Upload README.md

README.md

@@ -0,0 +1,198 @@
+---
+title: Qwen3 Text Generator
+emoji: 🤖
+colorFrom: blue
+colorTo: purple
+sdk: gradio
+sdk_version: "4.0.0"
+app_file: app.py
+pinned: false
+---
+
+# Qwen3 Text Generator
+
+A text generation application using the Qwen3 0.6B model trained on TinyStories dataset.
+
+✨ **The model automatically downloads from HuggingFace** - no need to upload large files! Model repository: [vuminhtue/qwen3_sentiment_tinystories](https://huggingface.co/vuminhtue/qwen3_sentiment_tinystories)
+
+## 🚀 Quick Start
+
+### Running Locally
+
+1. Make sure you have the required files:
+   - `app.py` - The Gradio interface (automatically downloads model)
+   - `Qwen3_model.py` - The model architecture
+   - `requirements.txt` - Python dependencies
+
+2. Install dependencies:
+```bash
+pip install -r requirements.txt
+```
+
+3. Run the app:
+```bash
+python app.py
+```
+
+4. The app will automatically download the model from HuggingFace (first run only)
+5. Open your browser to the URL shown (usually http://127.0.0.1:7860)
+
+## 📤 Deploying to HuggingFace Spaces
+
+### Step 1: Prepare Your Files
+
+You only need these **3 files** (total ~40KB):
+- `app.py` - Main application (downloads model automatically)
+- `Qwen3_model.py` - Model architecture
+- `requirements.txt` - Dependencies (includes huggingface_hub)
+- `README.md` - This file (optional but recommended)
+
+**Note**: The model (`Qwen3_200k_model_params.pt`) is automatically downloaded from the HuggingFace model repository, so you don't need to upload it!
+
+### Step 2: Create a HuggingFace Space
+
+1. Go to https://huggingface.co/new-space
+2. Fill in the details:
+   - **Space name**: Choose a name (e.g., "qwen3-text-generator")
+   - **License**: Select your preferred license
+   - **Select the SDK**: Choose **Gradio**
+   - **Space hardware**: Start with "CPU basic" (free)
+
+3. Click "Create Space"
+
+### Step 3: Upload Your Files
+
+You have two options:
+
+#### Option A: Using Git (Recommended)
+
+```bash
+# Clone your new space
+git clone https://huggingface.co/spaces/YOUR_USERNAME/YOUR_SPACE_NAME
+cd YOUR_SPACE_NAME
+
+# Copy your 3 files (model downloads automatically!)
+cp /path/to/app.py .
+cp /path/to/Qwen3_model.py .
+cp /path/to/requirements.txt .
+cp /path/to/README.md .  # Optional but recommended
+
+# Commit and push (fast - only ~40KB!)
+git add .
+git commit -m "Initial commit: Add Qwen3 text generator"
+git push
+```
+
+#### Option B: Using the Web Interface
+
+1. On your Space page, click "Files" → "Add file" → "Upload files"
+2. Drag and drop or select all your files
+3. Click "Commit to main"
+
+### Step 4: Wait for Build
+
+- HuggingFace will automatically build and deploy your app
+- This may take 5-10 minutes
+- You'll see build logs in the "App" tab
+
+### Step 5: Test Your App
+
+Once the build is complete, your app will be live at:
+`https://huggingface.co/spaces/YOUR_USERNAME/YOUR_SPACE_NAME`
+
+## 🎮 How to Use the App
+
+1. **Enter Starting Text**: Type the beginning of your story (e.g., "Once upon a time")
+
+2. **Adjust Max New Tokens**: 
+   - Controls how much text to generate
+   - 10-50: Short continuation
+   - 50-100: Medium paragraph
+   - 100-200: Long passage
+
+3. **Adjust Temperature**:
+   - 0.1-0.7: More predictable, focused text
+   - 0.8-1.0: Balanced creativity
+   - 1.1-2.0: Very creative, more random
+
+4. **Click Generate**: Watch as the model continues your story!
+
+## 📊 Model Information
+
+- **Architecture**: Qwen3 0.6B
+- **Parameters**: 596 million (unique parameters)
+- **Training Data**: TinyStories dataset
+- **Best For**: Simple narratives, children's stories, everyday situations
+
+## 🔧 Troubleshooting
+
+### Model Download Issues
+
+If the model doesn't download automatically:
+- Check that the model repository is public: https://huggingface.co/vuminhtue/qwen3_sentiment_tinystories
+- Verify internet connectivity on HuggingFace Spaces
+- Check the build logs for specific error messages
+- Make sure `requirements.txt` includes `huggingface_hub>=0.16.0`
+
+### Out of Memory Error
+
+If you get memory errors:
+1. Go to your Space settings
+2. Upgrade to a better hardware tier (may require payment)
+3. The Qwen3 0.6B model requires about 2-3GB RAM
+
+### App Not Loading
+
+1. Check the build logs in the "App" tab
+2. Make sure all files are uploaded correctly
+3. Verify `requirements.txt` has all necessary packages
+4. Check that file names match exactly (case-sensitive)
+
+## 💡 Tips for Better Results
+
+1. **Good Prompts**: Start with clear, simple sentences
+   - ✅ "Once upon a time, there was a little girl"
+   - ❌ "Explain quantum physics"
+
+2. **Temperature Selection**:
+   - Use lower temperature (0.5-0.7) for coherent stories
+   - Use higher temperature (1.0-1.5) for creative variety
+
+3. **Token Length**: 
+   - Start with 30-50 tokens to see the style
+   - Increase if you want longer passages
+
+## 📝 File Structure
+
+**Files to upload to HuggingFace Spaces:**
+```
+.
+├── app.py                          # Main Gradio application (downloads model)
+├── Qwen3_model.py                  # Model architecture and helpers
+├── requirements.txt                # Python dependencies (includes huggingface_hub)
+└── README.md                       # This file (optional)
+```
+
+**Model file** (`Qwen3_200k_model_params.pt`) is automatically downloaded from HuggingFace, no need to upload!
+
+## 🤝 Contributing
+
+Feel free to:
+- Report issues
+- Suggest improvements
+- Share your generated stories!
+
+## 📜 License
+
+This project uses the Qwen3 architecture. Please check the license for your specific use case.
+
+## 🙏 Acknowledgments
+
+- Qwen3 architecture from Alibaba Cloud
+- Training approach inspired by "LLMs from Scratch"
+- TinyStories dataset for training data
+
+---
+
+**Enjoy generating creative stories! 📚✨**
+