IDDM (Industrial, landscape, animate, latent diffusion), support LDM, DDPM, DDIM, PLMS, webui and distributed training. Pytorch实现扩散模型，生成模型，分布式训练

Global and local integrated gradient-based diffusion model for De novo drug design

Currently, social media has been playing an important role in the process of information diffusion. Exploring the pattern of message propagation on social network help us better prepare for natural disasters or human crises. So, we developed models, algorithms, and tools to generate simulated networks, analyze simulated networks, and simulate information diffusion on social network over time. This software provides the following functions: (1) Simulated social network generators. This network generator is designed to generate networks with different configuration, such as Random network, Full network, Circle Network, Prefattached network, SmallWord network, and so on. (2) Network analysis. Network analysis was designed to explore the characteristics of a network. Now this module can be used to analysis characters of network and nodes of network. (3) Community detection. In this tool, we integrated two algorithms for community detection, including CNM (Clauset-Newman-Moore community detection) and GN (Girvan-Newman community detection). (4) Simulating information diffusion on simulated network. In this module, we implemented two information diffusion models, independent cascade model (IC model) and linear threshold model (LT model). (5) Case study. This module provides functions to conduct case study. (6) Data preprocessing. Because there are some functions may need you load your data that organized as a special format. This module help user prepare data for further use.

This project implements a Diffusion-based image generation model with an integrated PatchGAN-style Discriminator to guide and stabilize training. Built with PyTorch, the architecture is designed for high-resolution and detail-preserving image synthesis tasks.

Developed an AI Image Generator web application using HTML, CSS, and JavaScript integrated with the Hugging Face API. Implemented text-to-image generation using pre-trained diffusion models. Designed a responsive UI to display generated images in real time. Improved user experience with loading states and error handling.

Global and local integrated gradient-based diffusion model for De novo drug design

Designed an AI-based sysytem using cGAN model to generate and redesign clothing images.Integrated with Stable Diffusion for text-to-image fashion design.