Update README.md

README.md

@@ -10,25 +10,15 @@ pinned: false
 ---
 # Shot Grammar Adapter — Proof (HF Space)
 
-This Space demonstrates that shot grammar JSON can drive transformer-based video generation.
+# Shot Grammar Adapter — Proof (Public Demo)
 
-- Left: pick a shot_id → view features + auto-prompt
-- Right: generated GIF/MP4
-- Batch mode: build a Proof Reel ZIP (clips + metadata.json)
+- Left: pick a `shot_id` → see features + auto-generated prompt  
+- Right: transformer-based video generation (GIF/MP4)  
+- Batch: build a small Proof Reel (ZIP)
 
-Model: set `VIDEO_MODEL_ID` (Space secret) e.g. damo-vilab/text-to-video-ms-1.7b
-Data: `shots_proof.jsonl`
+**Data:** 20–30-shot public subset (`shots_public_subset.jsonl`)  
+**Note:** This is a demo subset. © JsisOn — Redistribution or reverse engineering prohibited.  
+For partnerships or research access to extended data, contact: princess@batalstone.com
 
-Title: Shot Grammar Adapter — Demo Subset (20–30 shots)
-
-License: Research-only, no redistribution, no commercial use (custom EULA inside)
-
-Description: “Camera grammar (size/angle/motion/relation/duration) → transformer video control proof.”
-
-Data Fields: shot_id (anonymized), features{size,angle,motion,relation,duration}, prompt
-
-Intended Use: Evaluation / Non-commercial research; Not for training commercial models.
-
-Caveats: Duration rounded; IDs anonymized; subset only.
 
 Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference