Skip to content

Distribute and run LLMs with a single file.

Commits on main

Make benchmark tool work more reliably
Justine Tunney committed 2 years ago
8b2f8d8
Avoid crashing on llava ctrl-c
Justine Tunney committed 2 years ago
5c40565
Fix o/depend file build
Justine Tunney committed 2 years ago
aa58b3a
Add missing CPUID checks
Justine Tunney committed 2 years ago
91dd4d3
Restore quantize_row_q8_K()
Justine Tunney committed 2 years ago
50dd001
Add llama-bench command (cpu mode only)
Justine Tunney committed 2 years ago
9206719
Introduce bf16 cuda support
Justine Tunney committed 2 years ago
c0aa43e
Import upstream ggml-cuda fixes
Justine Tunney committed 2 years ago
0b5997d
Sync with llama.cpp upstream
Justine Tunney committed 2 years ago
d228e01
Fix typo in llama.h (#354)
Ikko Eltociear Ashimine committed 2 years ago
b3aa97d
Fix f16 cpuid check
Justine Tunney committed 2 years ago
87d4ce1
938cf72
Enable GGML_CUDA_FORCE_MMQ in tinyBLAS mode
Justine Tunney committed 2 years ago
00e4f72
Fix crash in llava-cli with ctrl-c
Justine Tunney committed 2 years ago
8d1afe4
Update llama.cpp (#427)
Zibri committed 2 years ago
64fbffc
Fix typo in README.md (#407)
Matt committed 2 years ago
3a0cf4e
d5f614c
c660d38
Update HF links to point to new Mozilla account
Stephen Hood committed 2 years ago
966dd23
a86e7ce
Update another logo
Justine Tunney committed 2 years ago
d4099fe
Redraw tinyBLAS ANSI art logo
Justine Tunney committed 2 years ago
d10eb32
Release llamafile v0.8.4
Justine Tunney committed 2 years ago
30cdd9c
Release llamafile v0.8.3
Justine Tunney committed 2 years ago
ae34574
Re-apply grammar parser memory bug fix 22aba95a1
Justine Tunney committed 2 years ago
3fe045f
Make vectorized expf() handle underflow
Justine Tunney committed 2 years ago
b5c6df6
Release llamafile v0.8.2
Justine Tunney committed 2 years ago
4ee1e39