ngxson/llama.cpp

Pull Requests Commits

ngxson committed 2 years ago

e9313f2e

ngxson committed 2 years ago

2ab3da68

maybe fix AddressSanitizer?

ngxson committed 2 years ago

040fddee

ngxson committed 2 years ago

852f6548

Merge branch 'master' into xsn/slot_state_machine

ngxson committed 2 years ago

ba0065fc

ngxson committed 2 years ago

e8e3e725

metrics : add n_busy_slots_per_decode

ngxson committed 2 years ago

69b398ce

fix passkey test

ngxson committed 2 years ago

fbebf650

add missing notify_one

ngxson committed 2 years ago

d3fedaa6

pop_deferred_task

ngxson committed 2 years ago

ec882cc1

readme : refactor API section + remove old hot topics

ggerganov committed 2 years ago

Verified b69a480a

add SLOT_STATE_DONE_PROMPT

ngxson committed 2 years ago

446d57d7

server : simplify state machine for slot

ngxson committed 2 years ago

2c81cde4

server : test script : add timeout for all requests (#9282)

ngxson committed 2 years ago

Verified 48baa61e

src: make tail invalid when kv cell is intersection for mamba (#9249)

kylo5aby committed 2 years ago

Verified f1485161

docker : fix missing binaries in full-cuda image (#9278)

slaren committed 2 years ago

Verified 048de848

ggml : add pthread includes on FreeBSD (#9258)

yurivict committed 2 years ago

Verified f771d064

server : refactor multitask handling (#9274)

ngxson committed 2 years ago

Verified 6e7d133a

llama-cli : remove duplicated log message (#9275)

nbcsm committed 2 years ago

Verified b60074f1

build(nix): Package gguf-py (#5664)

ditsuke committed 2 years ago

Verified 9c1ba557

llama : minor style

ggerganov committed 2 years ago

Verified c6d4cb46

llama : support RWKV v6 models (#8980)

MollySophia committed 2 years ago

Verified 8f1d81a0

nix: fix CUDA build - replace deprecated autoAddOpenGLRunpathHook

enolan committed 2 years ago

a47667cf

sgemm : improved Q4_0 and Q8_0 performance via 4xN and Mx4 gemm (#8908)

Srihari-mcw committed 2 years ago

Verified ea5d7478

llama : fix typo in xcda_array_view comment [no ci] (#9132)

danbev committed 2 years ago

Verified 49271efb

llama : fix llama_split_mode enum values in main_gpu document (#9057)

kou committed 2 years ago

Verified 0ab30f8d

Correct typo run_llama2.sh > run-llama2.sh (#9149)

MakeDecisionWorth committed 2 years ago

Verified cddae488

llava : the function "clip" should be int (#9237)

tc-mb committed 2 years ago

Verified 7ea8d80d

Threadpool: take 2 (#8672)

fmz committed 2 years ago

Verified 42c76d13

server : fix crash when error handler dumps invalid utf-8 json (#9195)

kaetemi committed 2 years ago

9f7d4bcf

Older