| title: ShellD Text-to-Image | |
| emoji: 🐚 | |
| colorFrom: green | |
| colorTo: pink | |
| sdk: gradio | |
| sdk_version: 6.20.0 | |
| app_file: app.py | |
| short_description: ShellD — lightweight DiT text-to-image (256×256) | |
| python_version: "3.12" | |
| startup_duration_timeout: 30m | |
| # ShellD — Shell Diffusion Text-to-Image | |
| A demo for [FlameF0X/ShellD](https://huggingface.co/FlameF0X/ShellD), a lightweight 67M-parameter Diffusion Transformer (DiT) that generates 256×256 images from natural language prompts. Uses a frozen MiniLM-L6-v2 text encoder and a custom VAE + DiT backbone with a 1000-step DDPM sampler. |