MediaPipe On-Device Text-to-Image Generation Solution Now Available for Android Developers- Google Developers Blog
Oct 9, 2023Paul Ruiz, Kris Tonthat

Summary
Introduces MediaPipe Image Generator, an early, experimental solution for on-device text-to-image generation on Android. It supports standard diffusion models (Stable Diffusion v1.5 architecture), controllable generation with conditioning images via a plugin system (facial structures, edge detection, depth awareness), and customization using Low-Rank Adaptation (LoRA) weights for specific concepts. Provides code examples for creating an ImageGenerator, passing prompts, iterations, and seeds, and extracting results. Mentions the ability to step through iterations manually. Links to conversion scripts, documentation, example app on GitHub, and Vertex AI for LoRA fine-tuning.
More from the Coral Edge AI Platform trail
- ML Drift: Next-Gen GPU AI/ML Inference at the Edge- Google Developers Blog
- Bring multimodal semantic search to the edge with EmbeddingGemma 2- Google Developers Blog
- EmbeddingGemma 2: The Developer Guide- Google Developers Blog
- Mastering Edge AI on Raspberry Pi with LiteRT and Gemma- Google Developers Blog
- LiteRT.js, Google's high performance Web AI Inference- Google Developers Blog