoRA Pilot | Next | Selected episodes staged locally with no train/test episode leakage. | Dataset JSONL/media manifests, LoRA adapter checkpoint, progress logs, held-out predictions, metrics, confusion matrices, and run report. | `dataset_manifest.json`, `training_metadata.json`, `progress.jsonl`, `metrics.json`, `predictions.jsonl`, `RUN_REPORT.md` |
 | Foundation-Model Selection Matrix | Next | The selected pilot episodes are prepared, or a 3-8 episode dry run is available for preprocessing checks. | Backbone registry, Cosmos 3 world-model branch plan, Qwen3-Omni baseline plan, OpenVLA/openpi/GR00T policy candidates, and model-specific evaluation additions. | `FOUNDATION_MODEL_PLAN.md`, `docs/data/foundation_model_plan.json`, `research_roadmap_interactive.json` |
 | 64-128 Episode Robustness Run | Planned | The selected-episode pilot trains and evaluates cleanly. | Split-by-session metrics, modality ablations, calibration/object/language error analysis, and sensitivity to missing views. | Held-out metrics by session, task, and modality; ablation tables; qualitative error analysis. |

XPERIENCE10M_DATASET_CARD_ALIGNMENT.md

@@ -222,4 +222,4 @@ When describing Xperience-10M in this repo, keep these limitations visible:
 | Episode layout uses six MP4 streams and `annotation.hdf5` | Used by sample inspection and pilot-readiness scripts |
 | Audio exists in MP4 streams | Represented in the current multimodal feature contract |
 | 4D reconstruction/world modeling are intended research directions | Represented by proxy/diagnostic tasks only |
-| Real model quality requires held-out multi-episode evaluation | Pending selected multi-episode staging, training, and held-out evaluation |
+| Real model quality requires held-out multi-episode evaluation | Pending selected multi-episode data preparation, training, and evaluation |

docs/data/artifact_index.json

@@ -1,6 +1,6 @@
 {
   "title": "Ropedia Xperience-10M Task Suite Artifact Index",
-  "generated_at_utc": "2026-06-04T07:39:12+00:00",
+  "generated_at_utc": "2026-06-04T16:27:07+00:00",
   "status": "pass",
   "artifact_count": 72,
   "missing": [],
@@ -40,8 +40,8 @@
       "surface": "repo_hf",
       "shows": "Gives first-pass readers a concise project shape before the detailed artifact trail.",
       "exists": true,
-      "bytes": 3596,
-      "sha256": "0a9557f57088644466b19e91202ec0304f9394e3abf69d7861a878d192d1c1aa"
+      "bytes": 3594,
+      "sha256": "8df692f2d6b092097d5ef7d5039e7529abe2fd0a80ff3fb23fa14260fb5f9de2"
     },
     {
       "id": "project_brief_json",
@@ -51,8 +51,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable first-reader project brief for the website and Hugging Face mirrors.",
       "exists": true,
-      "bytes": 3645,
-      "sha256": "66056460f2d93691d19e9af2f47acb57bab561f3b2697884549d144569b70fa7"
+      "bytes": 3637,
+      "sha256": "0e859cfd3fca3ac66c9dd67e04a404ab41757c248d8777ee9cf7da63ab11a2f7"
     },
     {
       "id": "project_status",
@@ -62,8 +62,8 @@
       "surface": "repo_hf",
       "shows": "Gives a compact current-state table for first-pass readers.",
       "exists": true,
-      "bytes": 7205,
-      "sha256": "aceb06898420b77589a84a48f7dd0a518183a82f4bf04be96d78d769ee4cbed6"
+      "bytes": 7143,
+      "sha256": "49653ef7bd5ef4ba040fe4c1c12b4d831be975226b1ea672634f75614c6f9bce"
     },
     {
       "id": "project_status_json",
@@ -73,8 +73,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable copy of the current project status for website and HF mirrors.",
       "exists": true,
-      "bytes": 9233,
-      "sha256": "f74b7265bebf28cf23688e81ea9aa392488ac110dcd9a8c9ebafb7f52935db35"
+      "bytes": 9176,
+      "sha256": "aa9ca62f30793c8d4c57e54133cb2beffeed14abff4446740b83218250279971"
     },
     {
       "id": "research_roadmap",
@@ -84,8 +84,8 @@
       "surface": "repo_hf",
       "shows": "Defines the staged path from public-sample task development to multi-episode held-out evaluation and larger omni-model extensions.",
       "exists": true,
-      "bytes": 6645,
-      "sha256": "6152c3981806beb454fd7e99a3e9e0e3a1637840571c267ec9315e78494df5f1"
+      "bytes": 6675,
+      "sha256": "314a2596075152da747ec9b983c2ac0110ec6213125005494f7a4d7d9d20f2e8"
     },
     {
       "id": "research_roadmap_json",
@@ -93,10 +93,10 @@
       "path": "docs/data/research_roadmap.json",
       "kind": "project_path",
       "surface": "website_hf",
-      "shows": "Machine-readable staged roadmap for the website and Hugging Face mirrors.",
+      "shows": "Machine-readable research roadmap for the website and Hugging Face mirrors.",
       "exists": true,
-      "bytes": 5727,
-      "sha256": "ac9be2d90db281a21df4956c792d8cd77648aa6c7dd4c12b5c3371fe951147b4"
+      "bytes": 5752,
+      "sha256": "b86d2bfdaeb0f3bd2e162164333122c5c1b4d09a2224143f37307f3b2d07708f"
     },
     {
       "id": "foundation_model_plan",
@@ -106,8 +106,8 @@
       "surface": "repo_hf",
       "shows": "Defines the post-data-gate backbone choices: Qwen3-Omni first, Cosmos 3 for world modeling, and VLA/policy models after action-target conversion.",
       "exists": true,
-      "bytes": 6562,
-      "sha256": "fcfdadd1b1b79c47ea99076d81b7624ac99c5d6a86df5b5c5e5ff8d16a1ec0f6"
+      "bytes": 6553,
+      "sha256": "816ff621d9e987425f545fd7471f1548a2bd427ccfb959704d2d4dc603b71d74"
     },
     {
       "id": "foundation_model_plan_json",
@@ -128,8 +128,8 @@
       "surface": "repo",
       "shows": "Defines the implemented scope, setup-stage items, and multi-episode prerequisites.",
       "exists": true,
-      "bytes": 11294,
-      "sha256": "b1643d23496ce868aec5d2dc6799013b99f86ee0af54f6175742bb585c6222f5"
+      "bytes": 11323,
+      "sha256": "5c1f41ab70424aa139fa3062cce2d5182995312f072560c403bc505c5cf3b302"
     },
     {
       "id": "project_packet",
@@ -139,8 +139,8 @@
       "surface": "website_hf",
       "shows": "Gives a short project path with scope status and public surfaces.",
       "exists": true,
-      "bytes": 7722,
-      "sha256": "06aa7bc21ee856c3d99cb95e30bb6f376eb8deed4289774740af4b9ed353749f"
+      "bytes": 7651,
+      "sha256": "f8b57892a6031620402554b1e3105115ef2fa27e0db65bb3263f0fd353947a3d"
     },
     {
       "id": "artifact_guide",
@@ -150,8 +150,8 @@
       "surface": "repo_hf",
       "shows": "Gives the human-readable map from project scope to data, tasks, platform mirrors, and scale-up status.",
       "exists": true,
-      "bytes": 16885,
-      "sha256": "755727a141df5a5e77eb9031d76f686d884e3ad11419c186d1f3c684f29a6aa0"
+      "bytes": 16897,
+      "sha256": "81c731cd65f7496abb275b9a60dcacca1407a132bba1748303854f2a1eb435df"
     },
     {
       "id": "official_dataset_card_alignment",
@@ -162,7 +162,7 @@
       "shows": "Aligns public dataset wording with the official gated Xperience-10M card, public sample card, HF API metadata, and current project coverage.",
       "exists": true,
       "bytes": 10625,
-      "sha256": "9e51442cb13cb4074d73fb212b49c56e6e4ae6dea10fd135a4020f7488e49be0"
+      "sha256": "1ad8d533b402091be66b320c36b4a0df5ec474906da4b3d079cbf860fd50dccf"
     },
     {
       "id": "official_dataset_card_alignment_json",
@@ -195,7 +195,7 @@
       "shows": "Machine-readable source-alignment pass/fail check for repo, website, and HF surfaces.",
       "exists": true,
       "bytes": 4432,
-      "sha256": "129d54143e4ea914e4c1adea6280871164a029da55ae3f668f01e09ebbe75225"
+      "sha256": "c12a17b949b50b4d246e2ea985ca3b872ed9891400b3c2e9fc4593dcff1bf269"
     },
     {
       "id": "source_alignment_validator",
@@ -227,8 +227,8 @@
       "surface": "repo_hf",
       "shows": "Defines the window unit, chronological split, task metrics, leakage controls, and current limitations.",
       "exists": true,
-      "bytes": 5877,
-      "sha256": "5761f8c81f543a5f8cd4d6a23d3a0986c1221eb2c05e362e01ff3c37407de08f"
+      "bytes": 5896,
+      "sha256": "51710edf5ac8d2f6737321089a4e6d0ae876dc26adf6edf33e2b8710acb8e4c2"
     },
     {
       "id": "evaluation_protocol_json",
@@ -238,8 +238,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable protocol generated from committed task metrics for website and HF mirrors.",
       "exists": true,
-      "bytes": 13604,
-      "sha256": "2a9cb2323eba460fb89ef54d149aad1e97902fd7f369813e1ab4f2d2f7b9a85b"
+      "bytes": 13624,
+      "sha256": "2e52de628b00d443117183b9276261cf30a5a07734e82fbd7fbbfa123bf28a82"
     },
     {
       "id": "evaluation_protocol_builder",
@@ -348,8 +348,8 @@
       "surface": "repo_hf",
       "shows": "Catalogs public figures, charts, modality thumbnails, dimensions, hashes, roles, and source scripts.",
       "exists": true,
-      "bytes": 4712,
-      "sha256": "bc1e1b695649270c64f654aa260c9d52cfc372a765ca6c993ef07145e8050f6e"
+      "bytes": 4947,
+      "sha256": "71e39f890e630923bf4a1699ee3069b0a8c29dd0cb7de637250dfeaa3b744eea"
     },
     {
       "id": "figure_index_json",
@@ -359,8 +359,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable visual asset index for website and Hugging Face mirrors.",
       "exists": true,
-      "bytes": 13434,
-      "sha256": "158dddd13e7868f35a919bfdd8d18a4f7cba52b872ab1ad68c7f321a21578fd9"
+      "bytes": 14115,
+      "sha256": "b6033bc57887e5d3421e661077194176c36dd6f08419a65fa44b245de8842ca2"
     },
     {
       "id": "figure_index_builder",
@@ -370,8 +370,8 @@
       "surface": "repo_hf",
       "shows": "Regenerates visual-asset hashes, dimensions, and source-script provenance.",
       "exists": true,
-      "bytes": 12971,
-      "sha256": "926f7efd274485b1c4772cea8cd61c96cca57ef04cddf4c23271de604042f135"
+      "bytes": 13397,
+      "sha256": "c5a4d3ba1764d9aaaf96c04b170bdc144d02b9b30230b411097544bf647786be"
     },
     {
       "id": "brand_assets_json",
@@ -436,8 +436,8 @@
       "surface": "repo_hf",
       "shows": "Keeps the repo, website, and Hugging Face cards aligned as one cohesive research project surface.",
       "exists": true,
-      "bytes": 1982,
-      "sha256": "4ddb758fc32354e74fae8aacef71e590a34c108b0856f7782fdc0bc71a7e79fc"
+      "bytes": 1939,
+      "sha256": "516ca8e7b67460a6151785bfa186165f37e394ab56114bab05481c878265ffe5"
     },
     {
       "id": "public_surface_qa_json",
@@ -448,7 +448,7 @@
       "volatile": true,
       "shows": "Machine-readable report for SEO/social metadata, accessible tab semantics, public links, project links, and clear project presentation.",
       "exists": true,
-      "bytes": 5632,
+      "bytes": 5591,
       "hash_policy": "existence_and_size_only"
     },
     {
@@ -459,8 +459,8 @@
       "surface": "repo_hf",
       "shows": "Regenerates the public presentation report before release.",
       "exists": true,
-      "bytes": 11863,
-      "sha256": "7311b7974bcc3866f08ca2f01bff13ea88898cae71b83c33247c2ffa32c37b84"
+      "bytes": 11986,
+      "sha256": "949fffe589d35db8ba342c44d462702acf00be129f4d3392a239496f797252fa"
     },
     {
       "id": "task_surface_integrity",
@@ -551,8 +551,8 @@
       "surface": "repo_hf",
       "shows": "Defines public reproduction commands, expected outputs, and non-reproducible scale-up boundaries.",
       "exists": true,
-      "bytes": 6346,
-      "sha256": "48d82bbdfeba88a63100f55e19ef7e1392c0e633b5f3f0c648444f9ed1c96e02"
+      "bytes": 6380,
+      "sha256": "19ca27c3771d22843ee8257aa1cd8ae66e17609a84f6d78e70c55deac2a6ce5b"
     },
     {
       "id": "reproducibility_matrix",
@@ -562,8 +562,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable reproduction steps with expected artifacts and public boundaries.",
       "exists": true,
-      "bytes": 5229,
-      "sha256": "928eb95b95d4028a7f21135584c37ff141833e26178cc918a874a6448f8c5ded"
+      "bytes": 5263,
+      "sha256": "8b476b240362a6f3fb2c33eef91f7766175fb05c1528dd6f052b9ed24d1fe330"
     },
     {
       "id": "artifact_index_builder",
@@ -573,8 +573,8 @@
       "surface": "repo_hf",
       "shows": "Generates the selective artifact catalog from local files.",
       "exists": true,
-      "bytes": 26573,
-      "sha256": "62d4c5dcdb8fe2757dde39af4b36d3c2e2601e50062bd156b11eff1663a9dd01"
+      "bytes": 26575,
+      "sha256": "7ce8429b8ec87d56f0efed41187f760efd71e98989b32dd7ecd2b43fdc85a368"
     },
     {
       "id": "publication_audit",
@@ -597,7 +597,7 @@
       "volatile": true,
       "shows": "Separates setup paths from completed held-out-episode results.",
       "exists": true,
-      "bytes": 18951,
+      "bytes": 18973,
       "hash_policy": "existence_and_size_only"
     },
     {
@@ -621,7 +621,7 @@
       "volatile": true,
       "shows": "Confirms local website links, anchors, JSON data files, and referenced images resolve.",
       "exists": true,
-      "bytes": 14746,
+      "bytes": 14923,
       "hash_policy": "existence_and_size_only"
     },
     {
@@ -654,8 +654,8 @@
       "surface": "website_hf",
       "shows": "Mirrors task metrics for the static dashboard.",
       "exists": true,
-      "bytes": 25369,
-      "sha256": "884c9caf1517126f4c93b5c903121d29806922ecbd42bc2c34e2e568b9af83e3"
+      "bytes": 25293,
+      "sha256": "22cbbaa5f0b3c889776e746257d99c6ceb6fd913c249d58353d641cac205d7d3"
     },
     {
       "id": "feature_manifest",
@@ -775,8 +775,8 @@
       "surface": "website_hf",
       "shows": "Shows the raw-episode to artifact pipeline with verified labels.",
       "exists": true,
-      "bytes": 707296,
-      "sha256": "3eb7767a6ec78ef50847b79fad16e2438e84f449782f3fe8367417bece1e289c"
+      "bytes": 704575,
+      "sha256": "c90723cc4b1bf5490269af2df594849030ae8d4cc8176e1d1eab96fabf9412f9"
     },
     {
       "id": "architecture_figure",
@@ -786,8 +786,8 @@
       "surface": "website_hf",
       "shows": "Shows the shared feature pipeline and minimal/neural head families.",
       "exists": true,
-      "bytes": 767403,
-      "sha256": "9e3b86acf7f7296eb7c4b1e5e74c7008f264ddce40485e0e923b1e2a2f2b8efe"
+      "bytes": 761507,
+      "sha256": "076c2e463ddce473e9138ac6f3615152d59031d6be2aa5c3d9ae1ace3d3f6c83"
     },
     {
       "id": "qwen_data_access_status",
@@ -797,8 +797,8 @@
       "surface": "repo_hf",
       "shows": "Summarizes the staging requirement before the held-out Qwen3-Omni pilot can report metrics.",
       "exists": true,
-      "bytes": 3142,
-      "sha256": "e55fbb9cadb6eb50fca8d787156e603067e99710b9b4c8efeafb638e5ea9064c"
+      "bytes": 3101,
+      "sha256": "8e8e80c0b5442f267ad7cc87a469c5a5b4ed8ec8e10578437ce51094e82b5764"
     },
     {
       "id": "multi_episode_access_status",
@@ -808,8 +808,8 @@
       "surface": "repo_hf",
       "shows": "Documents the public multi-episode access status and 32-episode pilot selection.",
       "exists": true,
-      "bytes": 2532,
-      "sha256": "7881cfca4a2b33712f52c55edea0db54b417169fff4fce636c1c7709ba43b95d"
+      "bytes": 2339,
+      "sha256": "14b8b629ddb548aa1ea39d0a8dc3c84fe3cf26931c6928f41b760d6f3c4c2470"
     },
     {
       "id": "citation",

docs/data/evidence_contract.json

@@ -14,13 +14,13 @@
     },
     {
       "id": "research_roadmap",
-      "claim": "The staged research path is explicit.",
+      "claim": "The research roadmap is explicit.",
       "status": "current",
       "evidence": [
         "RESEARCH_ROADMAP.md",
         "docs/data/research_roadmap.json"
       ],
-      "boundary": "connects public-sample task development to multi-episode staging, Qwen3-Omni LoRA, robustness runs, and larger omni-model extensions"
+      "boundary": "connects public-sample task development to multi-episode data preparation, Qwen3-Omni LoRA, robustness runs, and larger omni-model extensions"
     },
         {
             "id": "official_dataset_card_alignment",
@@ -171,18 +171,18 @@
         "results/omni_finetune/dataset_manifest.json",
         "results/omni_finetune/metrics_eval.json"
       ],
-      "boundary": "one episode and 128 train windows; full metrics require completed multi-episode staging and held-out evaluation"
+      "boundary": "one episode and 128 train windows; full metrics require completed multi-episode data preparation and held-out evaluation"
     },
     {
       "id": "multi_episode_staging",
-      "claim": "The Qwen3-Omni LoRA pilot is in multi-episode staging.",
-      "status": "data_staging",
+      "claim": "The Qwen3-Omni LoRA pilot is in selected multi-episode preparation.",
+      "status": "data_preparation",
       "evidence": [
         "results/omni_finetune/DATA_ACCESS_STATUS.md",
         "results/omni_finetune/MULTI_EPISODE_ACCESS_STATUS.md",
         "results/omni_finetune/source_discovery.json"
       ],
-      "boundary": "the gated full dataset is available for the selected pilot; held-out metrics come after manifest construction, training, and test evaluation"
+      "boundary": "the gated Xperience-10M dataset is available for the selected pilot; held-out metrics come after manifest construction, training, and test evaluation"
     },
     {
       "id": "scale_up_status_check",

docs/data/mirror_parity.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-04T16:04:34+00:00",
+  "generated_at_utc": "2026-06-04T16:34:37+00:00",
   "hf_root": "hf_publish",
   "summary": {
     "group_count": 101,
@@ -71,27 +71,27 @@
       "local": {
         "path": "repo:docs/data/artifact_index.json",
         "exists": true,
-        "bytes": 32301,
-        "sha256": "779d6fa86dba767cee4defe1c690c2ce59f8835f0621caa5a060ac6434bd9f20"
+        "bytes": 32303,
+        "sha256": "1892a5a8a54ceb58658045526579d25d47a69fd2ef17c44ec4fbb1d9dd207868"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/artifact_index.json",
           "exists": true,
-          "bytes": 32301,
-          "sha256": "779d6fa86dba767cee4defe1c690c2ce59f8835f0621caa5a060ac6434bd9f20"
+          "bytes": 32303,
+          "sha256": "1892a5a8a54ceb58658045526579d25d47a69fd2ef17c44ec4fbb1d9dd207868"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/artifact_index.json",
           "exists": true,
-          "bytes": 32301,
-          "sha256": "779d6fa86dba767cee4defe1c690c2ce59f8835f0621caa5a060ac6434bd9f20"
+          "bytes": 32303,
+          "sha256": "1892a5a8a54ceb58658045526579d25d47a69fd2ef17c44ec4fbb1d9dd207868"
         },
         "hf_model": {
           "path": "hf_model:metrics/artifact_index.json",
           "exists": true,
-          "bytes": 32301,
-          "sha256": "779d6fa86dba767cee4defe1c690c2ce59f8835f0621caa5a060ac6434bd9f20"
+          "bytes": 32303,
+          "sha256": "1892a5a8a54ceb58658045526579d25d47a69fd2ef17c44ec4fbb1d9dd207868"
         }
       },
       "failures": []
@@ -133,27 +133,27 @@
       "local": {
         "path": "repo:docs/data/evidence_contract.json",
         "exists": true,
-        "bytes": 12107,
-        "sha256": "4099cb6e56db9fafd74804a49f38c49f76da162ff8fad62bdeb2a830d202c423"
+        "bytes": 12147,
+        "sha256": "d027ab3ee0b147957b1196211261949ff607fc7a2bb43ff3b92ee46e5a902fa0"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/evidence_contract.json",
           "exists": true,
-          "bytes": 12107,
-          "sha256": "4099cb6e56db9fafd74804a49f38c49f76da162ff8fad62bdeb2a830d202c423"
+          "bytes": 12147,
+          "sha256": "d027ab3ee0b147957b1196211261949ff607fc7a2bb43ff3b92ee46e5a902fa0"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/evidence_contract.json",
           "exists": true,
-          "bytes": 12107,
-          "sha256": "4099cb6e56db9fafd74804a49f38c49f76da162ff8fad62bdeb2a830d202c423"
+          "bytes": 12147,
+          "sha256": "d027ab3ee0b147957b1196211261949ff607fc7a2bb43ff3b92ee46e5a902fa0"
         },
         "hf_model": {
           "path": "hf_model:metrics/evidence_contract.json",
           "exists": true,
-          "bytes": 12107,
-          "sha256": "4099cb6e56db9fafd74804a49f38c49f76da162ff8fad62bdeb2a830d202c423"
+          "bytes": 12147,
+          "sha256": "d027ab3ee0b147957b1196211261949ff607fc7a2bb43ff3b92ee46e5a902fa0"
         }
       },
       "failures": []
@@ -444,26 +444,26 @@
         "path": "repo:docs/data/publication_audit.json",
         "exists": true,
         "bytes": 7289,
-        "sha256": "c918d047780dedd2586a095fedc30bb083820447e4cc8eef826e631da29d69c7"
+        "sha256": "96474643a8610c61c8411ed3fa775b3e05f34e4859562df3cb19b7898d5f1bbb"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/publication_audit.json",
           "exists": true,
           "bytes": 7289,
-          "sha256": "c918d047780dedd2586a095fedc30bb083820447e4cc8eef826e631da29d69c7"
+          "sha256": "96474643a8610c61c8411ed3fa775b3e05f34e4859562df3cb19b7898d5f1bbb"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/publication_audit.json",
           "exists": true,
           "bytes": 7289,
-          "sha256": "c918d047780dedd2586a095fedc30bb083820447e4cc8eef826e631da29d69c7"
+          "sha256": "96474643a8610c61c8411ed3fa775b3e05f34e4859562df3cb19b7898d5f1bbb"
         },
         "hf_model": {
           "path": "hf_model:metrics/publication_audit.json",
           "exists": true,
           "bytes": 7289,
-          "sha256": "c918d047780dedd2586a095fedc30bb083820447e4cc8eef826e631da29d69c7"
+          "sha256": "96474643a8610c61c8411ed3fa775b3e05f34e4859562df3cb19b7898d5f1bbb"
         }
       },
       "failures": []
@@ -475,26 +475,26 @@
         "path": "repo:docs/data/public_surface_qa.json",
         "exists": true,
         "bytes": 5591,
-        "sha256": "9caca879dd69d6a87693ab6058903b2c2fb2389db0cb25f1cdff39a987b29135"
+        "sha256": "10b6eba408d6a88543ea93319b983e1abd42a4cf846c7cd0375e95d8ab18eea2"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/public_surface_qa.json",
           "exists": true,
           "bytes": 5591,
-          "sha256": "9caca879dd69d6a87693ab6058903b2c2fb2389db0cb25f1cdff39a987b29135"
+          "sha256": "10b6eba408d6a88543ea93319b983e1abd42a4cf846c7cd0375e95d8ab18eea2"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/public_surface_qa.json",
           "exists": true,
           "bytes": 5591,
-          "sha256": "9caca879dd69d6a87693ab6058903b2c2fb2389db0cb25f1cdff39a987b29135"
+          "sha256": "10b6eba408d6a88543ea93319b983e1abd42a4cf846c7cd0375e95d8ab18eea2"
         },
         "hf_model": {
           "path": "hf_model:metrics/public_surface_qa.json",
           "exists": true,
           "bytes": 5591,
-          "sha256": "9caca879dd69d6a87693ab6058903b2c2fb2389db0cb25f1cdff39a987b29135"
+          "sha256": "10b6eba408d6a88543ea93319b983e1abd42a4cf846c7cd0375e95d8ab18eea2"
         }
       },
       "failures": []
@@ -598,27 +598,27 @@
       "local": {
         "path": "repo:docs/data/research_roadmap.json",
         "exists": true,
-        "bytes": 5727,
-        "sha256": "ac9be2d90db281a21df4956c792d8cd77648aa6c7dd4c12b5c3371fe951147b4"
+        "bytes": 5752,
+        "sha256": "b86d2bfdaeb0f3bd2e162164333122c5c1b4d09a2224143f37307f3b2d07708f"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/research_roadmap.json",
           "exists": true,
-          "bytes": 5727,
-          "sha256": "ac9be2d90db281a21df4956c792d8cd77648aa6c7dd4c12b5c3371fe951147b4"
+          "bytes": 5752,
+          "sha256": "b86d2bfdaeb0f3bd2e162164333122c5c1b4d09a2224143f37307f3b2d07708f"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/research_roadmap.json",
           "exists": true,
-          "bytes": 5727,
-          "sha256": "ac9be2d90db281a21df4956c792d8cd77648aa6c7dd4c12b5c3371fe951147b4"
+          "bytes": 5752,
+          "sha256": "b86d2bfdaeb0f3bd2e162164333122c5c1b4d09a2224143f37307f3b2d07708f"
         },
         "hf_model": {
           "path": "hf_model:metrics/research_roadmap.json",
           "exists": true,
-          "bytes": 5727,
-          "sha256": "ac9be2d90db281a21df4956c792d8cd77648aa6c7dd4c12b5c3371fe951147b4"
+          "bytes": 5752,
+          "sha256": "b86d2bfdaeb0f3bd2e162164333122c5c1b4d09a2224143f37307f3b2d07708f"
         }
       },
       "failures": []
@@ -629,27 +629,27 @@
       "local": {
         "path": "repo:docs/data/research_roadmap_interactive.json",
         "exists": true,
-        "bytes": 131493,
-        "sha256": "a93086069f48fead2c65c105a66186ee30922e8e368d0013f28e89fa8dbb8032"
+        "bytes": 131507,
+        "sha256": "c63be99d3f575016c41295e5ba79fb732bd23c90d657c63bc147b55628c7407f"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/research_roadmap_interactive.json",
           "exists": true,
-          "bytes": 131493,
-          "sha256": "a93086069f48fead2c65c105a66186ee30922e8e368d0013f28e89fa8dbb8032"
+          "bytes": 131507,
+          "sha256": "c63be99d3f575016c41295e5ba79fb732bd23c90d657c63bc147b55628c7407f"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/research_roadmap_interactive.json",
           "exists": true,
-          "bytes": 131493,
-          "sha256": "a93086069f48fead2c65c105a66186ee30922e8e368d0013f28e89fa8dbb8032"
+          "bytes": 131507,
+          "sha256": "c63be99d3f575016c41295e5ba79fb732bd23c90d657c63bc147b55628c7407f"
         },
         "hf_model": {
           "path": "hf_model:metrics/research_roadmap_interactive.json",
           "exists": true,
-          "bytes": 131493,
-          "sha256": "a93086069f48fead2c65c105a66186ee30922e8e368d0013f28e89fa8dbb8032"
+          "bytes": 131507,
+          "sha256": "c63be99d3f575016c41295e5ba79fb732bd23c90d657c63bc147b55628c7407f"
         }
       },
       "failures": []
@@ -753,27 +753,27 @@
       "local": {
         "path": "repo:docs/data/scope_claims_audit.json",
         "exists": true,
-        "bytes": 18951,
-        "sha256": "639a9761e036a9d98496a334d0538a8b9924ebbd7f1c99c39b1c32a6fb2a799b"
+        "bytes": 19497,
+        "sha256": "5c32c92bc89c2cee451f2560c4ebed1edaa45064de19968912344bedb7bbd160"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/scope_claims_audit.json",
           "exists": true,
-          "bytes": 18951,
-          "sha256": "639a9761e036a9d98496a334d0538a8b9924ebbd7f1c99c39b1c32a6fb2a799b"
+          "bytes": 19497,
+          "sha256": "5c32c92bc89c2cee451f2560c4ebed1edaa45064de19968912344bedb7bbd160"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/scope_claims_audit.json",
           "exists": true,
-          "bytes": 18951,
-          "sha256": "639a9761e036a9d98496a334d0538a8b9924ebbd7f1c99c39b1c32a6fb2a799b"
+          "bytes": 19497,
+          "sha256": "5c32c92bc89c2cee451f2560c4ebed1edaa45064de19968912344bedb7bbd160"
         },
         "hf_model": {
           "path": "hf_model:metrics/scope_claims_audit.json",
           "exists": true,
-          "bytes": 18951,
-          "sha256": "639a9761e036a9d98496a334d0538a8b9924ebbd7f1c99c39b1c32a6fb2a799b"
+          "bytes": 19497,
+          "sha256": "5c32c92bc89c2cee451f2560c4ebed1edaa45064de19968912344bedb7bbd160"
         }
       },
       "failures": []
@@ -816,26 +816,26 @@
         "path": "repo:docs/data/source_alignment_audit.json",
         "exists": true,
         "bytes": 4432,
-        "sha256": "63fc3ea43c73a9c27707041449fa5fe29ed6246d096c585f64972066c7086e64"
+        "sha256": "96c7adc61c869fab71ef34ec2f6ec4f5f88af844509bd3d51d3818732d1f84b6"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/source_alignment_audit.json",
           "exists": true,
           "bytes": 4432,
-          "sha256": "63fc3ea43c73a9c27707041449fa5fe29ed6246d096c585f64972066c7086e64"
+          "sha256": "96c7adc61c869fab71ef34ec2f6ec4f5f88af844509bd3d51d3818732d1f84b6"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/source_alignment_audit.json",
           "exists": true,
           "bytes": 4432,
-          "sha256": "63fc3ea43c73a9c27707041449fa5fe29ed6246d096c585f64972066c7086e64"
+          "sha256": "96c7adc61c869fab71ef34ec2f6ec4f5f88af844509bd3d51d3818732d1f84b6"
         },
         "hf_model": {
           "path": "hf_model:metrics/source_alignment_audit.json",
           "exists": true,
           "bytes": 4432,
-          "sha256": "63fc3ea43c73a9c27707041449fa5fe29ed6246d096c585f64972066c7086e64"
+          "sha256": "96c7adc61c869fab71ef34ec2f6ec4f5f88af844509bd3d51d3818732d1f84b6"
         }
       },
       "failures": []
@@ -846,27 +846,27 @@
       "local": {
         "path": "repo:docs/data/summary_metrics.json",
         "exists": true,
-        "bytes": 25369,
-        "sha256": "884c9caf1517126f4c93b5c903121d29806922ecbd42bc2c34e2e568b9af83e3"
+        "bytes": 25293,
+        "sha256": "22cbbaa5f0b3c889776e746257d99c6ceb6fd913c249d58353d641cac205d7d3"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/summary_metrics.json",
           "exists": true,
-          "bytes": 25369,
-          "sha256": "884c9caf1517126f4c93b5c903121d29806922ecbd42bc2c34e2e568b9af83e3"
+          "bytes": 25293,
+          "sha256": "22cbbaa5f0b3c889776e746257d99c6ceb6fd913c249d58353d641cac205d7d3"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/summary_metrics.json",
           "exists": true,
-          "bytes": 25369,
-          "sha256": "884c9caf1517126f4c93b5c903121d29806922ecbd42bc2c34e2e568b9af83e3"
+          "bytes": 25293,
+          "sha256": "22cbbaa5f0b3c889776e746257d99c6ceb6fd913c249d58353d641cac205d7d3"
         },
         "hf_model": {
           "path": "hf_model:metrics/summary_metrics.json",
           "exists": true,
-          "bytes": 25369,
-          "sha256": "884c9caf1517126f4c93b5c903121d29806922ecbd42bc2c34e2e568b9af83e3"
+          "bytes": 25293,
+          "sha256": "22cbbaa5f0b3c889776e746257d99c6ceb6fd913c249d58353d641cac205d7d3"
         }
       },
       "failures": []
@@ -940,26 +940,26 @@
         "path": "repo:docs/data/website_integrity.json",
         "exists": true,
         "bytes": 14923,
-        "sha256": "20d63e8eb5b33ef301c6ed0bb30fdc343d4fbeaf0fdfbd2c577956f67cbb0056"
+        "sha256": "9a93b18174da0445db809132d0899f0213df99f2d99ad2a219bbb77c84657cf9"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/website_integrity.json",
           "exists": true,
           "bytes": 14923,
-          "sha256": "20d63e8eb5b33ef301c6ed0bb30fdc343d4fbeaf0fdfbd2c577956f67cbb0056"
+          "sha256": "9a93b18174da0445db809132d0899f0213df99f2d99ad2a219bbb77c84657cf9"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/website_integrity.json",
           "exists": true,
           "bytes": 14923,
-          "sha256": "20d63e8eb5b33ef301c6ed0bb30fdc343d4fbeaf0fdfbd2c577956f67cbb0056"
+          "sha256": "9a93b18174da0445db809132d0899f0213df99f2d99ad2a219bbb77c84657cf9"
         },
         "hf_model": {
           "path": "hf_model:metrics/website_integrity.json",
           "exists": true,
           "bytes": 14923,
-          "sha256": "20d63e8eb5b33ef301c6ed0bb30fdc343d4fbeaf0fdfbd2c577956f67cbb0056"
+          "sha256": "9a93b18174da0445db809132d0899f0213df99f2d99ad2a219bbb77c84657cf9"
         }
       },
       "failures": []
@@ -1692,21 +1692,21 @@
       "local": {
         "path": "repo:scripts/build_artifact_index.py",
         "exists": true,
-        "bytes": 26573,
-        "sha256": "62d4c5dcdb8fe2757dde39af4b36d3c2e2601e50062bd156b11eff1663a9dd01"
+        "bytes": 26575,
+        "sha256": "7ce8429b8ec87d56f0efed41187f760efd71e98989b32dd7ecd2b43fdc85a368"
       },
       "mirrors": {
         "hf_artifacts": {
           "path": "hf_artifacts:scripts/build_artifact_index.py",
           "exists": true,
-          "bytes": 26573,
-          "sha256": "62d4c5dcdb8fe2757dde39af4b36d3c2e2601e50062bd156b11eff1663a9dd01"
+          "bytes": 26575,
+          "sha256": "7ce8429b8ec87d56f0efed41187f760efd71e98989b32dd7ecd2b43fdc85a368"
         },
         "hf_model": {
           "path": "hf_model:scripts/build_artifact_index.py",
           "exists": true,
-          "bytes": 26573,
-          "sha256": "62d4c5dcdb8fe2757dde39af4b36d3c2e2601e50062bd156b11eff1663a9dd01"
+          "bytes": 26575,
+          "sha256": "7ce8429b8ec87d56f0efed41187f760efd71e98989b32dd7ecd2b43fdc85a368"
         }
       },
       "failures": []
@@ -2217,21 +2217,21 @@
       "local": {
         "path": "repo:docs/index.html",
         "exists": true,
-        "bytes": 173399,
-        "sha256": "88336fde84cd16d1b2e66e722e06817698cbcd26be985804ec52890b447e9fd1"
+        "bytes": 173405,
+        "sha256": "31aec0e4dcd4db5701916a75784dce65a2890839101a88cef43e67081853a6a9"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:index.html",
           "exists": true,
-          "bytes": 173399,
-          "sha256": "88336fde84cd16d1b2e66e722e06817698cbcd26be985804ec52890b447e9fd1"
+          "bytes": 173405,
+          "sha256": "31aec0e4dcd4db5701916a75784dce65a2890839101a88cef43e67081853a6a9"
         },
         "hf_artifacts_docs": {
           "path": "hf_artifacts:docs/index.html",
           "exists": true,
-          "bytes": 173399,
-          "sha256": "88336fde84cd16d1b2e66e722e06817698cbcd26be985804ec52890b447e9fd1"
+          "bytes": 173405,
+          "sha256": "31aec0e4dcd4db5701916a75784dce65a2890839101a88cef43e67081853a6a9"
         }
       },
       "failures": []
@@ -2875,27 +2875,27 @@
       "local": {
         "path": "repo:PROJECT_BRIEF.md",
         "exists": true,
-        "bytes": 3596,
-        "sha256": "0a9557f57088644466b19e91202ec0304f9394e3abf69d7861a878d192d1c1aa"
+        "bytes": 3594,
+        "sha256": "8df692f2d6b092097d5ef7d5039e7529abe2fd0a80ff3fb23fa14260fb5f9de2"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:PROJECT_BRIEF.md",
           "exists": true,
-          "bytes": 3596,
-          "sha256": "0a9557f57088644466b19e91202ec0304f9394e3abf69d7861a878d192d1c1aa"
+          "bytes": 3594,
+          "sha256": "8df692f2d6b092097d5ef7d5039e7529abe2fd0a80ff3fb23fa14260fb5f9de2"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:PROJECT_BRIEF.md",
           "exists": true,
-          "bytes": 3596,
-          "sha256": "0a9557f57088644466b19e91202ec0304f9394e3abf69d7861a878d192d1c1aa"
+          "bytes": 3594,
+          "sha256": "8df692f2d6b092097d5ef7d5039e7529abe2fd0a80ff3fb23fa14260fb5f9de2"
         },
         "hf_model": {
           "path": "hf_model:PROJECT_BRIEF.md",
           "exists": true,
-          "bytes": 3596,
-          "sha256": "0a9557f57088644466b19e91202ec0304f9394e3abf69d7861a878d192d1c1aa"
+          "bytes": 3594,
+          "sha256": "8df692f2d6b092097d5ef7d5039e7529abe2fd0a80ff3fb23fa14260fb5f9de2"
         }
       },
       "failures": []
@@ -2937,27 +2937,27 @@
       "local": {
         "path": "repo:RESEARCH_ROADMAP.md",
         "exists": true,
-        "bytes": 6668,
-        "sha256": "5f888e6a1eb64fa77f225b9474a4cffcff544419ac506dde1d388fac3fb9bd48"
+        "bytes": 6675,
+        "sha256": "314a2596075152da747ec9b983c2ac0110ec6213125005494f7a4d7d9d20f2e8"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:RESEARCH_ROADMAP.md",
           "exists": true,
-          "bytes": 6668,
-          "sha256": "5f888e6a1eb64fa77f225b9474a4cffcff544419ac506dde1d388fac3fb9bd48"
+          "bytes": 6675,
+          "sha256": "314a2596075152da747ec9b983c2ac0110ec6213125005494f7a4d7d9d20f2e8"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:RESEARCH_ROADMAP.md",
           "exists": true,
-          "bytes": 6668,
-          "sha256": "5f888e6a1eb64fa77f225b9474a4cffcff544419ac506dde1d388fac3fb9bd48"
+          "bytes": 6675,
+          "sha256": "314a2596075152da747ec9b983c2ac0110ec6213125005494f7a4d7d9d20f2e8"
         },
         "hf_model": {
           "path": "hf_model:RESEARCH_ROADMAP.md",
           "exists": true,
-          "bytes": 6668,
-          "sha256": "5f888e6a1eb64fa77f225b9474a4cffcff544419ac506dde1d388fac3fb9bd48"
+          "bytes": 6675,
+          "sha256": "314a2596075152da747ec9b983c2ac0110ec6213125005494f7a4d7d9d20f2e8"
         }
       },
       "failures": []
@@ -3093,26 +3093,26 @@
         "path": "repo:XPERIENCE10M_DATASET_CARD_ALIGNMENT.md",
         "exists": true,
         "bytes": 10625,
-        "sha256": "9e51442cb13cb4074d73fb212b49c56e6e4ae6dea10fd135a4020f7488e49be0"
+        "sha256": "1ad8d533b402091be66b320c36b4a0df5ec474906da4b3d079cbf860fd50dccf"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:XPERIENCE10M_DATASET_CARD_ALIGNMENT.md",
           "exists": true,
           "bytes": 10625,
-          "sha256": "9e51442cb13cb4074d73fb212b49c56e6e4ae6dea10fd135a4020f7488e49be0"
+          "sha256": "1ad8d533b402091be66b320c36b4a0df5ec474906da4b3d079cbf860fd50dccf"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:XPERIENCE10M_DATASET_CARD_ALIGNMENT.md",
           "exists": true,
           "bytes": 10625,
-          "sha256": "9e51442cb13cb4074d73fb212b49c56e6e4ae6dea10fd135a4020f7488e49be0"
+          "sha256": "1ad8d533b402091be66b320c36b4a0df5ec474906da4b3d079cbf860fd50dccf"
         },
         "hf_model": {
           "path": "hf_model:XPERIENCE10M_DATASET_CARD_ALIGNMENT.md",
           "exists": true,
           "bytes": 10625,
-          "sha256": "9e51442cb13cb4074d73fb212b49c56e6e4ae6dea10fd135a4020f7488e49be0"
+          "sha256": "1ad8d533b402091be66b320c36b4a0df5ec474906da4b3d079cbf860fd50dccf"
         }
       },
       "failures": []

docs/data/public_surface_qa.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Public Project Surface",
   "status": "pass",
-  "generated_at_utc": "2026-06-04T16:03:23+00:00",
+  "generated_at_utc": "2026-06-04T16:27:07+00:00",
   "scope": "Repo README, GitHub Pages HTML, Hugging Face Space card, artifact dataset card, and model card.",
   "checks": [
     {
@@ -18,7 +18,7 @@
         "website_integrity": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-04T16:02:00+00:00"
+          "generated_at_utc": "2026-06-04T16:19:06+00:00"
         },
         "rendered_site_check": {
           "exists": true,
@@ -33,7 +33,7 @@
         "source_alignment": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-04T16:03:23+00:00"
+          "generated_at_utc": "2026-06-04T16:18:24+00:00"
         },
         "scale_up_status": {
           "exists": true,
@@ -43,12 +43,12 @@
         "publication_package": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-04T16:02:01+00:00"
+          "generated_at_utc": "2026-06-04T16:18:29+00:00"
         },
         "mirror_parity": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-04T16:02:57+00:00"
+          "generated_at_utc": "2026-06-04T16:04:34+00:00"
         },
         "live_publication": {
           "exists": true,
@@ -101,7 +101,7 @@
       "reason": "Public copy should consistently present the project as Ropedia Xperience-10M, with the Qwen3-Omni scale-up status.",
       "marker_counts": {
         "Ropedia Xperience-10M Task Suite": 15,
-        "Xperience-10M": 136,
+        "Xperience-10M": 142,
         "12-task": 33,
         "Qwen3-Omni": 108,
         "128-episode pilot": 11

docs/data/publication_audit.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-04T16:03:25+00:00",
+  "generated_at_utc": "2026-06-04T16:33:03+00:00",
   "checks": [
     {
       "name": "required_publication_assets_present",

docs/data/research_roadmap.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Research Roadmap",
   "summary": "Staged path from the public-sample task lab to multi-episode held-out evaluation, foundation-model selection, and larger omni/world-model extensions.",
-  "current_decision_point": "Keep the public-sample task suite as the development harness, stage the selected official Xperience-10M relay for the held-out Qwen3-Omni pilot, then branch into Cosmos 3 world modeling and policy-model experiments after the data staging path is stable.",
+  "current_decision_point": "Keep the public-sample task suite as the development harness, prepare the selected official Xperience-10M episodes for the held-out Qwen3-Omni pilot, then branch into Cosmos 3 world modeling and policy-model experiments after the data preparation path is stable.",
   "phases": [
     {
       "id": "public_sample_task_lab",
@@ -30,7 +30,7 @@
       "id": "multi_episode_data_staging",
       "name": "Multi-Episode Data Staging",
       "status": "active",
-      "entry_condition": "Full-dataset access and enough storage for selected episodes.",
+      "entry_condition": "Gated dataset availability and enough storage for selected episodes.",
       "deliverables": [
         "128 selected episodes",
         "episode manifest",
@@ -73,7 +73,7 @@
       "id": "foundation_model_selection_matrix",
       "name": "Foundation-Model Selection Matrix",
       "status": "next",
-      "entry_condition": "The selected relay is staged or a 3-8 episode dry run is staged for preprocessing checks.",
+      "entry_condition": "The selected episodes are prepared or a 3-8 episode dry run is available for preprocessing checks.",
       "deliverables": [
         "backbone registry",
         "Cosmos 3 world-model branch plan",

docs/data/research_roadmap_interactive.json

@@ -2061,7 +2061,7 @@
     ],
     "status": "planning_artifact"
   },
-  "generated_at_utc": "2026-06-04T07:35:10+00:00",
+  "generated_at_utc": "2026-06-04T16:13:49+00:00",
   "omni_plan": {
     "adapter": "LoRA rank 16, alpha 32, dropout 0.05",
     "backbone": "Qwen/Qwen3-Omni-30B-A3B-Instruct",
@@ -2116,7 +2116,7 @@
         "held-out episode split",
         "source-discovery report"
       ],
-      "entry_condition": "Full-dataset access and enough storage for selected episodes.",
+      "entry_condition": "Gated dataset availability and enough storage for selected episodes.",
       "id": "multi_episode_data_staging",
       "name": "Multi-Episode Data Staging",
       "reader_takeaway": "The next scale decision is data staging, with train/test separation at the episode level.",
@@ -2161,7 +2161,7 @@
         "OpenVLA/openpi/GR00T policy-branch candidates",
         "model-specific evaluation additions"
       ],
-      "entry_condition": "Selected pilot episodes are prepared or a 3-8 episode dry run is available for preprocessing checks.",
+      "entry_condition": "The selected episodes are prepared or a 3-8 episode dry run is available for preprocessing checks.",
       "id": "foundation_model_selection_matrix",
       "name": "Foundation-Model Selection Matrix",
       "reader_takeaway": "Qwen3-Omni remains the first trainable held-out pilot; Cosmos 3 is the first world-model branch; VLA/policy models wait for explicit action targets.",
@@ -2211,14 +2211,14 @@
     }
   ],
   "scale_up": {
-    "access_status": "The gated full dataset is available for selected multi-episode pilot preparation.",
+    "access_status": "The gated Xperience-10M dataset is available for selected multi-episode pilot preparation.",
     "candidate_scan_top_level_sessions": 802,
     "estimated_bytes": 298188841943,
     "exclude": [
       "visualization.rrd"
     ],
     "selection_strategy": "stratified_round_robin_by_top_level_session",
-    "status": "selected_pilot_preparation",
+    "status": "selected_relay_in_progress",
     "target_episodes": 128,
     "valid_candidates": 12102
   },

docs/data/scope_claims_audit.json

@@ -1,13 +1,13 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-04T07:45:04+00:00",
+  "generated_at_utc": "2026-06-04T16:27:09+00:00",
   "summary": {
     "qwen3_omni_32_episode_claim": false,
     "dataset_manifest_num_episodes": 1,
     "dataset_manifest_num_samples": 128,
     "training_metadata_num_train_samples": 128,
     "source_discovery_ready_for_32_episode_pilot": false,
-    "historical_identifier_count": 159,
+    "historical_identifier_count": 157,
     "public_32_episode_status_file_count": 3,
     "failure_count": 0
   },
@@ -90,7 +90,7 @@
     {
       "name": "historical_32ep_identifiers_are_confined_to_readiness_artifacts",
       "status": "pass",
-      "detail": "historical identifiers found in result provenance files=159",
+      "detail": "historical identifiers found in result provenance files=157",
       "evidence": [
         "results/omni_finetune/"
       ]
@@ -171,24 +171,6 @@
       ],
       "example": "- Dataset: `results/omni_finetune/xperience10m_qwen3_omni_32ep_dataset/dataset.jsonl`"
     },
-    {
-      "classification": "historical_identifier_in_readiness_artifact",
-      "path": "results/omni_finetune/XPERIENCE10M_128_RELAY_AND_FINETUNE_PLAN.md",
-      "line": 42,
-      "patterns": [
-        "ropedia-episode-task-suite"
-      ],
-      "example": "export RELAY_WORKDIR=/path/to/ropedia-episode-task-suite"
-    },
-    {
-      "classification": "historical_identifier_in_readiness_artifact",
-      "path": "results/omni_finetune/XPERIENCE10M_128_RELAY_AND_FINETUNE_PLAN.md",
-      "line": 45,
-      "patterns": [
-        "ropedia-episode-task-suite"
-      ],
-      "example": "export TRAINING_REPO=/path/to/ropedia-episode-task-suite"
-    },
     {
       "classification": "historical_identifier_in_readiness_artifact",
       "path": "results/omni_finetune/config.yaml",
@@ -428,8 +410,30 @@
         "ropedia-episode-task-suite"
       ],
       "example": "{\"id\": \"xperience-10m-sample:qa:42\", \"episode_id\": \"xperience-10m-sample\", \"split\": \"train\", \"target\": \"episode_qa\", \"prompt_type\": \"json_episode_understanding\", \"center_window\": {\"start_frame\": 840, \"end_frame\": 859, \"num_frames\": 20}, \"media\": {\"video_paths\""
+    },
+    {
+      "classification": "historical_identifier_in_readiness_artifact",
+      "path": "results/omni_finetune/dataset.jsonl",
+      "line": 20,
+      "patterns": [
+        "qwen3_omni_32ep",
+        "xperience10m_qwen3_omni_32ep",
+        "ropedia-episode-task-suite"
+      ],
+      "example": "{\"id\": \"xperience-10m-sample:qa:43\", \"episode_id\": \"xperience-10m-sample\", \"split\": \"train\", \"target\": \"episode_qa\", \"prompt_type\": \"json_episode_understanding\", \"center_window\": {\"start_frame\": 860, \"end_frame\": 879, \"num_frames\": 20}, \"media\": {\"video_paths\""
+    },
+    {
+      "classification": "historical_identifier_in_readiness_artifact",
+      "path": "results/omni_finetune/dataset.jsonl",
+      "line": 21,
+      "patterns": [
+        "qwen3_omni_32ep",
+        "xperience10m_qwen3_omni_32ep",
+        "ropedia-episode-task-suite"
+      ],
+      "example": "{\"id\": \"xperience-10m-sample:qa:44\", \"episode_id\": \"xperience-10m-sample\", \"split\": \"train\", \"target\": \"episode_qa\", \"prompt_type\": \"json_episode_understanding\", \"center_window\": {\"start_frame\": 880, \"end_frame\": 899, \"num_frames\": 20}, \"media\": {\"video_paths\""
     }
   ],
-  "historical_identifier_total_count": 159,
+  "historical_identifier_total_count": 157,
   "failures": []
 }

docs/data/source_alignment_audit.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Source Alignment Note",
   "status": "pass",
-  "generated_at_utc": "2026-06-04T16:03:23+00:00",
+  "generated_at_utc": "2026-06-04T16:33:01+00:00",
   "alignment_json": "docs/data/xperience10m_dataset_card_alignment.json",
   "alignment_summary": {
     "full_dataset_repo": "ropedia-ai/xperience-10m",

docs/data/summary_metrics.json

@@ -1,9 +1,9 @@
 {
   "omni_relay": {
-    "status": "selected_relay_in_progress",
+    "status": "selected_episode_preparation",
     "dataset": "ropedia-ai/xperience-10m",
-    "staging": "accelerated_chunked_parallel_transfer_with_batch_prefetch",
-    "training_target": "external_multi_gpu_training_host",
+    "staging": "selected_episode_preparation",
+    "training_target": "multi_episode_training_pipeline",
     "selection_strategy": "stratified_round_robin_by_top_level_session",
     "target_episodes": 128,
     "selected_sessions": 128,
@@ -13,7 +13,7 @@
     "exclude": [
       "visualization.rrd"
     ],
-    "access_status": "Full-dataset access is granted; selected multi-episode relay is in progress with chunked parallel transfer and overlapping batch prefetch.",
+    "access_status": "The gated Xperience-10M dataset is available for selected multi-episode pilot preparation.",
     "current_scope": "The selected-episode Qwen3-Omni fine-tune requires completed data staging and held-out evaluation; the 32-episode Qwen3-Omni fine-tune requires gated data staging before any real held-out metric is reported."
   },
   "models": {

docs/data/website_integrity.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-04T16:02:00+00:00",
+  "generated_at_utc": "2026-06-04T16:33:03+00:00",
   "docs_root": "docs",
   "site_base": "/ropedia-xperience-10m-task-suite/",
   "summary": {
@@ -74,8 +74,8 @@
       "name": "project_overview_precedes_progress_ledger",
       "status": "pass",
       "reason": "The project overview should appear before the deeper progress ledger.",
-      "overview_index": 67288,
-      "evidence_index": 84398
+      "overview_index": 67294,
+      "evidence_index": 84399
     },
     {
       "name": "project_status_links_json",
@@ -149,9 +149,9 @@
       "name": "evaluation_protocol_between_overview_and_progress",
       "status": "pass",
       "reason": "The evaluation protocol should appear before the deeper evidence ledger.",
-      "overview_index": 67288,
-      "protocol_index": 81257,
-      "evidence_index": 84398
+      "overview_index": 67294,
+      "protocol_index": 81258,
+      "evidence_index": 84399
     },
     {
       "name": "evaluation_protocol_links_json",
@@ -244,7 +244,7 @@
   "json_files": [
     {
       "path": "data/artifact_index.json",
-      "bytes": 32301,
+      "bytes": 32303,
       "top_level_type": "dict"
     },
     {
@@ -264,7 +264,7 @@
     },
     {
       "path": "data/evidence_contract.json",
-      "bytes": 12107,
+      "bytes": 12147,
       "top_level_type": "dict"
     },
     {
@@ -349,12 +349,12 @@
     },
     {
       "path": "data/research_roadmap.json",
-      "bytes": 5727,
+      "bytes": 5752,
       "top_level_type": "dict"
     },
     {
       "path": "data/research_roadmap_interactive.json",
-      "bytes": 131493,
+      "bytes": 131507,
       "top_level_type": "dict"
     },
     {
@@ -364,7 +364,7 @@
     },
     {
       "path": "data/scope_claims_audit.json",
-      "bytes": 18951,
+      "bytes": 19497,
       "top_level_type": "dict"
     },
     {
@@ -379,7 +379,7 @@
     },
     {
       "path": "data/summary_metrics.json",
-      "bytes": 25369,
+      "bytes": 25293,
       "top_level_type": "dict"
     },
     {

docs/index.html

@@ -511,7 +511,7 @@
     #diagnostics { order: 17; }
     #evidence { order: 18; }
     #artifacts { order: 19; }
-    #omni-relay { order: 20; }
+    #omni-scale-up { order: 20; }
     #run { order: 21; }
     #suite { padding: 62px 0 76px; }
     #suite .wrap { width: min(1680px, calc(100% - 48px)); }
@@ -2197,7 +2197,7 @@
           <strong>Directions</strong>
           <span>four tracks and probes</span>
         </button>
-        <button type="button" class="project-tab" id="tab-resources" role="tab" data-tab-key="resources" data-default-section="artifacts" aria-selected="false" aria-pressed="false" aria-controls="evidence artifacts omni-relay run" tabindex="-1">
+        <button type="button" class="project-tab" id="tab-resources" role="tab" data-tab-key="resources" data-default-section="artifacts" aria-selected="false" aria-pressed="false" aria-controls="evidence artifacts omni-scale-up run" tabindex="-1">
           <strong>Resources</strong>
           <span>research artifacts and scale-up</span>
         </button>
@@ -2234,7 +2234,7 @@
             </article>
             <article class="brief-card">
               <strong>What comes next</strong>
-              <p>The next model-quality stage is a held-out episode pilot over the selected multi-episode relay, with no train/test episode leakage and a completed omni-model evaluation report.</p>
+              <p>The next model-quality stage is a held-out episode pilot over selected multi-episode data, with no train/test episode leakage and a completed omni-model evaluation report.</p>
             </article>
           </div>
           <div class="brief-grid" aria-label="Research capability map">
@@ -3063,7 +3063,7 @@
             </button>
             <button type="button" class="content-tab" id="artifact-tab-scale-up" role="tab" data-panel-target="artifact-panel-scale-up" aria-selected="false" aria-pressed="false" aria-controls="artifact-panel-scale-up" tabindex="-1">
               <strong>Scale-Up</strong>
-              <span>relay and Omni path</span>
+              <span>Omni scale-up path</span>
             </button>
             <button type="button" class="content-tab" id="artifact-tab-checks" role="tab" data-panel-target="artifact-panel-checks" aria-selected="false" aria-pressed="false" aria-controls="artifact-panel-checks" tabindex="-1">
               <strong>Checks</strong>
@@ -3142,7 +3142,7 @@
       </div>
     </section>
 
-    <section id="omni-relay" data-project-tab="resources" role="tabpanel" aria-labelledby="tab-resources" tabindex="-1">
+    <section id="omni-scale-up" data-project-tab="resources" role="tabpanel" aria-labelledby="tab-resources" tabindex="-1">
       <div class="wrap">
         <div class="section-head">
           <h2>Qwen3-Omni pilot is in data preparation.</h2>
@@ -3257,7 +3257,7 @@ python scripts/validate_publication_package.py</code></pre>
       { id: "diagnostics", label: "Diagnostic Charts" },
       { id: "artifacts", label: "Research Artifacts" },
       { id: "evidence", label: "Research Progress" },
-      { id: "omni-relay", label: "Omni Scale-Up" },
+      { id: "omni-scale-up", label: "Omni Scale-Up" },
       { id: "run", label: "Reproduce" }
     ];
     const sectionLabels = Object.fromEntries(sectionConfig.map((section) => [section.id, section.label]));

index.html

@@ -511,7 +511,7 @@
     #diagnostics { order: 17; }
     #evidence { order: 18; }
     #artifacts { order: 19; }
-    #omni-relay { order: 20; }
+    #omni-scale-up { order: 20; }
     #run { order: 21; }
     #suite { padding: 62px 0 76px; }
     #suite .wrap { width: min(1680px, calc(100% - 48px)); }
@@ -2197,7 +2197,7 @@
           <strong>Directions</strong>
           <span>four tracks and probes</span>
         </button>
-        <button type="button" class="project-tab" id="tab-resources" role="tab" data-tab-key="resources" data-default-section="artifacts" aria-selected="false" aria-pressed="false" aria-controls="evidence artifacts omni-relay run" tabindex="-1">
+        <button type="button" class="project-tab" id="tab-resources" role="tab" data-tab-key="resources" data-default-section="artifacts" aria-selected="false" aria-pressed="false" aria-controls="evidence artifacts omni-scale-up run" tabindex="-1">
           <strong>Resources</strong>
           <span>research artifacts and scale-up</span>
         </button>
@@ -2234,7 +2234,7 @@
             </article>
             <article class="brief-card">
               <strong>What comes next</strong>
-              <p>The next model-quality stage is a held-out episode pilot over the selected multi-episode relay, with no train/test episode leakage and a completed omni-model evaluation report.</p>
+              <p>The next model-quality stage is a held-out episode pilot over selected multi-episode data, with no train/test episode leakage and a completed omni-model evaluation report.</p>
             </article>
           </div>
           <div class="brief-grid" aria-label="Research capability map">
@@ -3063,7 +3063,7 @@
             </button>
             <button type="button" class="content-tab" id="artifact-tab-scale-up" role="tab" data-panel-target="artifact-panel-scale-up" aria-selected="false" aria-pressed="false" aria-controls="artifact-panel-scale-up" tabindex="-1">
               <strong>Scale-Up</strong>
-              <span>relay and Omni path</span>
+              <span>Omni scale-up path</span>
             </button>
             <button type="button" class="content-tab" id="artifact-tab-checks" role="tab" data-panel-target="artifact-panel-checks" aria-selected="false" aria-pressed="false" aria-controls="artifact-panel-checks" tabindex="-1">
               <strong>Checks</strong>
@@ -3142,7 +3142,7 @@
       </div>
     </section>
 
-    <section id="omni-relay" data-project-tab="resources" role="tabpanel" aria-labelledby="tab-resources" tabindex="-1">
+    <section id="omni-scale-up" data-project-tab="resources" role="tabpanel" aria-labelledby="tab-resources" tabindex="-1">
       <div class="wrap">
         <div class="section-head">
           <h2>Qwen3-Omni pilot is in data preparation.</h2>
@@ -3257,7 +3257,7 @@ python scripts/validate_publication_package.py</code></pre>
       { id: "diagnostics", label: "Diagnostic Charts" },
       { id: "artifacts", label: "Research Artifacts" },
       { id: "evidence", label: "Research Progress" },
-      { id: "omni-relay", label: "Omni Scale-Up" },
+      { id: "omni-scale-up", label: "Omni Scale-Up" },
       { id: "run", label: "Reproduce" }
     ];
     const sectionLabels = Object.fromEntries(sectionConfig.map((section) => [section.id, section.label]));

results/omni_finetune/DATA_ACCESS_STATUS.md

@@ -11,16 +11,15 @@ held-out multi-episode experiment.
 | Current public local sample | 1 episode |
 | Full dataset access | Granted; metadata-only Hugging Face audit completed |
 | Current full-dataset metadata snapshot | 12,102 complete visible HF episodes across 802 complete sessions |
-| Current selected relay | 128 metadata-balanced episodes, 96/16/16 train/val/test |
-| Current staged multi-episode data | 128-episode relay started with chunked parallel transfer and batch prefetch; staging not complete yet |
+| Current selected pilot | 128 metadata-balanced episodes, 96/16/16 train/val/test |
+| Current multi-episode data state | Selected episodes are being prepared; preprocessing and held-out evaluation are not complete yet |
 | Current Qwen3-Omni artifacts | Setup-stage sample run, not held-out multi-episode model metrics |
 | Public raw-data redistribution | Not included |
 
-The current 128-episode relay is an operational data-staging step. It uses
-chunked parallel transfer for staged files and overlapping batch prefetch for
-future relay batches. It should not be described as a completed fine-tune or
-evaluated model until all selected episodes are staged, audited, preprocessed,
-trained, and evaluated on held-out sessions.
+The selected 128-episode pilot is the next model-quality milestone. It should
+not be described as a completed fine-tune or evaluated model until the selected
+episodes are available, checked for modality coverage, preprocessed into
+train/validation/test examples, trained, and evaluated on held-out sessions.
 
 ## Episode Requirement
 
@@ -56,7 +55,7 @@ training yet.
 - `results/omni_finetune/FULL_DATASET_METADATA_AUDIT.md`
 - `results/omni_finetune/full_dataset_metadata_audit.json`
 - `results/omni_finetune/XPERIENCE10M_128_EPISODE_SELECTION.md`
-- `results/omni_finetune/XPERIENCE10M_128_RELAY_AND_FINETUNE_PLAN.md`
+- `results/omni_finetune/XPERIENCE10M_128_DATA_PREPARATION_AND_FINETUNE_PLAN.md`
 - `results/omni_finetune/xperience10m_128_episode_selection.json`
 - `results/omni_finetune/xperience10m_128_episode_download_files.txt`
 - `results/omni_finetune/source_discovery.json`

results/omni_finetune/FULL_DATASET_METADATA_AUDIT.md

@@ -109,7 +109,7 @@ Metadata-only analysis of the gated Hugging Face dataset. No MP4, HDF5, RRD, or
 ## Download and Compute Recommendation
 
 - This metadata listing check can run on any machine with Hugging Face access.
-- If the training host cannot reach Hugging Face, download on an HF-reachable relay host, then transfer staged episode folders to the training host.
+- If the training host cannot reach Hugging Face, download on an HF-reachable machine, then transfer staged episode folders to the training host.
 - For training downloads, include `annotation.hdf5` plus the six MP4 streams; exclude `visualization.rrd` unless Rerun visualization is specifically needed.
 - For the first real training pilot, prefer 32 complete episodes from different top-level sessions and avoid selecting only the tiny outlier episodes.
 - The training host is used after staged data exists: manifest validation, preprocessing, LoRA training, and held-out evaluation.

results/omni_finetune/MULTI_EPISODE_ACCESS_STATUS.md

@@ -1,10 +1,9 @@
 # Multi-Episode Access Status
 
 Current status: access to the gated full `ropedia-ai/xperience-10m` dataset is
-granted, and a metadata-only Hugging Face audit has been completed. A
-128-episode metadata-balanced relay has started with chunked parallel transfer
-and overlapping batch prefetch, but the selected multi-episode data has not
-been fully staged, audited, trained, or evaluated yet.
+granted, and a metadata-only Hugging Face audit has been completed. A selected
+128-episode pilot has been defined, but the multi-episode fine-tune has not
+completed preprocessing, training, or held-out evaluation yet.
 
 This file records the public data-access status and pilot requirements. It does
 not include local-machine aliases, private paths, SSH hosts, or token locations.
@@ -18,7 +17,7 @@ not include local-machine aliases, private paths, SSH hosts, or token locations.
 | Strategy | stratified round-robin across top-level session UUIDs |
 | Metadata-audited visible complete episodes | 12,102 |
 | Metadata-audited complete sessions | 802 |
-| Current selected relay | 128 metadata-balanced episodes |
+| Current selected pilot | 128 metadata-balanced episodes |
 | Recommended split | 96 train / 16 val / 16 test |
 | Recommended estimated download | 277.71 GiB excluding `visualization.rrd` |
 | Representative 32-episode estimate | ~70.5 GiB at median episode size |
@@ -29,11 +28,9 @@ not include local-machine aliases, private paths, SSH hosts, or token locations.
 
 The current Qwen3-Omni artifacts come from the locally available sample data.
 The held-out model-quality run starts after selected complete episodes are
-downloaded, transferred through the relay path, validated locally, audited for
-content balance, and preprocessed into train/val/test examples. The current
-relay uses chunked parallel file transfer and overlapping prefetch to reduce
-staging idle time; this is a data-staging improvement, not a model-quality
-result.
+available, validated for modality coverage, reviewed for content balance, and
+preprocessed into train/validation/test examples. This is still a preparation
+stage, not a model-quality result.
 
 A real held-out pilot can be claimed only after:
 
@@ -55,7 +52,7 @@ The current 128-episode metadata-balanced download plan is:
 
 `results/omni_finetune/XPERIENCE10M_128_EPISODE_SELECTION.md`
 
-The older machine-generated source discovery blocker remains a pre-access local
-staging record:
+The older machine-generated source discovery report remains a pre-access local
+planning record:
 
 `results/omni_finetune/DATA_BLOCKER_REPORT.md`

results/omni_finetune/XPERIENCE10M_128_DATA_PREPARATION_AND_FINETUNE_PLAN.md

@@ -0,0 +1,205 @@
+# Xperience-10M 128-Episode Data Preparation and Fine-Tune Plan
+
+This is the executable plan for moving from metadata selection to real
+multi-episode training. It does not claim model-quality results until data is
+downloaded, staged, audited, trained, and evaluated on held-out sessions.
+
+## Current Preflight
+
+| Host | Role | Status |
+| --- | --- | --- |
+| HF-reachable download host | Dataset download and transfer | Needs Hugging Face access and enough scratch storage for one batch |
+| Training host | Persistent data + training | Needs enough storage for the staged selection and the training/eval stack |
+
+Conclusion: use a Hugging Face reachable machine for dataset downloads and the
+training machine as the persistent data store. The selected episodes are
+prepared in bounded batches, transferred, validated, and then used for the
+held-out Qwen3-Omni LoRA pilot.
+
+Current execution status:
+
+- a 128-episode data-preparation job has been launched on an HF-reachable host,
+- staged-file transfer is active,
+- later batches are scheduled after storage checks,
+- no multi-episode model-quality training result is claimed yet.
+
+## Selected Data
+
+- Selection file: `results/omni_finetune/xperience10m_128_episode_selection.json`
+- Download list: `results/omni_finetune/xperience10m_128_episode_download_files.txt`
+- Episodes: 128
+- Sessions: 128 unique sessions
+- Split: 96 train / 16 val / 16 test
+- Files: 896 training files
+- Excluded: `visualization.rrd`
+- Estimated training-host storage: 277.71 GiB excluding RRD
+
+## Transfer Setup
+
+Define host-specific paths outside the public artifact:
+
+```bash
+export RELAY_WORKDIR=/path/to/ropedia-episode-task-suite
+export RELAY_ROOT=/path/to/xperience10m_relay
+export TRAINING_HOST=<training-user>@<training-host>
+export TRAINING_REPO=/path/to/ropedia-episode-task-suite
+export TRAINING_DATA_ROOT=/path/to/xperience10m_128
+```
+
+Create a dedicated download-to-training SSH key:
+
+```bash
+ssh <relay-host> 'mkdir -p ~/.ssh && chmod 700 ~/.ssh && test -f ~/.ssh/xperience10m_relay_ed25519 || ssh-keygen -t ed25519 -N "" -f ~/.ssh/xperience10m_relay_ed25519 -C xperience10m-relay-to-training'
+ssh <relay-host> 'cat ~/.ssh/xperience10m_relay_ed25519.pub'
+```
+
+Append that public key to the training host `~/.ssh/authorized_keys`, then verify from the download host:
+
+```bash
+ssh <relay-host> 'ssh -i ~/.ssh/xperience10m_relay_ed25519 -o BatchMode=yes -o StrictHostKeyChecking=accept-new <training-user>@<training-host> hostname'
+```
+
+## Copy Minimal Repo Files to Download Host
+
+```bash
+ssh <relay-host> 'mkdir -p "$RELAY_WORKDIR"'
+rsync -av \
+  scripts/omni/relay_xperience10m_selection.py \
+  scripts/omni/parallel_chunk_transfer.py \
+  results/omni_finetune/xperience10m_128_episode_selection.json \
+  <relay-host>:"$RELAY_WORKDIR"/
+```
+
+## Transfer Dry Run
+
+```bash
+ssh <relay-host> '
+cd "$RELAY_WORKDIR" &&
+python3 relay_xperience10m_selection.py \
+  --selection-json xperience10m_128_episode_selection.json \
+  --relay-root "$RELAY_ROOT" \
+  --batch-max-gib 40 \
+  --batch-max-episodes 16 \
+  --transfer-host "$TRAINING_HOST" \
+  --transfer-root "$TRAINING_DATA_ROOT" \
+  --ssh-key ~/.ssh/xperience10m_relay_ed25519 \
+  --transfer-mode chunked \
+  --chunk-parallel 8 \
+  --chunk-size-mib 8 \
+  --chunk-threshold-mib 8 \
+  --delete-after-transfer \
+  --dry-run
+'
+```
+
+## Start Data Preparation
+
+Run in a persistent terminal or `tmux` session on the download host:
+
+```bash
+export HF_TOKEN=...
+cd "$RELAY_WORKDIR"
+python3 relay_xperience10m_selection.py \
+  --selection-json xperience10m_128_episode_selection.json \
+  --relay-root "$RELAY_ROOT" \
+  --batch-max-gib 40 \
+  --batch-max-episodes 16 \
+  --transfer-host "$TRAINING_HOST" \
+  --transfer-root "$TRAINING_DATA_ROOT" \
+  --ssh-key ~/.ssh/xperience10m_relay_ed25519 \
+  --transfer-mode chunked \
+  --chunk-parallel 8 \
+  --chunk-size-mib 8 \
+  --chunk-threshold-mib 8 \
+  --delete-after-transfer
+```
+
+Batch sizing is intentionally conservative. A 40 GiB batch size keeps restarts
+and partial-transfer cleanup cheaper than treating the full 277.71 GiB selection
+as one unit. Later batches should start after disk headroom is checked.
+
+## Training-Host Data Validation
+
+After transfer completes:
+
+```bash
+cd "$TRAINING_REPO"
+python3 scripts/omni/discover_xperience10m_sources.py \
+  --workspace "$TRAINING_REPO" \
+  --data-root "$TRAINING_DATA_ROOT" \
+  --output results/omni_finetune/source_discovery_128.json \
+  --report-output results/omni_finetune/DATA_BLOCKER_REPORT_128.md \
+  --target-episodes 128 \
+  --skip-modelscope \
+  --skip-huggingface
+```
+
+Then build the episode manifest:
+
+```bash
+python3 scripts/omni/build_episode_manifest.py \
+  --workspace "$TRAINING_REPO" \
+  --data-root "$TRAINING_DATA_ROOT" \
+  --max-episodes 128 \
+  --train-fraction 0.75 \
+  --val-fraction 0.125 \
+  --test-fraction 0.125 \
+  --split-seed 7 \
+  --output results/omni_finetune/episode_manifest_128.json
+```
+
+## Content Rebalance Gate
+
+Parse staged annotations before training:
+
+```bash
+python3 scripts/omni/audit_staged_xperience10m_content.py \
+  --data-root "$TRAINING_DATA_ROOT" \
+  --selection-json results/omni_finetune/xperience10m_128_episode_selection.json \
+  --output-json results/omni_finetune/staged_content_audit_128.json \
+  --output-csv results/omni_finetune/staged_content_audit_128.csv \
+  --report-output results/omni_finetune/STAGED_CONTENT_AUDIT_128.md
+```
+
+If a category dominates train, val, or test, swap episodes before training.
+
+## Training Order
+
+### 1. Qwen3-Omni LoRA Baseline
+
+Use this as the first real multi-episode SFT run because the repo already has
+working Qwen3-Omni training/eval scripts.
+
+Expected dataset:
+
+- 128 episodes
+- 32,768 max windows at 256 windows per episode
+- held-out sessions in val/test
+
+### 2. Cosmos3-Nano Compatibility
+
+Cosmos3-Nano should be treated as a second branch:
+
+- first run inference compatibility on a few staged clips,
+- then adapt data format for Cosmos video/action tasks,
+- then run post-training only after Qwen3-Omni and content audit pass.
+
+Good Cosmos tasks:
+
+- video + text -> physical reasoning text,
+- video + text -> future state/action label,
+- video + action/text -> future video,
+- video + text -> action trajectory proxy.
+
+Do not start with Cosmos3-Super. Cosmos3-Nano is the practical first target;
+Super is for a later run after data format, metrics, and compute are stable.
+
+## Acceptance Gates
+
+- 128 selected episodes staged on the training host.
+- No `visualization.rrd` in training data.
+- 128 unique sessions preserved.
+- Train/val/test session leakage is zero.
+- Content audit reviewed before training.
+- Qwen3-Omni eval runs on held-out sessions.
+- Cosmos3-Nano branch starts with compatibility, not immediate full fine-tune.

scripts/build_artifact_index.py

@@ -63,7 +63,7 @@ ARTIFACTS = [
         "path": "docs/data/research_roadmap.json",
         "kind": "project_path",
         "surface": "website_hf",
-        "shows": "Machine-readable staged roadmap for the website and Hugging Face mirrors.",
+        "shows": "Machine-readable research roadmap for the website and Hugging Face mirrors.",
     },
     {
         "id": "foundation_model_plan",

scripts/generate_visualizations.py

@@ -25,10 +25,10 @@ ASSETS = DOCS / "assets"
 CHARTS = ASSETS / "charts"
 
 OMNI_RELAY = {
-    "status": "selected_relay_in_progress",
+    "status": "selected_episode_preparation",
     "dataset": "ropedia-ai/xperience-10m",
-    "staging": "accelerated_chunked_parallel_transfer_with_batch_prefetch",
-    "training_target": "external_multi_gpu_training_host",
+    "staging": "selected_episode_preparation",
+    "training_target": "multi_episode_training_pipeline",
     "selection_strategy": "stratified_round_robin_by_top_level_session",
     "target_episodes": 128,
     "selected_sessions": 128,
@@ -36,7 +36,7 @@ OMNI_RELAY = {
     "valid_candidates": 12102,
     "estimated_bytes": 298188841943,
     "exclude": ["visualization.rrd"],
-    "access_status": "Full-dataset access is granted; selected multi-episode relay is in progress with chunked parallel transfer and overlapping batch prefetch.",
+    "access_status": "The gated Xperience-10M dataset is available for selected multi-episode pilot preparation.",
     "current_scope": "The selected-episode Qwen3-Omni fine-tune requires completed data staging and held-out evaluation; the 32-episode Qwen3-Omni fine-tune requires gated data staging before any real held-out metric is reported.",
 }