Skip to content

1.0.0
Compare
Choose a tag to compare
@atiorh atiorh released this 14 Jun 15:48
· 69 commits to main since this release
1.0.0
7f9c58a
  • 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
Assets 2
Loading