ggml-org/llama.cpp

Pull Requests Commits
speculative: add --n-gpu-layers-draft option (#3063)
sozforex committed 2 years ago
Verified 84e72365
arm64 support for windows (#3007)
es0m committed 2 years ago
Verified b52b29ab
CUDA: fix LoRAs (#3130)
JohannesGaessler committed 2 years ago
Verified 4f7cd6ba
CUDA: fix mul_mat_q not used for output tensor (#3127)
JohannesGaessler committed 2 years ago
Verified 89e89599
CUDA: lower GPU latency + fix Windows performance (#3110)
JohannesGaessler committed 2 years ago
Verified d54a4027
cmake : support build for iOS/tvOS (#3116)
jhen0409 committed 2 years ago
Verified 1b0d0925
CUDA: add device number to error messages (#3112)
JohannesGaessler committed 2 years ago
Verified 8a4ca9af
metal : PP speedup (#3084)
ikawrakow committed 2 years ago
Verified f31b6f4e
convert: remove most of the n_mult usage in convert.py (#3098)
Green-Sky committed 2 years ago
Verified 6eeb4d90
metal : support for Swift (#3078)
kchro3 committed 2 years ago
Verified 21ac3a15
metal : support build for iOS/tvOS (#3089)
jhen0409 committed 2 years ago
Verified 4fd54779
flake : add train-text-from-scratch to flake.nix (#3042)
takov751 committed 2 years ago
Verified ec2a24fe
readme : fix typo (#3043)
eltociear committed 2 years ago
Verified 7d99aca7
metal : Q3_K speedup (#2995)
ikawrakow committed 2 years ago
Verified ba7ffbb2
examples : make n_ctx warning work again (#3066)
cebtenzzre committed 2 years ago
Verified e64f5b55
readme : update hot tpoics
ggerganov committed 2 years ago
Verified 94f10b91
sync : ggml (CUDA GLM RoPE + POSIX) (#3082)
ggerganov committed 2 years ago
Verified b3e9852e
build : do not use _GNU_SOURCE gratuitously (#2035)
przemoc committed 2 years ago
Verified cb6c44c5
docker : add git to full-cuda.Dockerfile main-cuda.Dockerfile (#3044)
polym committed 2 years ago
Verified a21baeb1
Update deprecated GGML TheBloke links to GGUF (#3079)
SleepyYui committed 2 years ago
Verified 6ff712a6
ggml-alloc : correctly check mmap return value for errors (#3075)
slaren committed 2 years ago
Verified ebc96086
enable CPU HBM (#2603)
jikunshang committed 2 years ago
Verified 7f412dab
convert : fix F32 ftype not being saved (#3048)
cebtenzzre committed 2 years ago
Verified 6336d834
fix some warnings from gcc and clang-tidy (#3038)
cebtenzzre committed 2 years ago
Verified 00d62adb
make : improve test target (#3031)
cebtenzzre committed 2 years ago
Verified 4fa2cc17
make : fix CPPFLAGS (#3035)
cebtenzzre committed 2 years ago
Verified 5ffab089
llama-bench : use two tokens in the warmup run for prompt evals (#3059)
slaren committed 2 years ago
Verified 15b67a66
metal : parallel RoPE on Metal (#3024)
ikawrakow committed 2 years ago
Verified be8c9c24
metal : correct fix of kernel_norm (#3060)
ikawrakow committed 2 years ago
Verified be6beeb8
metal : fix kernel_norm (fixes Falcon on Metal) (#3057)
ggerganov committed 2 years ago
Verified c4f49664
Terms Of Use Privacy Policy Refund Policy Impressum