New
1.0.0
- 6-bit weight compression using coremltools
- Improved attention implementation (
SPLIT_EINSUM_V2) which yields up to 30% improved Neural Engine performance - Multilingual text encoder support
- New benchmarks for iPhone, iPad and Mac
Unclaimed project
Are you a maintainer of ml-stable-diffusion? Claim this project to take control of your public changelog and roadmap.
SPLIT_EINSUM_V2) which yields up to 30% improved Neural Engine performance