Update README.md

README.md

@@ -18,21 +18,6 @@ pipeline_tag: image-classification
 [![NeurIPS](https://img.shields.io/badge/NeurIPS_2025-1st_Place_CUMSUM-4b44ce.svg)](https://neurips.cc/virtual/2025/loc/san-diego/135896)
 [![GitHub](https://img.shields.io/badge/GitHub-satformer-181717.svg?logo=github)](https://github.com/leharris3/satformer)
 
-SaTformer is a Vision Transformer adapted for spatio-temporal precipitation nowcasting from geostationary satellite (HRIT) imagery. It won **1st place** in the NeurIPS 2025 CUMSUM challenge.
-
-## Model Details
-
-| Parameter | Value |
-|---|---|
-| Architecture | Vision Transformer (adapted from TimeSformer) |
-| Attention | Joint space-time (ST²) |
-| Embedding dim | 512 |
-| Depth | 12 blocks |
-| Heads | 8 (dim 64) |
-| Input | 4 frames x 11 channels x 32x32 |
-| Output | 64 precipitation bins (classification) |
-| Patch size | 4x4 |
-
 ## Usage
 
 ```python