Add files using upload-large-folder tool

FIGURE_INDEX.md

@@ -19,7 +19,7 @@ Public figures, diagrams, charts, and derived modality thumbnails. Raw Xperience
 | Qwen3-Omni LoRA training pipeline | `docs/assets/qwen3_omni_lora_pipeline.png` | 1536 x 1024 | `docs/assets/qwen3_omni_lora_pipeline.prompt.md` | Detailed raw-data-to-adapter flow for staged Xperience-10M Qwen3-Omni LoRA training. |
 | Spatial intelligence slide diagram | `docs/assets/foundation-pipelines/spatial-intelligence-pipeline.png` | 2560 x 1920 | `scripts/render_foundation_pipeline_diagrams.py` | High-resolution slide diagram for the spatial intelligence pipeline track. |
 | Human-video world model slide diagram | `docs/assets/foundation-pipelines/human-video-world-model-pipeline.png` | 2560 x 1920 | `scripts/render_foundation_pipeline_diagrams.py` | High-resolution slide diagram for the human-video world-model pipeline track. |
-| Vision-language-action slide diagram | `docs/assets/foundation-pipelines/vision-language-action-pipeline.png` | 2560 x 1919 | `scripts/render_foundation_pipeline_diagrams.py` | High-resolution slide diagram for the VLA/action-policy pipeline track. |
+| Vision-language-action slide diagram | `docs/assets/foundation-pipelines/vision-language-action-pipeline.png` | 2560 x 1920 | `scripts/render_foundation_pipeline_diagrams.py` | High-resolution slide diagram for the VLA/action-policy pipeline track. |
 | Minimal and neural task architecture map | `docs/assets/task_architectures.png` | 1800 x 2450 | `scripts/render_overview_figures.py` | Minimal and neural heads for the original task contracts and shared feature contracts. |
 | Video modality thumbnail | `docs/assets/modalities/video.jpg` | 880 x 520 | `scripts/export_modality_atlas_assets.py` | Derived thumbnail for synchronized camera streams. |
 | Audio modality thumbnail | `docs/assets/modalities/audio.png` | 880 x 520 | `scripts/export_modality_atlas_assets.py` | Derived waveform thumbnail for the MP4 AAC stream. |

PROJECT_README.md

@@ -672,7 +672,7 @@ scripts/
   tier2_task_suite.py              # historical-name builder for tasks 13-20
   build_unified_task_suite.py       # builds TASK_SUITE_20.md and task_suite_20.json
   build_unified_task_model_radar.py # builds the unified 20-axis model comparison chart
-  build_task_method_20_gap_audit.py # builds the explicit 144/180 scored-cell gap ledger
+  build_task_method_20_gap_audit.py # builds the explicit 145/180 scored-cell gap ledger
   task_walkthroughs.py              # human-readable task-card and walkthrough-storyboard metadata
   generate_visualizations.py        # refreshes SVG charts + summary JSON
   render_task_suite_infographic.py  # renders the task-suite presentation PNG
@@ -718,7 +718,7 @@ docs/
   data/single_episode_task_model_radar.json # 1-episode split radar values
   data/episode128_task_model_radar.json # 128-episode split radar values
   data/task_method_20_result_matrix.json # 9-method x 20-task result matrix
-  data/task_method_20_gap_audit.json # explicit 144/180 scored-cell gap ledger
+  data/task_method_20_gap_audit.json # explicit 145/180 scored-cell gap ledger
   data/evidence_contract.json       # machine-readable project scope
   data/artifact_index.json          # compact project-artifact catalog
   data/live_publication_status.json # live GitHub/HF publication verification
@@ -1172,11 +1172,13 @@ so the public claims stay precise:
 
 High-resolution slide diagrams for the three tracks are published in
 [`docs/assets/foundation-pipelines`](docs/assets/foundation-pipelines). Spatial
-intelligence and human-video world modeling now use the clean slide PNGs, while
-the VLA card stays on the earlier restored slide source until a matching clean
-VLA slide PNG is supplied. The 2026-06-18 image refresh verified that the newly
-uploaded Spatial and Human-video PNGs are already the committed clean sources;
-the third uploaded file duplicated the Spatial slide. These images are
+intelligence and human-video world modeling use the clean slide PNGs supplied
+for publication and are exported as 2560-pixel public images. The 2026-06-19
+refresh verified that the latest uploaded Spatial and Human-video PNGs are
+byte-identical to the committed clean source cache. The VLA card is now a
+clean 2560-pixel deterministic slide redraw from the original presentation-photo
+source, because the third uploaded clean PNG duplicated the Spatial slide
+instead of providing a VLA export. These images are
 communication assets, not completed model-quality evidence; the exact task,
 training, and evaluation contracts remain in the Markdown and JSON files.
 

THREE_FOUNDATION_PIPELINES.md

@@ -14,16 +14,16 @@ inertial signals, object/contact annotations, and language captions.
 | Human-video world models | Can the model predict what happens next? | Observed video/audio/sensor windows, hand/body motion, object/contact state, action/subtask labels, future windows. | Future-state and future-action probes over the existing split, then Cosmos-style or latent world-model training with separate dynamics metrics. | Partially evidenced through current future-task probes and Cosmos-style branch artifacts; still needs stronger visual/latent future metrics. |
 | Vision-language-action models | Can the model turn what it sees and reads into action? | Egocentric video, language captions, hand/body motion, contacts, objects, procedure/subtask labels. | Observation-language-to-action target conversion, action-chunk scoring, policy-token baselines, then VLA/policy model fine-tuning. | Feasible but gated by action-target conversion; do not claim policy quality until action tokens, normalization, and held-out policy metrics exist. |
 
-## Published Direction Photos
+## Published Direction Figures
 
 The repo and public mirrors include three high-resolution direction images from
 the original direction slides. Spatial intelligence and human-video world
 modeling use the clean high-resolution slide PNGs supplied for publication and
 are exported as 2560-pixel public assets. The 2026-06-19 refresh verified the
 latest two clean PNGs as byte-identical to the committed source-slide cache. The
-VLA card stays on the restored original presentation-photo source because the
-third uploaded image duplicates the Spatial intelligence slide. They are
-communication assets, not
+VLA card is a clean deterministic redraw from the original presentation-photo
+source because the third uploaded image duplicates the Spatial intelligence
+slide. They are communication assets, not
 evidence of completed model-quality training. The exact technical scope remains
 the text and JSON contract in this document and
 `docs/data/three_foundation_pipelines.json`.
@@ -35,9 +35,9 @@ the text and JSON contract in this document and
 | Vision-language-action models | `docs/assets/foundation-pipelines/vision-language-action-pipeline.png` | `docs/assets/foundation-pipelines/source-photos/vision-language-action-source.jpg` |
 
 The deterministic restoration script is
-`scripts/render_foundation_pipeline_diagrams.py`; it remains available for the
-photo-derived VLA card, while the two clean slide PNGs are used directly after
-public-resolution packaging.
+`scripts/render_foundation_pipeline_diagrams.py`; it uses the two clean slide
+PNGs directly and renders the VLA card from the presentation-photo content so
+the three public direction figures have a consistent slide style.
 
 ## 1. Spatial Intelligence Pipeline
 

assets/foundation-pipelines/README.md

@@ -8,8 +8,9 @@ diagrams. They are used for the pipeline tracks documented in
 They replace the earlier concept-art images and keep the public visuals tied to
 the original direction slides. Spatial intelligence and human-video world
 modeling use the clean slide PNGs supplied for publication and are exported as
-2560-pixel public assets; VLA stays on the restored original presentation-photo
-source until a matching clean VLA slide PNG is supplied.
+2560-pixel public assets. VLA is now a clean 2560-pixel deterministic redraw
+from the original presentation-photo source because the latest third clean PNG
+duplicated the Spatial slide.
 They are still **pipeline communication assets**, not evidence of completed
 foundation-model quality. Exact technical claims live in the surrounding
 Markdown, JSON, and website labels.

docs/assets/foundation-pipelines/README.md

@@ -8,8 +8,9 @@ diagrams. They are used for the pipeline tracks documented in
 They replace the earlier concept-art images and keep the public visuals tied to
 the original direction slides. Spatial intelligence and human-video world
 modeling use the clean slide PNGs supplied for publication and are exported as
-2560-pixel public assets; VLA stays on the restored original presentation-photo
-source until a matching clean VLA slide PNG is supplied.
+2560-pixel public assets. VLA is now a clean 2560-pixel deterministic redraw
+from the original presentation-photo source because the latest third clean PNG
+duplicated the Spatial slide.
 They are still **pipeline communication assets**, not evidence of completed
 foundation-model quality. Exact technical claims live in the surrounding
 Markdown, JSON, and website labels.

docs/assets/foundation-pipelines/prompts.md

@@ -8,9 +8,9 @@ manifests and mirrors already link here as the provenance note.
 Update on 2026-06-19: the latest supplied clean Spatial intelligence and
 Human-video world model PNGs are byte-identical to the committed source-slide
 cache and are published as 2560-pixel public images. The third uploaded image
-duplicates the Spatial intelligence PNG, so the Vision-language-action card
-continues to use the restored original presentation photo until a clean VLA
-slide PNG is supplied.
+duplicates the Spatial intelligence PNG, so the Vision-language-action card is
+published as a clean deterministic slide redraw from the original VLA
+presentation-photo content.
 
 | Track | Source | Enhanced public PNG |
 | --- | --- | --- |
@@ -25,8 +25,11 @@ Restoration is deterministic and local:
 - Autocontrast and moderate brightness/color/contrast correction.
 - Lanczos resize to a 2560-pixel public width.
 - Gentle sharpening and unsharp masking.
+- For VLA only, a deterministic clean slide redraw preserves the visible
+  presentation content from the source photo while matching the clean black and
+  lime public-slide style.
 
-The restoration script deliberately does not synthesize, redraw, or hallucinate
-slide text. Technical task/training/evaluation claims are maintained in
+The restoration script deliberately avoids hallucinated model claims or
+non-source concept art. Technical task/training/evaluation claims are maintained in
 `THREE_FOUNDATION_PIPELINES.md` and
 `docs/data/three_foundation_pipelines.json`.

docs/data/artifact_index.json

@@ -1,6 +1,6 @@
 {
   "title": "Ropedia Xperience-10M Task Suite Artifact Index",
-  "generated_at_utc": "2026-06-18T16:39:29+00:00",
+  "generated_at_utc": "2026-06-18T17:18:54+00:00",
   "status": "pass",
   "artifact_count": 213,
   "missing": [],
@@ -136,8 +136,8 @@
       "surface": "repo_hf",
       "shows": "Frames spatial intelligence, human-video world modeling, and vision-language-action as three pipeline tracks with explicit inputs, outputs, maturity, and next evidence gates.",
       "exists": true,
-      "bytes": 8179,
-      "sha256": "ed390388e996fbe9b1b663cb98420ea658a59930b453aa7215859e9a05c2d0ef"
+      "bytes": 8234,
+      "sha256": "bad51748b2f9461d1246730131ee348714cfef146aae7080348f6d6b8d72ab22"
     },
     {
       "id": "three_foundation_pipelines_json",
@@ -147,8 +147,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable pipeline-track contract for the website and Hugging Face mirrors.",
       "exists": true,
-      "bytes": 10531,
-      "sha256": "30a261a912b172fee7433d81e2b36390554afa8f2bed009d5c126f3f2583b077"
+      "bytes": 10520,
+      "sha256": "be07a1a44100047181636efe4fce53248d61df6ec0b76f341652032e45ef252e"
     },
     {
       "id": "spatial_intelligence_slide_diagram",
@@ -180,8 +180,8 @@
       "surface": "website_hf",
       "shows": "High-resolution slide diagram for the vision-language-action training pipeline direction.",
       "exists": true,
-      "bytes": 2911014,
-      "sha256": "d4704ee28f747067c440845905cf2cacf6cbbf3fd5d17418ba16993f617ade29"
+      "bytes": 156793,
+      "sha256": "43f2c9f741319abb64d652aa3cf52d35d740b6890efc7be099c701ebd01b4018"
     },
     {
       "id": "spatial_intelligence_source_slide",
@@ -1017,7 +1017,7 @@
       "shows": "Catalogs public figures, charts, modality thumbnails, dimensions, hashes, roles, and source scripts.",
       "exists": true,
       "bytes": 6983,
-      "sha256": "84c4040a5a4dadeae164599fc937814cbb9d6a129593bd7c1bbd0614e36fa993"
+      "sha256": "48ea04c063df0745f2a31483d15baa71d420906b2ad7ce15fdb10760f41907e6"
     },
     {
       "id": "figure_index_json",
@@ -1027,8 +1027,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable visual asset index for website and Hugging Face mirrors.",
       "exists": true,
-      "bytes": 19441,
-      "sha256": "9b872e3f7e6c62f93cb3e38748f5918005a9b3f6ef3033eddbab9b7afa980d4a"
+      "bytes": 19440,
+      "sha256": "f9616f6254547e0483b43d498fb4256546e4d5f844f5934934ba78a8adba70bb"
     },
     {
       "id": "figure_index_builder",
@@ -1310,7 +1310,7 @@
       "volatile": true,
       "shows": "Confirms prepared GitHub/HF Space/artifact/model mirrors share the same critical data, figure, website HTML, and validator files.",
       "exists": true,
-      "bytes": 922005,
+      "bytes": 923179,
       "hash_policy": "existence_and_size_only"
     },
     {

docs/data/figure_index.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Figure Index",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T16:36:15+00:00",
+  "generated_at_utc": "2026-06-18T17:18:54+00:00",
   "scope": "Public figures, diagrams, charts, and derived modality thumbnails. Raw Xperience-10M videos, annotations, RRD files, and Qwen weights are excluded.",
   "figure_count": 29,
   "figures": [
@@ -149,12 +149,12 @@
       "source_script": "scripts/render_foundation_pipeline_diagrams.py",
       "surface": "README, website, HF Space, artifact dataset, model card",
       "exists": true,
-      "bytes": 2911014,
-      "sha256": "d4704ee28f747067c440845905cf2cacf6cbbf3fd5d17418ba16993f617ade29",
+      "bytes": 156793,
+      "sha256": "43f2c9f741319abb64d652aa3cf52d35d740b6890efc7be099c701ebd01b4018",
       "dimensions": {
         "format": "PNG",
         "width": 2560,
-        "height": 1919
+        "height": 1920
       },
       "source_script_exists": true
     },

docs/data/public_surface_qa.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Public Project Surface",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T16:39:28+00:00",
+  "generated_at_utc": "2026-06-18T17:18:53+00:00",
   "scope": "Repo README, GitHub Pages HTML, Hugging Face Space card, artifact dataset card, and model card.",
   "checks": [
     {
@@ -43,12 +43,12 @@
         "publication_package": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T16:31:49+00:00"
+          "generated_at_utc": "2026-06-18T16:39:43+00:00"
         },
         "mirror_parity": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T15:39:07+00:00"
+          "generated_at_utc": "2026-06-18T16:41:09+00:00"
         }
       },
       "failures": {}

docs/data/publication_audit.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-18T16:39:43+00:00",
+  "generated_at_utc": "2026-06-18T17:19:27+00:00",
   "checks": [
     {
       "name": "required_publication_assets_present",
@@ -237,8 +237,8 @@
     "hf_artifact_bundle": {
       "root": "hf_publish/artifacts",
       "exists": true,
-      "file_count": 2673,
-      "text_file_count": 1146,
+      "file_count": 2674,
+      "text_file_count": 1147,
       "largest_file": {
         "path": "results/omni_finetune/xperience10m_128ep_dense_multiscale_hierarchical_v1_20260608/dense_multiscale_windows.jsonl",
         "bytes": 135591061
@@ -248,8 +248,8 @@
     "hf_model_bundle": {
       "root": "hf_publish/model",
       "exists": true,
-      "file_count": 3146,
-      "text_file_count": 1314,
+      "file_count": 3147,
+      "text_file_count": 1315,
       "largest_file": {
         "path": "results/omni_finetune/xperience10m_128ep_dense_multiscale_hierarchical_v1_20260608/dense_multiscale_windows.jsonl",
         "bytes": 135591061

docs/data/three_foundation_pipelines.json

@@ -6,13 +6,13 @@
   "diagram_assets": {
     "status": "published_high_resolution_slide_diagrams",
     "asset_root": "docs/assets/foundation-pipelines",
-    "source": "Clean direction-slide PNGs where supplied, plus restored original presentation photos where a clean slide PNG is not yet available",
+    "source": "Clean direction-slide PNGs where supplied, plus a deterministic VLA slide redraw from the original presentation-photo source",
     "source_slide_root": "docs/assets/foundation-pipelines/source-slides",
     "source_photo_root": "docs/assets/foundation-pipelines/source-photos",
     "provenance_file": "docs/assets/foundation-pipelines/prompts.md",
     "renderer_script": "scripts/render_foundation_pipeline_diagrams.py",
     "diagram_type": "direction_slide_diagram",
-    "source_update": "2026-06-19: the latest supplied clean Spatial intelligence and Human-video world model PNGs are byte-identical to the committed source-slide cache and are published as 2560-pixel public images. The third uploaded file duplicates the Spatial intelligence slide, so Vision-language-action intentionally remains the restored original presentation-photo source until a clean VLA slide PNG is supplied.",
+    "source_update": "2026-06-19: the latest supplied clean Spatial intelligence and Human-video world model PNGs are byte-identical to the committed source-slide cache and are published as 2560-pixel public images. The third uploaded file duplicates the Spatial intelligence slide, so Vision-language-action is published as a clean 2560-pixel deterministic redraw from the original VLA presentation-photo content.",
     "note": "Images are slide-diagram communication assets for pipeline tracks. Technical claims remain governed by the Markdown/JSON contracts and verified metrics."
   },
   "shared_principles": [

docs/data/website_integrity.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-18T16:38:48+00:00",
+  "generated_at_utc": "2026-06-18T17:19:13+00:00",
   "docs_root": "docs",
   "site_base": "/ropedia-xperience-10m-task-suite/",
   "summary": {
@@ -81,7 +81,7 @@
       "status": "pass",
       "reason": "The project overview should appear before the deeper progress ledger.",
       "overview_index": 95702,
-      "evidence_index": 132075
+      "evidence_index": 132076
     },
     {
       "name": "project_status_links_json",
@@ -160,8 +160,8 @@
       "status": "pass",
       "reason": "The evaluation protocol should appear before the deeper evidence ledger.",
       "overview_index": 95702,
-      "protocol_index": 128256,
-      "evidence_index": 132075
+      "protocol_index": 128257,
+      "evidence_index": 132076
     },
     {
       "name": "evaluation_protocol_links_json",
@@ -301,7 +301,7 @@
     },
     {
       "path": "data/artifact_index.json",
-      "bytes": 116109,
+      "bytes": 116108,
       "top_level_type": "dict"
     },
     {
@@ -331,7 +331,7 @@
     },
     {
       "path": "data/figure_index.json",
-      "bytes": 19441,
+      "bytes": 19440,
       "top_level_type": "dict"
     },
     {
@@ -351,7 +351,7 @@
     },
     {
       "path": "data/mirror_parity.json",
-      "bytes": 922005,
+      "bytes": 923179,
       "top_level_type": "dict"
     },
     {
@@ -516,7 +516,7 @@
     },
     {
       "path": "data/three_foundation_pipelines.json",
-      "bytes": 10531,
+      "bytes": 10520,
       "top_level_type": "dict"
     },
     {
@@ -664,9 +664,9 @@
     {
       "path": "assets/foundation-pipelines/vision-language-action-pipeline.png",
       "exists": true,
-      "bytes": 2911014,
+      "bytes": 156793,
       "width": 2560,
-      "height": 1919,
+      "height": 1920,
       "format": "PNG"
     },
     {

docs/index.html

@@ -3608,7 +3608,7 @@
         </div>
         <div class="foundation-pipeline-grid" aria-label="Three high-resolution foundation direction slide diagrams">
           <article class="foundation-pipeline-card">
-            <img src="assets/foundation-pipelines/spatial-intelligence-pipeline.png?v=foundation-slides-v7" alt="High-resolution slide diagram showing the Spatial intelligence models direction for Xperience-10M.">
+            <img src="assets/foundation-pipelines/spatial-intelligence-pipeline.png?v=foundation-slides-v8" alt="High-resolution slide diagram showing the Spatial intelligence models direction for Xperience-10M.">
             <div class="foundation-pipeline-body">
               <span>High-resolution direction slide</span>
               <h3>Spatial intelligence models</h3>
@@ -3620,7 +3620,7 @@
             </div>
           </article>
           <article class="foundation-pipeline-card">
-            <img src="assets/foundation-pipelines/human-video-world-model-pipeline.png?v=foundation-slides-v7" alt="High-resolution slide diagram showing the Human-video world models direction for Xperience-10M.">
+            <img src="assets/foundation-pipelines/human-video-world-model-pipeline.png?v=foundation-slides-v8" alt="High-resolution slide diagram showing the Human-video world models direction for Xperience-10M.">
             <div class="foundation-pipeline-body">
               <span>High-resolution direction slide</span>
               <h3>Human-video world models</h3>
@@ -3632,9 +3632,9 @@
             </div>
           </article>
           <article class="foundation-pipeline-card">
-            <img src="assets/foundation-pipelines/vision-language-action-pipeline.png?v=foundation-slides-v7" alt="High-resolution slide diagram showing the Vision-language-action models direction for Xperience-10M.">
+            <img src="assets/foundation-pipelines/vision-language-action-pipeline.png?v=foundation-slides-v8" alt="High-resolution slide diagram showing the Vision-language-action models direction for Xperience-10M.">
             <div class="foundation-pipeline-body">
-              <span>Restored direction slide</span>
+              <span>Clean VLA direction slide</span>
               <h3>Vision-language-action models</h3>
               <p>Train VLA or policy-compatible heads only after converting egocentric video, captions, hand/body motion, contacts, objects, and procedures into traceable action tokens, chunks, and object-conditioned action targets.</p>
               <div class="foundation-pipeline-links">

index.html

@@ -3608,7 +3608,7 @@
         </div>
         <div class="foundation-pipeline-grid" aria-label="Three high-resolution foundation direction slide diagrams">
           <article class="foundation-pipeline-card">
-            <img src="assets/foundation-pipelines/spatial-intelligence-pipeline.png?v=foundation-slides-v7" alt="High-resolution slide diagram showing the Spatial intelligence models direction for Xperience-10M.">
+            <img src="assets/foundation-pipelines/spatial-intelligence-pipeline.png?v=foundation-slides-v8" alt="High-resolution slide diagram showing the Spatial intelligence models direction for Xperience-10M.">
             <div class="foundation-pipeline-body">
               <span>High-resolution direction slide</span>
               <h3>Spatial intelligence models</h3>
@@ -3620,7 +3620,7 @@
             </div>
           </article>
           <article class="foundation-pipeline-card">
-            <img src="assets/foundation-pipelines/human-video-world-model-pipeline.png?v=foundation-slides-v7" alt="High-resolution slide diagram showing the Human-video world models direction for Xperience-10M.">
+            <img src="assets/foundation-pipelines/human-video-world-model-pipeline.png?v=foundation-slides-v8" alt="High-resolution slide diagram showing the Human-video world models direction for Xperience-10M.">
             <div class="foundation-pipeline-body">
               <span>High-resolution direction slide</span>
               <h3>Human-video world models</h3>
@@ -3632,9 +3632,9 @@
             </div>
           </article>
           <article class="foundation-pipeline-card">
-            <img src="assets/foundation-pipelines/vision-language-action-pipeline.png?v=foundation-slides-v7" alt="High-resolution slide diagram showing the Vision-language-action models direction for Xperience-10M.">
+            <img src="assets/foundation-pipelines/vision-language-action-pipeline.png?v=foundation-slides-v8" alt="High-resolution slide diagram showing the Vision-language-action models direction for Xperience-10M.">
             <div class="foundation-pipeline-body">
-              <span>Restored direction slide</span>
+              <span>Clean VLA direction slide</span>
               <h3>Vision-language-action models</h3>
               <p>Train VLA or policy-compatible heads only after converting egocentric video, captions, hand/body motion, contacts, objects, and procedures into traceable action tokens, chunks, and object-conditioned action targets.</p>
               <div class="foundation-pipeline-links">

results/omni_finetune/model_output_task_probes_20260616/action_object_relation/cosmos3_super_reasoner/metrics.json

@@ -8,7 +8,7 @@
     "predictions_csv": "results/omni_finetune/model_output_task_probes_20260616/action_object_relation/cosmos3_super_reasoner/predictions.csv"
   },
   "excluded_rows_without_true_relation": 2,
-  "generated_at_utc": "2026-06-18T16:30:30+00:00",
+  "generated_at_utc": "2026-06-18T15:25:22+00:00",
   "labels": [
     "Adjust canned food on shelf :: canned food | cardboard box | store shelf",
     "Adjust item on shelf :: shelf | stationery package",

results/omni_finetune/model_output_task_probes_20260616/action_object_relation/qwen3_omni_v6_lora/metrics.json

@@ -8,7 +8,7 @@
     "predictions_csv": "results/omni_finetune/model_output_task_probes_20260616/action_object_relation/qwen3_omni_v6_lora/predictions.csv"
   },
   "excluded_rows_without_true_relation": 18,
-  "generated_at_utc": "2026-06-18T16:30:30+00:00",
+  "generated_at_utc": "2026-06-18T15:25:21+00:00",
   "labels": [
     "Adjust canned food on shelf :: canned food | cardboard box | store shelf",
     "Adjust item on shelf :: hand | packaged item | shelf",

results/omni_finetune/model_output_task_probes_20260616/long_horizon_next_action/cosmos3_nano_future_window/metrics.json

@@ -7,7 +7,7 @@
   },
   "dataset_contract": "xperience10m_future_window_world_model_v0",
   "excluded_rows_without_true_action": 0,
-  "generated_at_utc": "2026-06-18T16:30:30+00:00",
+  "generated_at_utc": "2026-06-18T15:25:22+00:00",
   "horizon_windows": 5,
   "known_limitation": "The verified package records horizon_windows rather than a raw wall-clock horizon. This score should be read as the Cosmos-Nano future-window branch for task 13, not as independent proof of an exact five-second raw-video target.",
   "labels": [

results/omni_finetune/model_output_task_probes_20260616/time_to_transition/cosmos3_super_reasoner/metrics.json

@@ -5,7 +5,7 @@
   },
   "cap_frames": 200,
   "excluded_rows_without_true_action": 0,
-  "generated_at_utc": "2026-06-18T16:30:30+00:00",
+  "generated_at_utc": "2026-06-18T15:25:22+00:00",
   "known_limitation": "This is a derived action-sequence probe, not evidence of a separately trained time-regression head. It is included because task 20's target is deterministically derivable from a sequence of action labels.",
   "metric_direction": "lower",
   "metric_key": "time_to_transition_mae",