ngxson/llama.cpp

Pull Requests Commits

also correct --pca-iter

ngxson committed 1 year ago

cc6eec0c

cvector: fix CI + correct help message

ngxson committed 1 year ago

bc8cb239

codecov : remove (#8004)

ggerganov committed 1 year ago

Verified a04a953c

[SYCL] refactor (#6408)

airMeng committed 1 year ago

Verified 623494a4

tokenizer : BPE fixes (#7530)

jaime-m-p committed 1 year ago

Verified 37bef894

Only use FIM middle token if it exists (#7648)

CISC committed 1 year ago

Verified 91c188d6

Fix no gcc pragma on Windows (#7751)

jojorne committed 1 year ago

Verified 84f6de17

Allow compiling with CUDA without CUDA runtime installed (#7989)

drepper committed 1 year ago

Verified 61665277

chore: clean useless beam search param (#7985)

thxCode committed 1 year ago

Verified b96f9afb

readme : update UI list (#7943)

abgulati committed 1 year ago

Verified 11937781

ggerganov committed 1 year ago

Verified 5326bcce

whisper : use ggml_backend_sched (whisper/2239)

ggerganov committed 1 year ago

Verified e6ecc2be

update: support Qwen2-57B-A14B (#7835)

legraphista committed 1 year ago

Verified a94e6ff8

Make updates to type cast based on compiler instead of OS (#7851)

Srihari-mcw committed 1 year ago

Verified 5b6da187

llama : disable FA if KV head size do not match (#7982)

ggerganov committed 1 year ago

Verified 7c26775a

Add Nix and Flox install instructions (#7899)

bryanhonof committed 1 year ago

Verified b473e950

sched : offload_op also requires supports_op (#7977)

slaren committed 1 year ago

Verified 99052cd2

fix: divide 0 exception in mamba (#7932)

thxCode committed 1 year ago

Verified c637fcd3

Implement non-mapped async IO for CUDA on Windows. (#7896)

mtavenrath committed 1 year ago

Verified 6a2f0b34

rpc : fix load/store misaligned addresses (#7948)

ggerganov committed 1 year ago

Verified 21be9cab

gguf-dump.py: add --markdown dump output (#7853)

mofosyne committed 1 year ago

Verified 006167aa

[SYCL] Update README-sycl.md for Chapter "Recommended release" and "News" (#7946)

arthw committed 1 year ago

Verified df68d4fa

Add support for sqrt on CUDA (#7953)

calvin-laurenson committed 1 year ago

Verified 43b35e38

cuda : fix bounds check for src0 rows in MMVQ kernel (whisper/2231)

ggerganov committed 1 year ago

Verified 19b7a836

ggml : fix and optimize ppc64le (ggml/849)

penghongbo committed 1 year ago

Verified b5fcf8ef

ggml : remove duplicate include of ggml-common.h (ggml/853)

danbev committed 1 year ago

Verified 398105ff

flake.lock: Update (#7951)

ggerganov committed 1 year ago

Verified bc6c457f

unicode : avoid char32_t (#7957)

ggerganov committed 1 year ago

Verified 52399254

readme : update UI list [no ci] (#7958)

hopkins385 committed 1 year ago

Verified 6fe1c627

ggml : fix handling of zero blocks in IQ quants (#7955)

ggerganov committed 1 year ago

Verified cddaf028

Older