1.58-bit FLUX
- Published
- Source
- arXiv
- Paper number
- 006
- Field
- Image Generation
- arXiv ID
- 2412.18653
Key points
- The method quantizes FLUX.1-dev into ternary -1, 0, and +1 weights, reaching a 1.58-bit representation for a large text-to-image model.
- The quantization method is data-free, relying on self-supervision from the original model instead of an external image dataset.
- Reported efficiency gains include 7.7x less storage, 5.1x less inference memory, and lower latency through custom low-bit kernels.
- The benchmarks suggest that image quality at 1024×1024 remains similar despite the extreme compression.
Paper links
External research summaries. These are not HDATF publications or measured product results.