readme fixed

README.md

@@ -1,6 +1,19 @@
+---
+title: Motivational Quote Generator
+emoji: 🌟
+colorFrom: blue
+colorTo: purple
+sdk: gradio
+sdk_version: 4.44.0
+app_file: app.py
+pinned: false
+license: mit
+hardware: cpu-basic
+---
+
 # 🌟 Motivational Quote Generator
 
-A fine-tuned Llama 3.2 3B model that generates motivational quotes and advice. This model has been specifically trained on a curated dataset of inspirational content to provide guidance on various life topics.
+A Llama 3.2 3B model optimized for generating motivational quotes and advice. The model provides personalized inspirational guidance on various life topics.
 
 **🚀 Try it live: [https://huggingface.co/spaces/balazsthomay/motivational-quote-generator](https://huggingface.co/spaces/balazsthomay/motivational-quote-generator)**
 
@@ -9,14 +22,13 @@ A fine-tuned Llama 3.2 3B model that generates motivational quotes and advice. T
 - **Personalized Advice**: Get motivational quotes tailored to your specific situation
 - **Multiple Topics**: Covers perseverance, leadership, success, personal growth, and more
 - **Adjustable Creativity**: Control the temperature for more or less creative responses
-- **Fast Generation**: Optimized with LoRA fine-tuning for efficient inference
+- **Fast Generation**: Optimized with 4-bit quantization for efficient inference
 
 ## 🛠️ Technical Details
 
 - **Base Model**: Llama 3.2 3B-Instruct (4-bit quantized)
-- **Fine-tuning Method**: LoRA (Low-Rank Adaptation)
-- **Training**: 2000 iterations on curated motivational quotes dataset
-- **Framework**: MLX for efficient Apple Silicon inference
+- **Framework**: PyTorch with Transformers
+- **Optimization**: BitsAndBytesConfig for memory efficiency
 
 ## 💡 Usage
 
@@ -27,16 +39,6 @@ Simply enter a topic you'd like advice about, such as:
 
 The model will generate a personalized motivational response to help inspire and guide you.
 
-## 🔧 Model Configuration
-
-- **LoRA Rank**: 8
-- **Training Iterations**: 2000
-- **Max Sequence Length**: 2048 tokens
-
-## 📊 Training Data
-
-The model was trained on a carefully curated dataset of motivational quotes, with theme-based labeling using Ollama for improved contextual understanding.
-
 ---
 
-*Built with ❤️ using MLX and Gradio*
+*Built with ❤️ using PyTorch and Gradio*