ggml-org/llama.cpp

Pull Requests Commits
convert : refactor vocab selection logic (#6355)
cebtenzzre committed 2 years ago
Verified be55134a
llava : fix MobileVLM (#6364)
ZiangWu-77 committed 2 years ago
Verified 66ba5602
llama : fix command-r inference when omitting outputs (#6367)
compilade committed 2 years ago
Verified 0308f5e3
ci: bench: fix master not schedule, fix commit status failed on external repo (#6365)
phymbert committed 2 years ago
Verified 28cb9a09
doc: fix outdated default value of batch size (#6336)
Sunt-ing committed 2 years ago
Verified cfc4d75d
server : stop gracefully on SIGTERM (#6348)
EZForever committed 2 years ago
Verified 6902cb7f
nix: removed unnessesary indentation
hutli committed 2 years ago
d2d8f389
nix: moved blas availability check to package inputs so it is still overridable
hutli committed 2 years ago
d39b308e
using blas.meta.available to check host platform
hutli committed 2 years ago
c8739766
only using explicit blas if hostPlatform is allowed
hutli committed 2 years ago
dbb03e2b
nix: .#windows: proper cross-compilation set-up
SomeoneSerge committed 2 years ago
e9f17dc3
nix: package: don't introduce the dependency on python
SomeoneSerge committed 2 years ago
22a462cc
nix: .#widnows: init
hutli committed 2 years ago
f6a0f5c6
doc: fix typo in MobileVLM-README.md (#6181)
ZiangWu-77 committed 2 years ago
Verified d0e2f641
[SYCL] fix set main gpu crash (#6339)
NeoZhangJianyu committed 2 years ago
Verified 25f4a613
server: continuous performance monitoring and PR comment (#6283)
phymbert committed 2 years ago
Verified a016026a
nix: ci: dont test cuda and rocm (for now)
SomeoneSerge committed 2 years ago
53c7ec53
ggml : fix bounds checking of zero size views (#6347)
slaren committed 2 years ago
Verified e5b89a44
make : whitespace
ggerganov committed 2 years ago
Verified 3a034597
embedding : show full embedding for single prompt (#6342)
howlger committed 2 years ago
Verified 1e13987f
[SYCL] Fix batched impl for NVidia GPU (#6164)
AidanBeltonS committed 2 years ago
Verified e82f9e2b
Make IQ1_M work for QK_K = 64 (#6327)
ikawrakow committed 2 years ago
Verified cbc83436
common : change --no-penalize-nl to --penalize-nl (#6334)
CISC committed 2 years ago
Verified e562b971
llama2c : open file as binary (#6332)
ggerganov committed 2 years ago
Verified 2ab4f00d
readme : add php api bindings (#6326)
mcharytoniuk committed 2 years ago
Verified 1740d6dd
server: public: use relative routes for static files (#6325)
EZForever committed 2 years ago
Verified 0642b22c
[SYCL] fix no file in win rel (#6314)
NeoZhangJianyu committed 2 years ago
Verified a4f569e8
wpm : portable unicode tolower (#6305)
cebtenzzre committed 2 years ago
Verified 32c8486e
llama : greatly reduce output buffer memory usage (#6122)
compilade committed 2 years ago
Verified 557410b8
IQ1_M: 1.75 bpw quantization (#6302)
ikawrakow committed 2 years ago
Verified 55c1b2a3
Terms Of Use Privacy Policy Refund Policy Impressum