ggml-org/llama.cpp

Pull Requests Commits

llama : fix type of KQ_mask and KQ_pos

ggerganov committed 2 years ago

Verified 3a468e6f

Merge branch 'master' into gg/flash-attn

ggerganov committed 2 years ago

Verified 9495d398

common : add HF arg helpers (#6234)

ggerganov committed 2 years ago

Verified 80bd33bc

llama : correction of the attn.v.weight quantization for IQ3_XS (#6209)

Nexesenex committed 2 years ago

Verified e80f06d2

tests : conditional python & node json schema tests (#6207)

ochafik committed 2 years ago

Verified f77a8ffd

json-schema-to-grammar : fix order of props + non-str const/enum (#6232)

ochafik committed 2 years ago

Verified 72114edf

cuda : add LLAMA_CUDA_NO_PEER_COPY to workaround broken ROCm p2p copy (#6208)

slaren committed 2 years ago

Verified 2f0e81e0

readme : add RecurseChat to the list of UIs (#6219)

xyc committed 2 years ago

Verified 29ab270e

server : fix n_keep always showing as 0 in response (#6211)

kaetemi committed 2 years ago

Verified 6b8bb3a3

server : enable continuous batching by default (#6231)

ggerganov committed 2 years ago

Verified 68e210b3

metal : proper assert for mat-mat memory alignment (#6225)

ggerganov committed 2 years ago

Verified b3e94f26

ci : add CURL flag for the mac builds (#6214)

Vaibhavs10 committed 2 years ago

Verified b2075fd6

metal : pad n_ctx by 32 (#6177)

ggerganov committed 2 years ago

Verified 95d576b4

add blog link (#6222)

NeoZhangJianyu committed 2 years ago

Verified 59c17f02

Fix params underscore convert to dash. (#6203)

dranger003 committed 2 years ago

Verified fa046eaf

server : update readme doc from `slot_id` to `id_slot` (#6213)

kaetemi committed 2 years ago

Verified be07a032

cuda : disable host register by default (#6206)

slaren committed 2 years ago

Verified d0a71233

Corrected typo to wrong file (#6199)

semidark committed 2 years ago

Verified f372c49c

tests : disable system() calls (#6198)

ggerganov committed 2 years ago

Verified 924ce1dc

cuda : fix LLAMA_CUDA_F16 build (#6197)

slaren committed 2 years ago

Verified 03a8f8fa

ggml : same IQ4_NL quantization for CPU/CUDA/Metal (#6196)

ikawrakow committed 2 years ago

Verified cfd3be76

json-schema-to-grammar improvements (+ added to server) (#5978)

ochafik committed 2 years ago

Verified 5b7b0ac8

ci : fix indentation error (#6195)

Vaibhavs10 committed 2 years ago

Verified 1943c019

build : add mac pre-build binaries (#6182)

Vaibhavs10 committed 2 years ago

Verified 5e43ba87

Add ability to use Q5_0, Q5_1, and IQ4_NL for quantized K cache (#6183)

ikawrakow committed 2 years ago

Verified 76aa30a2

Add nvidia and amd backends (#6157)

AidanBeltonS committed 2 years ago

Verified c5b8595e

cuda : fix conflict with std::swap (#6186)

slaren committed 2 years ago

Verified 42e21c68

cuda : print the returned error when CUDA initialization fails (#6185)

slaren committed 2 years ago

Verified 1c51f98a

llava : update MobileVLM-README.md (#6180)

ZiangWu-77 committed 2 years ago

Verified f9c7ba34

llava : add MobileVLM_V2 backup (#6175)

ZiangWu-77 committed 2 years ago

Verified 272935b2

Older