ggerganov/llama.cpp

Pull Requests Commits

speculative : add --draft CLI arg

ggerganov committed 1 year ago

Verified 847896ab

speculative : print encoding speed

ggerganov committed 1 year ago

Verified a15ca746

speculative : initial example

ggerganov committed 1 year ago

Verified c82c808d

perplexity : fix ETA by warming up the model with an empty run

ggerganov committed 1 year ago

Verified 8f429fa5

gguf(python): Fix special vocab handling when id < 0 (#2984)

KerfuffleV2 committed 1 year ago

Verified 6519e9c9

metal : restore 363f0bf and fix reduce in F16_F32 kernels (#2986)

ggerganov committed 1 year ago

Verified b7f2aa9e

cov : disable comment in PRs (#2989)

alonfaraj committed 1 year ago

Verified 73a12a63

llama : fix bpe tokenize from byte (#2889)

opparco committed 1 year ago

Verified 37301347

metal : revert 6af0bab until we fix it

ggerganov committed 1 year ago

Verified d9151e6f

cov : add Code Coverage and codecov.io integration (#2928)

alonfaraj committed 1 year ago

Verified afc43d5f

opencl : fix a bug in ggml_cl_pool_malloc() for ggml_cl_mul_mat_f32() (#2955)

rchardx committed 1 year ago

Verified 6460f758

metal : more optimizations (#2959)

ikawrakow committed 1 year ago

Verified ca82cf7b

swift : add support for k-quants (#2983)

kchro3 committed 1 year ago

Verified 6a31a3bd

convert.py : BPE fixes (#2938)

KerfuffleV2 committed 1 year ago

Verified cff7b0bf

docs : add `catai` to `README.md` (#2967)

ido-pluto committed 1 year ago

Verified 340af42f

examples : fix gpt-neox (#2943)

mmnga committed 1 year ago

Verified c42f0ec6

swift : add missing c file to Package.swift (#2978)

kchro3 committed 1 year ago

Verified 2753415a

make : support overriding CFLAGS/CXXFLAGS/CPPFLAGS/LDFLAGS (#2886)

cebtenzzre committed 1 year ago

Verified bc054af9

logging: Fix creating empty file even when disabled (#2966)

KerfuffleV2 committed 1 year ago

Verified 3358c381

readme : update clblast instructions (#2903)

bandoti committed 1 year ago

Verified 52315a42

metal : show all Metal device instances in the system (#2952)

knweiss committed 1 year ago

Verified 8b56b4f2

k-quants : fix build on armv7 (android only) (#2920)

jhen0409 committed 1 year ago

Verified 21f3d1be

server : avoid aniprompt in probabilities of final response (#2849)

jhen0409 committed 1 year ago

Verified 571083f5

cuda : vsubss4 for older versions of ROCm/clang (#2942)

Engininja2 committed 1 year ago

Verified f04d0028

readme : quick start command fix (#2908)

gklab committed 1 year ago

Verified 69fdbb9a

Allow quantize to only copy tensors, some other improvements (#2931)

KerfuffleV2 committed 1 year ago

Verified 5d6f19f1

llama2c : rename function

ggerganov committed 1 year ago

Verified 0d589366

make : use unaligned vector moves on MinGW (#2945)

cebtenzzre committed 1 year ago

Verified 6c9c2342

minor : add const qualifiers (#2853)

mendax0110 committed 1 year ago

Verified ee8654bc

docs : add java-llama.cpp to README.md (#2935)

kherud committed 1 year ago

Verified 49bb9cbe

Older