Hugging Face's logo

marcosremar2
/
comic-generator-multilingual

Model card Files
xet
 Community
comic-generator-multilingual / project_summary.md
Comic Developer
Initial clean commit - Comic generator with SDXL
4595df6
preview code
|
raw
history blame contribute delete
7.09 kB

HD French - Comic Story Generator Project Summary

Project Overview

This project generates comic stories with AI-generated characters and scenes using machine learning models. The system can create complete comic stories with multiple panels based on different themes. It now includes a web interface for easier story creation.

Current Status

The project has been successfully deployed to a remote server with NVIDIA RTX 3090 GPU and is fully functional. A new web interface has been added for easier access, along with scripts to manage the server and SSH tunnel.

Project Structure 

├── config.js              # Main configuration file
├── .env                   # API keys (not committed to Git)
├── .env.example          # Example API key configuration
├── package.json          # Node.js project configuration
├── README.md             # Project documentation
├── server.js             # Web server for frontend interface
├── start_server_remote.sh # Script to start web server on remote server
├── create_ssh_tunnel.sh   # Script to create SSH tunnel
├── create_ssh_tunnel_bg.sh # Script to create SSH tunnel in background
├── start_web_interface.sh # Script to start web interface (combined)
├── web/                  # Web interface files
│   ├── index.html        # Main web interface
│   ├── styles.css        # Styling for web interface
│   └── script.js         # Frontend logic
├── src/                  # Node.js source code
│   ├── index.js          # Main entry point (API version)
│   ├── index.demo.js     # Demo version
│   ├── utils/
│   │   └── fireworks.js  # Fireworks AI API client
│   ├── models/
│   │   ├── characterGenerator.js  # Character generation
│   │   └── personagene.js         # Personagene creation
│   └── comics/
│       └── storyGenerator.js      # Comic story generation
├── Python Scripts (on server)
│   ├── stable_diffusion_generator.py  # Image generation with Stable Diffusion
│   ├── comic_story_generator.py       # Complete comic story generation
│   └── comic_viewer.py               # HTML viewer generation

Server Deployment

  • Server: vast.ai instance
  • IP: 207.102.87.207
  • Port: 53887
  • User: root
  • Workspace Directory: /workspace/hd-french-setup
  • GPU: NVIDIA RTX 3090

Key Components

1. Node.js Backend (Local Development)

  • Configuration management
  • Fireworks AI API integration
  • Character and story generation logic
  • Demo mode for testing without API keys
  • Web server for frontend interface

2. Python ML Components (Server)

  • Stable Diffusion Generator: Creates individual images from text prompts
  • Comic Story Generator: Creates complete 5-panel comic stories
  • Comic Viewer: Generates HTML pages to view comic stories

3. Web Interface

  • HTML/CSS/JS Frontend: User-friendly interface for story creation
  • Web Server: Serves frontend files and handles requests

4. Management Scripts

  • start_server_remote.sh: Starts web server on remote server in background
  • create_ssh_tunnel.sh: Creates SSH tunnel to access web interface
  • create_ssh_tunnel_bg.sh: Creates SSH tunnel in background
  • start_web_interface.sh: Combined script to start web interface (server + tunnel)

Recent Accomplishments

Comic Story Generated

  • Title: "The Quest for the Lost Artifact"
  • Theme: Adventure
  • Character: A brave elf warrior with silver armor and a magical sword
  • Setting: Ancient ruins in a mystical forest
  • Panels Generated: 5
  • Directory: comic_1756920589

Web Interface Added

  • Feature: Browser-based interface for story creation
  • Technology: HTML, CSS, JavaScript
  • Access: Run npm run web and visit http://localhost:8081

Management Scripts Added

  • Feature: Easy management of remote server and SSH tunnel
  • Technology: Bash scripts
  • Access: Run ./start_server_remote.sh and ./create_ssh_tunnel_bg.sh

Files Created on Server

  1. comic_1756920589/panel_1.png - Establishing shot
  2. comic_1756920589/panel_2.png - Character close-up
  3. comic_1756920589/panel_3.png - Action scene
  4. comic_1756920589/panel_4.png - Climactic scene
  5. comic_1756920589/panel_5.png - Resolution
  6. comic_1756920589/story_metadata.json - Story metadata
  7. comic_1756920589/comic_story.html - HTML viewer

How to Access the Generated Comic

SSH Connection 

ssh -p 53887 root@207.102.87.207
cd /workspace/hd-french-setup

View Generated Files 

ls -la comic_1756920589/

Download Files Locally 

scp -P 53887 root@207.102.87.207:/workspace/hd-french-setup/comic_1756920589/* .

Running the System

Local Demo (No API Required) 

npm run demo

Web Interface 

# Start server on remote server
./start_server_remote.sh

# Create SSH tunnel (choose one):
./create_ssh_tunnel.sh     # Keeps terminal occupied
./create_ssh_tunnel_bg.sh  # Runs in background

# Then visit http://localhost:8082 in your browser

Server-Side Comic Generation 

# SSH into server
ssh -p 53887 root@207.102.87.207

# Navigate to project directory
cd /workspace/hd-french-setup

# Generate new comic story
python3 comic_story_generator.py [theme] [title]

# Generate HTML viewer for existing comic
python3 comic_viewer.py comic_1756920589

Supported Story Themes

  1. Adventure - Default theme
  2. Mystery - Detective stories
  3. Romance - Love stories

Performance Information

  • Image Generation Time: 1-2 seconds per panel on RTX 3090
  • Complete Story Generation: 2-3 minutes for 5 panels
  • Model Used: Stable Diffusion v1.5 (public access)

Future Improvements

  1. Integration with Flux Schnell model (requires Hugging Face access)
  2. Enhanced web interface with real API integration
  3. Custom character creation tools
  4. Additional story themes and templates
  5. User account system for saving stories

Troubleshooting

Common Issues

  1. Hugging Face Access: Flux Schnell requires authorized access
  2. GPU Memory: Large models may require significant VRAM
  3. Network Latency: SSH connections may have delays
  4. Web Interface: Ensure port 8081 is available

Solutions

  1. Use Stable Diffusion as alternative (already implemented)
  2. Monitor GPU usage with nvidia-smi
  3. Use background processes for long-running tasks
  4. Check for port conflicts when running web server

Git Repository

The project is version controlled with Git. Current commit: 0047d5b6

Next Steps

  1. Continue developing additional features
  2. Test with different story themes
  3. Optimize image generation parameters
  4. Enhance web interface with real API integration
  5. Add user account system for saving stories

Document created: September 3, 2025 Project Status: Fully functional with active server deployment, web interface, and management scripts