ggml-org/llama.cpp

Pull Requests Commits

server: benchmark: chat/completions scenario and other llm servers comparison (#5941)

phymbert committed 1 year ago

Verified 621e86b3

server : print chat template info

ggerganov committed 1 year ago

Verified 77d1ac7e

perplexity : support using multiple sequences to allow larger batch sizes (#5946)

slaren committed 1 year ago

Verified d894f352

readme : update hot topics

ggerganov committed 1 year ago

Verified 098dbaab

ggml : fix unnecessary f32 -> f16 -> f32 casts (mmla) (#5951)

ggerganov committed 1 year ago

Verified 8380ecfb

server : fix metrics init (#5964)

ggerganov committed 1 year ago

Verified 58308a0e

ggml : remove old quantization functions (#5942)

ggerganov committed 1 year ago

Verified 5b097973

server : clarify some items in the readme (#5957)

ggerganov committed 1 year ago

Verified 97c09585

server : normalize embeddings (#5956)

redlion0929 committed 1 year ago

Verified fb215c38

tests : gitignore ggml-common.h

ggerganov committed 1 year ago

Verified 2c4f566c

server : fix passing prompt as tokens (#5955)

z80maniac committed 1 year ago

Verified 0db32bea

ggml : add ggml-common.h to deduplicate shared code (#5940)

ggerganov committed 1 year ago

Verified 8a3012a4

server : simplify logic for empty prompts (#5953)

ggerganov committed 1 year ago

Verified 9674aaf3

Server: reorganize some http logic (#5939)

ngxson committed 1 year ago

Verified 950ba1ab

server : add SSL support (#5926)

gabe-l-hart committed 1 year ago

Verified e1fa9569

server: tests: add truncated prompt tests, better kv cache size (#5933)

phymbert committed 1 year ago

Verified fd72d2d2

llama : support Mamba Selective State Space Models (#5328)

compilade committed 1 year ago

Verified c2101a2e

llama : fix quantization of shared token_embd (#5944)

compilade committed 1 year ago

Verified 515f7d0d

server: metrics: add llamacpp:prompt_seconds_total and llamacpp:tokens_predicted_seconds_total, reset bucket only on /metrics. Fix values cast to int. Add Process-Start-Time-Unix header. (#5937)

phymbert committed 1 year ago

Verified 76e86882

llama : assume tied weights if lm_head/output weights is missing (#5824)

dmahurin committed 1 year ago

Verified e457fb35

server : fix EOS token detection with disabled cache (#5938)

ggerganov committed 1 year ago

Verified af37fd8b

log : fix MSVC compile errors (#5643)

uextm committed 1 year ago

Verified 581ed5c4

llama-bench : add embeddings option (#5924)

ggerganov committed 1 year ago

Verified 6cdabe65

Revert "[SYCL] fix error when set main gpu to non-zero (#5901)" (#5918)

NeoZhangJianyu committed 1 year ago

Verified 89fb735f

server : add `/v1/completions` endpoint (#5914)

mscheong01 committed 1 year ago

Verified 55a2a900

server : refactor (#5882)

ggerganov committed 1 year ago

Verified 2002bc96

[SYCL] fix error when set main gpu to non-zero (#5901)

NeoZhangJianyu committed 1 year ago

Verified ceca1aef

ggml : use SYS_get_cpu if SYS_getcpu is not defined (#5906)

cebtenzzre committed 1 year ago

Verified e04e04f8

ggml : use `uint8x16_t` return type for `ggml_vqtbl1q_u8` (#5894)

bobqianic committed 1 year ago

Verified e25fb4b1

convert : remove AWQ remnants (#5768)

ggerganov committed 1 year ago

Verified 1e35d619

Newer Older