Update README.md

README.md

@@ -19,7 +19,13 @@ metrics:
 
 # SwipeALot Base Model
 
-Multimodal transformer for swipe keyboard prediction. Trained on the [futo-org/swipe.futo.org](https://huggingface.co/datasets/futo-org/swipe.futo.org) dataset.
+Multimodal, multi-objective transformer for swipe keyboard prediction. Trained on the [futo-org/swipe.futo.org](https://huggingface.co/datasets/futo-org/swipe.futo.org) dataset.
+
+This model is trained with the following objectives:
+- Masked character prediction
+- Masked path prediction
+- Text length prediction (CLS token)
+- Path/text embedding (SEP token)
 
 <p align="center">
   <img src="https://cdn-uploads.huggingface.co/production/uploads/65ff92ea467d83751a727538/OV87xy-_ID0TqKW0bfvVq.png" style="width: 400px;">
@@ -78,9 +84,7 @@ print(f"Predicted word length: {predicted_length}")
 - Path coordinates must be normalized to [0, 1] range for x, y
 - Timestamps must be normalized to [0, 1] range
 - Paths longer than 128 points are downsampled using linear interpolation
-- Paths shorter than 128 points are zero-padded
 - Text longer than 48 characters is truncated with EOS preserved
-- Text shorter than 48 characters is zero-padded
 
 ## Capabilities