Skip to content

0.2.0

Compare
Choose a tag to compare
@marella marella released this 21 May 02:05
· 121 commits to main since this release

Changes

  • Add support for LLaMA, MPT models
  • Change repetition_penalty default value from 1.0 to 1.1

Breaking Changes

  • Update GGML library which has breaking changes to quantization formats. Old models have to be re-quantized.

Some of the latest quantized models are available here.