DEPLOYMENT.md

HuggingFace Space Deployment Guide

This directory contains files for deploying the DGA classifier to HuggingFace Spaces.

Files

• app.py - Main Gradio application
• requirements.txt - Python dependencies
• README.md - Space description and metadata

Deployment Steps

1. Upload Model to HuggingFace Hub

First, upload your trained model:

# Make sure you have huggingface_hub installed
pixi add huggingface_hub

# Login to HuggingFace (get token from https://huggingface.co/settings/tokens)
pixi run huggingface-cli login

# Upload the model
cd ..  # Go back to project root
pixi run python upload_to_hf.py

2. Create HuggingFace Space

1. Go to https://huggingface.co/spaces
2. Click "Create new Space"
3. Fill in the details:
   • Space name: dga-detector (or any name you prefer)
   • License: MIT
   • SDK: Gradio
   • Hardware: CPU Basic (free) or GPU if you need faster inference
   • Visibility: Public

3. Upload Files to Space

You can either:

Option A: Use Git (Recommended)

# Clone your Space repository
git clone https://huggingface.co/spaces/ccss17/dga-detector
cd dga-detector

# Copy files from this directory
cp ../hf_space/* .

# Update app.py with your model name
# Edit line 34: MODEL_NAME = "YOUR_USERNAME/dga-transformer-encoder"

# Commit and push
git add .
git commit -m "Initial commit: DGA detector demo"
git push

Option B: Use Web UI

1. Go to your Space page
2. Click "Files" tab
3. Upload app.py, requirements.txt, and README.md

4. Update Model Name

In app.py, line 34, change:

MODEL_NAME = "ccss17/dga-transformer-encoder"

to:

MODEL_NAME = "YOUR_USERNAME/dga-transformer-encoder"

5. Wait for Build

HuggingFace will automatically:

• Install dependencies from requirements.txt
• Download your model from HuggingFace Hub
• Build and deploy the Gradio app

This usually takes 2-5 minutes.

6. Test Your Space

Once deployed, your Space will be available at:

https://huggingface.co/spaces/YOUR_USERNAME/dga-detector

Troubleshooting

Model not found

• Make sure you uploaded the model to HuggingFace Hub first
• Check that the model name in app.py matches your repository

Build fails

• Check the "Logs" tab in your Space
• Ensure requirements.txt has all necessary dependencies
• Make sure you're using compatible versions

Slow inference

• Consider upgrading to GPU hardware (Settings > Hardware)
• Free GPU is available: T4 small

Customization

Change theme

In app.py, line 179:

with gr.Blocks(title="DGA Domain Classifier", theme=gr.themes.Soft()) as demo:

Try other themes: gr.themes.Base(), gr.themes.Default(), gr.themes.Glass()

Add more examples

Edit the EXAMPLES list in app.py (line 160)

Modify UI

Edit the Gradio Blocks structure in app.py (lines 179-340)

Resources

• HuggingFace Spaces Documentation
• Gradio Documentation
• Model Repository