Update model card: OdaxAI S.R.L. branding, full address and email

README.md

@@ -25,7 +25,9 @@ datasets:
 [![GitHub](https://img.shields.io/badge/GitHub-odaxai%2FVenice--H1-181717?logo=github)](https://github.com/odaxai/Venice-H1)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 
-**Nicolò Savioli, Ph.D.** — OdaxAI Research
+**Nicolò Savioli, Ph.D.**  
+OdaxAI S.R.L., Abano Terme (PD), Italy  
+nicolo.savioli@odaxai.com
 
 ---
 
@@ -36,7 +38,7 @@ Venice-H1 is a lightweight, backbone-decoupled re-ranking module for Referring I
 - **Multi-Scale Grid Signatures**: 4×4, 8×8, 16×16 spatial pooling → 675-dim descriptors
 - **Failure Gate**: binary classifier predicting whether Query 0 is suboptimal
 - **Gain Predictor**: IoU-regression head estimating improvement per alternative query
-- **Backbone-decoupled**: works on cached features, no retraining of DeRIS
+- **Backbone-decoupled**: works on cached features, no retraining of DeRIS required
 
 **Architecture**: 3-layer pre-norm Transformer encoder, 8 heads, hidden_dim=512  
 **Parameters**: 11,296,258 (11.3M) — matches paper exactly
@@ -95,7 +97,6 @@ print("Parameters:", sum(v.numel() for v in ckpt["model"].values() if hasattr(v,
 **Full inference pipeline** (requires DeRIS-L features):
 
 ```python
-# Clone repo
 # git clone https://github.com/odaxai/Venice-H1
 # cd Venice-H1 && pip install -r requirements.txt
 
@@ -103,12 +104,10 @@ from venice_h1.model.reranker import VeniceH1Reranker
 import torch
 from huggingface_hub import hf_hub_download
 
-# Load checkpoint
 ckpt_path = hf_hub_download(repo_id="OdaxAI/venice-h1", filename="venice_h1_deris_l.pt")
 ckpt = torch.load(ckpt_path, map_location="cpu", weights_only=False)
 cfg = ckpt["config"]
 
-# Build model (note: architecture uses original naming convention)
 model = VeniceH1Reranker(
     query_feat_dim=cfg["query_feat_dim"],
     hidden_dim=cfg["hidden_dim"],
@@ -118,11 +117,11 @@ model = VeniceH1Reranker(
 )
 model.eval()
 
-# features: (B, N_queries, 936) — query_embed(256) + grid_sig(675) + mask_stats(4) + det_score(1)
-# Inference
+# features: (B, N_queries, 936)
+# query_embed(256) + grid_sig(675) + mask_stats(4) + det_score(1)
 with torch.no_grad():
     out = model(features, det_scores, mask_means)
-    p_fail = out["p_fail"]        # (B,) — failure probability
+    p_fail = out["p_fail"]       # (B,) — failure probability
     selected = model.rerank(features, det_scores, mask_means)  # (B,) — best query index
 ```
 
@@ -135,7 +134,10 @@ with torch.no_grad():
 | `venice_h1_deris_l.pt` | **Trained checkpoint** — 11.3M params, DeRIS-L backbone |
 | `venice_h1_deris_l_metrics.json` | Full evaluation metrics |
 | `config.yaml` | Training hyperparameters |
-| `README.md` | This model card |
+| `venice_h1/` | Python package (model code) |
+| `train.py` | Training script |
+| `evaluate.py` | Evaluation script |
+| `scripts/extract_features.py` | Feature extraction from DeRIS |
 
 ---
 
@@ -165,7 +167,7 @@ with torch.no_grad():
   author  = {Savioli, Nicol\`{o}},
   journal = {arXiv preprint arXiv:XXXX.XXXXX},
   year    = {2025},
-  note    = {OdaxAI Research},
+  institution = {OdaxAI S.R.L., Abano Terme (PD), Italy},
 }
 ```
 
@@ -173,4 +175,5 @@ with torch.no_grad():
 
 ## License
 
-MIT License. © 2025 Nicolò Savioli — OdaxAI Research
+MIT License. © 2025 OdaxAI S.R.L., Abano Terme (PD), Italy.  
+All research conducted by Nicolò Savioli, Ph.D.