Update README.md

README.md

@@ -1,24 +1,40 @@
 ---
+language: en
 license: apache-2.0
-pipeline_tag: text-classification
+base_model: distilbert-base-uncased
 tags:
-- text-classification
-- fraud-detection
-- distilbert
-inference:
-  parameters:
-    top_k: 1
-language:
-- en
+  - text-classification
+  - fraud-detection
+  - transformer
+  - distilbert
+  - huggingface
+pipeline_tag: text-classification
+widget:
+  - text: "We require an urgent refund for the suspicious transaction on our account."
 ---
 
-# 🔍 fraud-model-aura
+# 🕵️‍♂️ Fraud Model Aura (KS-Vijay)
+
+This model uses **DistilBERT** to classify whether a given grievance or complaint text contains **fraudulent intent or behavior**. It is trained as part of an intelligent **Grievance Redressal Platform** to auto-detect fraud-related issues in startup complaints.
+
+## 🧠 Use Case
+
+Detects if the complaint relates to fraud:
+- `Fraud`
+- `Legitimate`
+
+This helps startups or service providers to **flag, escalate, or triage suspicious reports** quickly.
 
-This is a fine-tuned `distilbert` model built to classify whether a given input text is **fraudulent or not**.
+## 🔍 Model Summary
 
-## 🧠 Model Use Case
+- **Model Type:** Text Classification
+- **Architecture:** DistilBERT (uncased)
+- **Output Labels:** `Fraud`, `Legitimate`
+- **Weights Format:** `safetensors`
+- **Dataset:** Custom (based on complaints.csv)
+- **Training Framework:** PyTorch using 🤗 `transformers`
 
-Designed to detect fraudulent activities based on textual complaints, descriptions, or reports.
+## 📥 Example Input
 
-### Try it in the widget above 👆  
-Enter an input like:
+```text
+"I think someone is misusing our company’s KYC information to open fake accounts."