LeanModels/DFloat11
Lossless compression framework that reduces LLM and diffusion model size by ~30% while preserving identical outputs.
Not currently ranked — collecting fresh signals.
star history
DFloat11 is a lossless compression framework targeting LLMs and diffusion models for efficient GPU inference. It reduces model size by approximately 30% through specialized compression algorithms while guaranteeing bit-for-bit identical outputs to the original uncompressed model. The framework supports popular models including FLUX.1, Qwen3-8B, Wan2.1, and HiDream-I1, and includes CPU offloading capabilities to enable inference on resource-constrained hardware with reduced VRAM requirements.
Frequently asked
- What is LeanModels/DFloat11?
- Lossless compression framework that reduces LLM and diffusion model size by ~30% while preserving identical outputs.
- Is DFloat11 open source?
- Yes — LeanModels/DFloat11 is open source, released under the Apache-2.0 license.
- What language is DFloat11 written in?
- LeanModels/DFloat11 is primarily written in Python.
- How popular is DFloat11?
- LeanModels/DFloat11 has 649 stars on GitHub.
- Where can I find DFloat11?
- LeanModels/DFloat11 is on GitHub at https://github.com/LeanModels/DFloat11.