Skip to content

Distribute and run LLMs with a single file.

Commits on main

Support CUDA without cuBLAS (#82)
Jōshin committed 2 years ago
72e1c72
Release llamafile v0.3
Justine Tunney committed 2 years ago
1f17930
Avoid bind() conflicts on port 8080 w/ server
Justine Tunney committed 2 years ago
c1c8683
Remove some heavyweight asserts
Justine Tunney committed 2 years ago
330c4dc
Allow --grammar to be used on --image prompts
Justine Tunney committed 2 years ago
1cd334f
Revert "Make grammar faster"
Justine Tunney committed 2 years ago
e3f2d88
Add i386 sanity check for Windows users
Justine Tunney committed 2 years ago
fd05108
Add CLI chatbot example to README
Justine Tunney committed 2 years ago
ff02315
Make memory map loader go much faster on MacOS
Justine Tunney committed 2 years ago
cd0fae0
Make LLaVA more shell script friendly
Justine Tunney committed 2 years ago
6cb036c
Recognize cache_prompt parameter in OpenAI API
Georgi Gerganov committed 2 years ago
8cb9fd8
Make grammar faster
Justine Tunney committed 2 years ago
7a53c17
Support optional tensors
Georgi Gerganov committed 2 years ago
97f6431
Support attention_bias on LLaMA architecture
CausalLM committed 2 years ago
dd1e0cd
Fix integer overflow during quantization
Georgi Gerganov committed 2 years ago
55467d9
Make ctrl-c work better
Justine Tunney committed 2 years ago
baf216a
Slightly speed up LLaVA runtime dispatch on Intel
Justine Tunney committed 2 years ago
b4dea04
Explain how to summarize HTML URLs and images
Justine Tunney committed 2 years ago
3bc37ad
Remove stdout logging in llava-cli
Justine Tunney committed 2 years ago
c895a44
Fix more modelines (#73)
Jōshin committed 2 years ago
5ea929c
Fix modelines (#72)
Jōshin committed 2 years ago
c43d896
Add mkdir statements to `make install`
Justine Tunney committed 2 years ago
ca1f665
Add Windows CUDA instructions (#69)
recursionbane committed 2 years ago
69ecde2
Have makefile download cosmocc automatically
Justine Tunney committed 2 years ago
ff1b437
Put finishing touches on zipalign tool
Justine Tunney committed 2 years ago
76d47c0
Fix llamafile man page
Justine Tunney committed 2 years ago
2a80151
Write man pages for all commands
Justine Tunney committed 2 years ago
7a3e557
Fix output path in llamafile-quantize
Justine Tunney committed 2 years ago
c8cd8e1
Add `make install` build rule
Justine Tunney committed 2 years ago
762ad79
Reduce 2mb SVG in README to 166kb transparent PNG
Justine Tunney committed 2 years ago
2000b66