Commit graph

3 commits

Author SHA1 Message Date
Justine Tunney
e7eb0b3070
Make more ML improvements
- Fix UX issues with llama.com
- Do housekeeping on libm code
- Add more vectorization to GGML
- Get GGJT quantizer programs working well
- Have the quantizer keep the output layer as f16c
- Prefetching improves performance 15% if you use fewer threads
2023-05-16 08:07:23 -07:00
Justine Tunney
59e1c245d1 Get more Python tests passing (#141) 2021-08-16 15:26:31 -07:00
Justine Tunney
839e2f4cfb Add hyperbolic math functions
See #61
2021-02-28 01:27:50 -08:00