Update README.md

README.md

@@ -52,16 +52,14 @@ print(embedding.shape)  # (1, hidden_dim)
 
 ## 🧩 Model Overview
 
-| **Property** | **Description** |
-|:-------------|:----------------|
-| **Architecture** | Sentence-BERT (`all-MiniLM-L6-v2` backbone) |
-| **Pooling Strategy** | Weighted mean aggregation over transcript chunks |
-| **Max Tokens per Chunk** | 512 |
-| **Input Representation** | Transcript + talk title + publication year |
-| **Training Objective** | Contrastive learning (DPR-style) using binary similarity loss |
-| **Training Data** | [Talk2Ref dataset](https://huggingface.co/datasets/s8frbroy/talk2ref) – transcripts of 6,279 scientific talks |
-| **Task** | Encode scientific talks into a shared semantic space with their cited papers |
-| **Paired Model** | [Talk2Ref Cited Paper Encoder](https://huggingface.co/s8frbroy/talk2ref_ref_key_cited_paper_encoder) |
+| Property | Description |
+|-----------|-------------|
+| **Architecture** | Sentence-BERT (all-MiniLM-L6-v2 backbone) |
+| **Pooling** | Mean pooling |
+| **Max sequence length** | 512 tokens |
+| **Training data** | Talk2Ref dataset (≈ 43 k cited papers linked to 6 k talks) |
+| **Objective** | Contrastive binary (DPR-style) loss |
+| **Task** | Encode cited papers into a shared semantic space with talk transcripts |
 
 ---