Update README.md

README.md

@@ -1,40 +1,35 @@
+---
+title: Handwritten Text Recognition
+emoji: ✍️
+colorFrom: blue
+colorTo: purple
+sdk: gradio
+sdk_version: "4.44.0"
+python_version: "3.13"
+app_file: app.py
+pinned: false
+---
+
 # ✍️ Handwritten Line Text Recognition (HTR)
 
 This project implements a **Handwritten Line Text Recognition (HTR)** system using Deep Learning with TensorFlow.
 
-The model is based on a **CRNN (CNN + BiLSTM + CTC Loss)** architecture and can recognize handwritten English text lines from images.
-
-Deployed with ❤️ on Hugging Face Spaces using Gradio.
-
----
-
-## 🚀 Demo
-
-Upload an image of a handwritten text line and the model will predict the text.
+The model is based on a **CRNN (CNN + BiLSTM + CTC)** architecture and can recognize handwritten English text lines from images.
 
-Supported:
-- English characters (A–Z, a–z)
-- Numbers (0–9)
-- Space
+Deployed on Hugging Face Spaces using Gradio.
 
 ---
 
-## 🧠 Model Architecture
-
-The architecture follows a standard CRNN pipeline:
-
-Image → CNN → Feature Map → BiLSTM → Dense → CTC Decoding
+## 🚀 How It Works
 
-### Components:
+Image → CNN → BiLSTM → Dense → CTC Decode → Text Output
 
-- **CNN layers**
-  - Feature extraction from handwritten image
-- **Bidirectional LSTM**
-  - Sequence modeling
-- **CTC (Connectionist Temporal Classification)**
-  - Handles variable-length sequences without alignment
+The model:
+- Extracts visual features using Convolutional layers
+- Models sequence dependencies with Bidirectional LSTM
+- Uses CTC decoding to predict variable-length text
 
 ---
 
-## 📂 Project Structure
+## 📂 Project Files