Update README.md (#1)

- Update README.md (2d12cf2c673224331a1eb24b5af26400730d6a36)

README.md

@@ -2,8 +2,24 @@
 license: apache-2.0
 datasets:
 - flwrlabs/pacs
+language:
+- en
+base_model:
+- google/siglip2-base-patch16-224
+pipeline_tag: image-classification
+library_name: transformers
+tags:
+- PACS-DG
+- domain generalization
+- SigLIP2
 ---
 
+![4.png](https://cdn-uploads.huggingface.co/production/uploads/65bb837dbfb878f46c77de4c/2M1HRenGKvzLJiAdaexKs.png)
+
+# **PACS-DG-SigLIP2**
+
+> **PACS-DG-SigLIP2** is a vision-language encoder model fine-tuned from **google/siglip2-base-patch16-224** for **multi-class domain generalization** classification. It is trained to distinguish visual domains such as **art paintings**, **cartoons**, **photos**, and **sketches** using the **SiglipForImageClassification** architecture.
+
 ```py
 Classification Report:
               precision    recall  f1-score   support
@@ -38,4 +54,87 @@ id2label = {str(i): label for i, label in enumerate(labels)}
 
 # Print the mapping
 print(id2label)
-```
+```
+
+---
+
+## **Label Space: 4 Domain Categories**
+
+The model predicts the most probable visual domain from the following:
+
+```
+Class 0: "art_painting"
+Class 1: "cartoon"
+Class 2: "photo"
+Class 3: "sketch"
+```
+
+---
+
+## **Install dependencies**
+
+```bash
+pip install -q transformers torch pillow gradio
+```
+
+---
+
+## **Inference Code**
+
+```python
+import gradio as gr
+from transformers import AutoImageProcessor, SiglipForImageClassification
+from PIL import Image
+import torch
+
+# Load model and processor
+model_name = "prithivMLmods/PACS-DG-SigLIP2"  # Update to your actual model path on Hugging Face
+model = SiglipForImageClassification.from_pretrained(model_name)
+processor = AutoImageProcessor.from_pretrained(model_name)
+
+# Label map
+id2label = {
+    "0": "art_painting",
+    "1": "cartoon",
+    "2": "photo",
+    "3": "sketch"
+}
+
+def classify_pacs_image(image):
+    image = Image.fromarray(image).convert("RGB")
+    inputs = processor(images=image, return_tensors="pt")
+    
+    with torch.no_grad():
+        outputs = model(**inputs)
+        logits = outputs.logits
+        probs = torch.nn.functional.softmax(logits, dim=1).squeeze().tolist()
+    
+    prediction = {
+        id2label[str(i)]: round(probs[i], 3) for i in range(len(probs))
+    }
+
+    return prediction
+
+# Gradio Interface
+iface = gr.Interface(
+    fn=classify_pacs_image,
+    inputs=gr.Image(type="numpy"),
+    outputs=gr.Label(num_top_classes=4, label="Predicted Domain Probabilities"),
+    title="PACS-DG-SigLIP2",
+    description="Upload an image to classify its visual domain: Art Painting, Cartoon, Photo, or Sketch."
+)
+
+if __name__ == "__main__":
+    iface.launch()
+```
+
+---
+
+## **Intended Use**
+
+The **PACS-DG-SigLIP2** model is designed to support tasks in **domain generalization**, particularly:
+
+- **Cross-domain Visual Recognition** – Identify the domain style of an image.
+- **Robust Representation Learning** – Aid in training or evaluating models on domain-shifted inputs.
+- **Dataset Characterization** – Use as a tool to explore domain imbalance or drift.
+- **Educational Tools** – Help understand how models distinguish between stylistic image variations.