Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,52 @@
+---
+library_name: pytorch
+tags:
+  - table-retrieval
+  - embedding-adapter
+  - centroid-adapter
+---
+
+# Universal Representation Adapter — splade
+
+Lightweight **BottleneckResidualAdapter** trained on top of
+[splade](https://huggingface.co/splade) embeddings to produce
+representation-invariant table embeddings.
+
+## Architecture
+
+```
+z = e + α · Up( Dropout( GELU( Down( LN(e) ) ) ) )
+```
+
+| Hyperparameter | Value |
+|---|---|
+| Embedding dim `d` | 30522 |
+| Bottleneck rank `r` | 512 |
+| Residual scale `α` | 0.01 |
+| Use bias | True |
+
+Trained on: WTQ, WIKISQL
+
+## Usage
+
+```python
+import torch
+from huggingface_hub import hf_hub_download
+import json
+
+# --- option A: use the from_pretrained helper in this repo ---
+# (copy BottleneckResidualAdapter + from_pretrained from push_to_hub.py)
+adapter = BottleneckResidualAdapter.from_pretrained("KBhandari11/centroid-adapter-subset-splade")
+e = torch.randn(1, 30522)   # your backbone embedding
+z = adapter(e)                    # representation-invariant embedding
+
+# --- option B: hf_hub_download one-liner ---
+from safetensors.torch import load_file
+weights_path = hf_hub_download("KBhandari11/centroid-adapter-subset-splade", "model.safetensors")
+cfg_path     = hf_hub_download("KBhandari11/centroid-adapter-subset-splade", "config.json")
+with open(cfg_path) as f:
+    cfg = json.load(f)
+adapter = BottleneckResidualAdapter(**cfg)
+adapter.load_state_dict(load_file(weights_path))
+adapter.eval()
+```