Update README.md

README.md

@@ -13,18 +13,17 @@ This repository contains the model weights for CausalPFN, a transformer-based in
 
 CausalPFN is a pre-trained model for amortized causal effect estimation via in-context learning. It allows for accurate estimation of conditional average treatment effects (CATE) and average treatment effects (ATE) without requiring model retraining for each new dataset.
 
-The model is based on a transformer architecture with:
-- Long-context in-context learning
-- Uncertainty quantification and calibration
+The model is based on a transformer architecture with uncertainty quantification and calibration.
 
 ## Requirements
 
 - Python ≥ 3.10
-- torch ≥ 2.5.1
-- numpy ≥ 1.26.4
-- tqdm ≥ 4.67.1
+- torch ≥ 2.0
+- numpy ≥ 1.24
+- tqdm
 - faiss-cpu ≥ 1.9.0
 - scikit-learn ≥ 1.5.2
+- huggingface_hub
 
 ## Installation