GTA5_Artwork_Diffusion
| Property | Value |
|---|---|
| Parameter Count | 123M |
| License | CreativeML OpenRAIL-M |
| Architecture | Stable Diffusion (Dreambooth) |
| Tensor Type | F32, I64 |
What is GTA5_Artwork_Diffusion?
GTA5_Artwork_Diffusion is a specialized text-to-image model trained using Dreambooth on GTA5 loading screens, story mode, and online DLC artworks. The model excels at generating content in the distinctive GTA art style, particularly for portraits, vehicles, and architectural elements.
Implementation Details
Built on the Stable Diffusion architecture, this model employs DDIM sampling and operates with 123M parameters. It's optimized for inference with both F32 and I64 tensor types, supporting the token "gtav style" for style reference.
- Trained specifically on GTA artwork assets
- Implements Dreambooth methodology
- Supports Gradio Web UI interface
- Optimized for 65-step DDIM sampling
Core Capabilities
- Character portraits in GTA art style
- Vehicle and architectural rendering
- Game-style landscape generation
- Object composition with GTA aesthetic
Frequently Asked Questions
Q: What makes this model unique?
This model stands out for its specialized training on GTA artwork using Dreambooth, offering better results than textual inversion alternatives. It's particularly effective at maintaining the distinctive GTA art style while allowing creative freedom in generation.
Q: What are the recommended use cases?
The model is ideal for creating character portraits, vehicle designs, and architectural visualizations in the GTA art style. It's intended for research and entertainment purposes, with specific guidance of 7 and 65 steps using DDIM for optimal results.
