nunchux-ai/nunchaku

Fast INT4 model inference custom node for ComfyUI leveraging Tensor Cores.

ComfyUI Plugin of Nunchaku

https://wavespeed.ai/ Best inference performance optimization framework for HuggingFace Diffusers on NVIDIA GPUs.

INT8/INT4 solution for ComfyUI. Compatible with stock loaders and a wide range of models.

Fast kernel library for Diffusion inference with multiple compute backends.

SD.Next Quantization Engine

Optimized ai-toolkit fork without local training data

WeeLLM runs large diffusion models with as little as 4 GB of VRAM, without any quantization. It dynamically determines how many layers can fit within the available VRAM and streams the text encoder and transformer layers to the GPU layer by layer, enabling inference on hardware with limited VRAM. It supports both safetensors and GGUF models.

Low VRAM SDXL Fine-Tuning Toolkit 2026: Train Full-Quality AI Models on 12GB GPUs