ngxson/llama.cpp

Pull Requests Commits

ngxson committed 1 year ago

89c2af90

add "model_path" to /props

ngxson committed 1 year ago

1949f68f

fix CI by adding safe_json_to_str

ngxson committed 1 year ago

65d2e6d6

remove task inf_type

ngxson committed 1 year ago

090a1134

add tests for /props and /slots

ngxson committed 1 year ago

e721f4c6

Merge branch 'master' into xsn/refactor_server_struct_input

ngxson committed 1 year ago

6bf6e306

ggml : disable iq4_nl interleave size 8 (#10709)

ggerganov committed 1 year ago

Verified d9c3ba2b

server : various fixes (#10704)

ggerganov committed 1 year ago

Verified ce4a7b84

add test for slots endpoint

ngxson committed 1 year ago

9bb1ae6b

ggml : refactor online repacking (#10446)

Djip007 committed 1 year ago

Verified 19d8762a

server : fix free of spec context and batch (#10651)

ggerganov committed 1 year ago

Verified c2a16c0b

Vulkan: VK_KHR_cooperative_matrix support to speed up prompt processing (#10597)

0cc4m committed 1 year ago

Verified 3df784b3

metal : Extend how Llama.cpp locates metal resources (#10676)

ormandi committed 1 year ago

Verified 86a19349

convert : add support for Roberta embeddings (#10695)

Ssukriti committed 1 year ago

Verified 784a14aa

convert : add custom attention mapping

ggerganov committed 1 year ago

Verified c5ede384

server : (refactor) no more json in server_task input

ngxson committed 1 year ago

db97c8b1

common : bring back --no-warmup to server (#10686)

ngxson committed 1 year ago

Verified f162d45a

server : (refactoring) do not rely on JSON internally (#10643)

ngxson committed 1 year ago

Verified 6c5bc062

fix(server) : not show alert when DONE is received (#10674)

pminev committed 1 year ago

Verified 7736837d

vulkan: Add VK_NV_cooperative_matrix2 support for mul_mat and flash attention (#10206)

jeffbolznv committed 1 year ago

Verified c9c6e01d

llama : add Minerva 7B model support (#10673)

Riccorl committed 1 year ago

Verified 6fe62478

ggerganov committed 1 year ago

Verified 0cd182eb

ggml: add `GGML_SET` Metal kernel + i32 CPU kernel (ggml/1037)

PABannier committed 1 year ago

Verified a8cbab20

ggml : add `GGML_PAD_REFLECT_1D` operation (ggml/1034)

PABannier committed 1 year ago

Verified c2082d93

py : update outdated copy-paste instructions [no ci] (#10667)

danbev committed 1 year ago

Verified d405804b

Update deprecation-warning.cpp (#10619)

aryantandon01 committed 1 year ago

Verified f112d198

server : fix speculative decoding with context shift (#10641)

ggerganov committed 1 year ago

Verified 1da7b765

ggml : add predefined list of CPU backend variants to build (#10626)

slaren committed 1 year ago

Verified 59f4db10

ggml-cpu : fix HWCAP2_I8MM value (#10646)

slaren committed 1 year ago

Verified 28035408

Fix HF repo commit to clone lora test models (#10649)

ltoniazzi committed 1 year ago

Verified 253b7fde

Older