Update README.md

README.md

@@ -29,10 +29,6 @@ The default inference command predicts the 3D conditions itself and generates 33
 
 The standard renderer-training recipe also generates its conditions with the frozen DeXplicit 3D model. It uses recorded future RGB only as supervision and rejects measured or unlabelled conditioning bundles.
 
-## Existing weight history
-
-The existing DeXplicit renderer weights are the evaluated high/low adapter pair at step 3,264. They were trained and originally evaluated using measured future 3D controls. The updated predicted-3D training code does not change that checkpoint history, and the original scores are not claimed for predicted conditioning. Fine-tuning and evaluating a new checkpoint is a separate experiment. The optional compact variant has its own adapter and conditioning interface.
-
 ```bash
 python -m dexplicit.download --output-dir weights
 python -m dexplicit.infer --help