Update README.md

README.md

@@ -78,30 +78,9 @@ as JPEG.
 
 ## Composition
 
-Trajectory groups:
 
-| Trajectory group | Pairs | Share |
-|---|---:|---:|
-| `pure_translation` | 292 | 38.4% |
-| `pure_rotation` | 186 | 24.5% |
-| `compound` | 282 | 37.1% |
+![Clipboard_Screenshot_1787922858](https://cdn-uploads.huggingface.co/production/uploads/654c6845bac6e6e49895a5b5/W4AMIdkJPdbn0OfOS3Yvy.png)
 
-Visual styles:
-
-| Style | Pairs | Share |
-|---|---:|---:|
-| `photo` | 365 | 48.0% |
-| `game_anime` | 334 | 43.9% |
-| `traditional` | 61 | 8.0% |
-
-Videos come from 9 generation systems compared pairwise. Ground-truth
-verdicts are close to balanced between the two positions.
-
-| Axis | left | right | tie |
-|---|---:|---:|---:|
-| `action` | 365 | 370 | 25 |
-| `appearance` | 367 | 371 | 22 |
-| `motion` | 358 | 363 | 39 |
 
 ## Results
 
@@ -131,28 +110,8 @@ axis a predictor does not model.
 
 
 ## Usage
-See https://github.com/CodeGoat24/WorldReward for the full evaluation protocol, including how chunk-level
-judgements aggregate to a pair-level verdict.
-
-```bash
-pip install worldreward
-
-python scripts/download_bench.py --output-dir data/WorldReward-Bench
-python scripts/preprocess.py \
-    --pairs data/WorldReward-Bench/bench.jsonl \
-    --output-root outputs/rendered_chunks
-python scripts/run_inference.py \
-    --render-root outputs/rendered_chunks \
-    --output outputs/chunk_predictions.json \
-    --model CodeGoat24/WorldReward-9B
-python worldreward-bench/score.py \
-    --bench data/WorldReward-Bench/bench.jsonl \
-    --predictions outputs/chunk_predictions.json
-```
-
-## Licence
+See https://github.com/CodeGoat24/WorldReward for the full evaluation protocol.
 
-Apache 2.0.
 
 ## Citation