ggerganov/llama.cpp

Pull Requests Commits
metal : add poc for normalized Q4_0 and Q4_1
ggerganov committed 1 year ago
Verified b4e70822
Merge branch 'master' into norm-quants
ggerganov committed 1 year ago
Verified 9ffe54ed
convert.py : use dir name to name the llama
ggerganov committed 1 year ago
Verified b532a69b
examples : fix underscore in beam-search + .gitignore (close #2900)
ggerganov committed 1 year ago
Verified c90d135e
gguf : add workflow for Pypi publishing (#2896)
monatis committed 1 year ago
Verified 0d1c7061
make : add test and update CI (#2897)
alonfaraj committed 1 year ago
Verified 95092944
docs : add `node-llama-cpp` to `README.md` (#2885)
giladgd committed 1 year ago
Verified 35092fb5
convert : various script cleanups/fixes + merges and special token handling (#2842)
KerfuffleV2 committed 1 year ago
Verified dc07dc49
llm.vim : stop generation at multiple linebreaks, bind to <F2> (#2879)
chaihahaha committed 1 year ago
Verified ad9ddcff
main : log file (#2748)
staviq committed 1 year ago
Verified 8341a259
tests : add a C compliance test (#2848)
cebtenzzre committed 1 year ago
Verified 84940895
ggml : add view_src and view_offs to ggml_tensor for views (#2874)
slaren committed 1 year ago
Verified 06abf8ee
remove outdated references to -eps and -gqa from README (#2881)
slaren committed 1 year ago
Verified c03a243a
Tell users attmepting to run perplexity with too few tokens to use more (#2882)
ikawrakow committed 1 year ago
Verified fa3582f5
10X faster BPE tokenizer (#2876)
ikawrakow committed 1 year ago
Verified e37e69dc
py : fix "usage" messages (#2873)
maddes8cht committed 1 year ago
Verified 53885d72
convert.py : fix baichuan7B support (#2870)
jameswu2014 committed 1 year ago
Verified bcce96ba
readme : add react-native binding (#2869)
jhen0409 committed 1 year ago
Verified 74e0caeb
make : fix clang tests build, add missing examples (#2859)
cebtenzzre committed 1 year ago
Verified d4b5e16c
metal : add option to disable debug logs (close #2764)
ggerganov committed 1 year ago
Verified 3a007648
scripts : add pipefail
ggerganov committed 1 year ago
611363ac
added `struct` to llama_dump_timing_info_yaml's `llama_context` (#2857)
MarcusDunn committed 1 year ago
Verified 95b6e521
train : mem usage and other improvements (#2439)
xaedes committed 1 year ago
Verified 44c117f4
llama-bench : set locale to utf8 (#2832)
slaren committed 1 year ago
Verified 43033b7b
YAML result logging + preset script (#2657)
JohannesGaessler committed 1 year ago
Verified 6b73ef12
make : fix tests build (#2855)
alonfaraj committed 1 year ago
Verified 75fafcbc
llama.cpp : fix wrong vsnprintf call in MS compiler (#2856)
grahameth committed 1 year ago
Verified be475f60
ggml : tiny ggml_vec_dot_q4_K_q8_K AVX2 improvement (#2819)
hydroo committed 1 year ago
Verified 3af6b863
ggml : sync (mem align to header + conv_transpose_2d fixes + ggml_alloc) (#2852)
ggerganov committed 1 year ago
Verified 35feac65
CUDA: fix RoPE asserts, block sizes (#2833)
JohannesGaessler committed 1 year ago
Verified 92b1bbd2
Terms Of Use Privacy Policy Refund Policy Impressum