MUMBAI, India, Oct. 5 -- Intellectual Property India has published a patent application (202611100954 A) filed by Indian Institute Of Information Technology Allahabad on August 20, 2026, for A System And Method For Prompt Complexity-Based Adaptive Inference Scheduling In Diffusion-Based Generative Models.
Inventors include Anuj Gupta; Jagriti Singh; and Prof. Shekhar Verma.
The application for the patent was published on October 02, 2026, under issue no. 40/2026.
Abstract: The present disclosure proposes a system (200) and method for prompt complexity-based adaptive inference scheduling in diffusion-based generative models through selective encoder execution and cached encoder-feature reuse during an iterative denoising process. The system (200) comprises one or more computing devices (202) collectively having at least one processor (204) and at least one memory (206) storing executable instructions. The at least one processor (204) executes the executable instructions stored in the at least one memory (206) to implement plurality of modules (208) that comprises a prompt complexity estimation module (210), a schedule generation module (220), and a diffusion inference engine (230). The system (200) supports operation with latent diffusion models, U-Net-based diffusion models, diffusion transformers, text-to-image generation systems, image-editing systems, video-generation systems, audio-generation systems, and multimodal generative systems.
Disclaimer: Curated by HT Syndication.
