Home
Softono
y

yanwenkun

Professional software vendor delivering innovative solutions on the Softono platform. Specialized in both open-source and proprietary software development.

Total Products
2

Software by yanwenkun

ComfyUI Windows Portable
Open Source

ComfyUI Windows Portable

ComfyUI-Windows-Portable is a standalone distribution of ComfyUI for Windows designed for NVIDIA GPUs from 2018 or newer. This all-in-one package eliminates complex setup by resolving all dependencies and coming pre-loaded with over 40 custom nodes and 300 mutually compatible Python packages. Key pre-installed components include performance optimization libraries such as SageAttention, FlashAttention, xFormers, and Nunchaku, as well as compiled native packages like insightface and dlib that typically require difficult manual installation. Unlike the official standalone version, this release integrates these tools out of the box to streamline image generation workflows. The package explicitly excludes large stable diffusion model files to keep initial download size manageable, requiring users to place their preferred models in the designated folder. It supports a DIY approach via GitHub pipelines, allowing users to fork the repository and build custom versions without configuring CI/CD. Users must extract the

LLM Tools & Chat UIs
555 Github Stars
ComfyUI Docker
Open Source

ComfyUI Docker

ComfyUI-Docker provides official Docker images and startup scripts for ComfyUI, a node-based graphical interface for generative AI workflows. This project simplifies the deployment of ComfyUI on Linux systems using containerization. It offers a variety of optimized image tags tailored to specific hardware architectures and usage needs. For NVIDIA GPU users, images support CUDA 13.0 and CUDA 12.6, with recommendations for the cu130-slim-v2 tag for beginners and cu130-megapak-pt211 for an all-in-one experience including development kits and numerous custom nodes. The package also includes support for AMD GPUs via ROCm 6 and 7, Intel GPUs using XPU, and a nightly build for testing the latest PyTorch features. The software includes detailed volume mounting examples to manage model storage, cache, user inputs, and outputs effectively. By leveraging PyTorch builds optimized for modern GPU architectures like Blackwell, Hopper, and Ada Lovelace, these images ensure high performance and stability. The repository serve

AI & Machine Learning DevOps & Infrastructure
1.5K Github Stars