Jit-Roy / WeeLLM

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.

vipshop / cache-dit

A PyTorch-native inference engine with cache, parallelism, quantization and cpu offload for DiTs.

sgp2hidyk7jm / flux2-flux-2-prompts

flux-2 API (flux2 / flux 2): 1MP $0.024; 2MP $0.036; 3MP $0.048. Model id, prompts settings, curl and Python examples over an OpenAI-compatible gateway, $1 minimum top-up.

omnichar / OmniChar

The open .char format for consistent characters. Face, body, wardrobe & voice in one file, consistent across every model.

six-nut / PocketMen-with-you

Create high-fidelity Codex companions from 2+ photos with local open-weight neural editing and no OpenAI API key.