ngxson/llama.cpp

Pull Requests Commits

ngxson committed 1 year ago

59e60910

fix rms_norm_act_2d

ngxson committed 1 year ago

3161c0c8

add MultiQueryAttentionBlock

ngxson committed 1 year ago

325cbe76

ngxson committed 1 year ago

96fd71dd

conversion script

ngxson committed 1 year ago

f8f73e86

ngxson committed 1 year ago

f4e70342

fix f_attention_scale

ngxson committed 1 year ago

d42071e2

add kq weighted norm

ngxson committed 1 year ago

e5fe2154

update text conversion

ngxson committed 1 year ago

d8589f8b

fix merge conflict

ngxson committed 1 year ago

4d3d0aeb

Merge branch 'master' into gemma3n_text

ngxson committed 1 year ago

4bfbfa1c

server : add server parameters for draft model cache type (#13782)

aa956 committed 1 year ago

Verified d67341dc

build : suppress gcc15 compile warnings (#14261)

fanyang89 committed 1 year ago

Verified 456af35e

sycl: Cleanup codepaths in Get Rows in sycl backend (#14215)

ShanoToni committed 1 year ago

Verified 600e3e9b

llama-bench : add --no-warmup flag (#14224) (#14270)

bashayer-alrumahi committed 1 year ago

Verified fffcce53

convert : fix remote option in Windows (#14100)

pqnet committed 1 year ago

Verified 5fc78568

llamafile : support s390x SIMD instruction set (#14273)

taronaeo committed 1 year ago

Verified faed5a5f

Vulkan: Set device max size for host memory to avoid OOM warning and fallback to CPU buffer (#14249)

0cc4m committed 1 year ago

Verified 10bb545c

memory : Hybrid recurrent cache (#13979)

gabe-l-hart committed 1 year ago

Verified edc4a29e

metal : add mean kernel (#14267)

ggerganov committed 1 year ago

Verified ed3290ab

docs: add s390x build documentation (#14264)

taronaeo committed 1 year ago

Verified 8d947136

ggml-cpu: reduce asm calls for hsum (#14037)

taronaeo committed 1 year ago

Verified 50d22279

ggml-cpu: fix uncaught underscore terminators (#14023)

taronaeo committed 1 year ago

Verified 6231c5cd

ggml: Add Apple support for GGML_CPU_ALL_VARIANTS (#14258)

chaxu01 committed 1 year ago

Verified ef035803

mtmd : refactor llava-uhd preprocessing logic (#14247)

ngxson committed 1 year ago

Verified 413977de

llama-chat : fix multiple system message for gemma, orion (#14246)

ngxson committed 1 year ago

Verified 95402553

convert : fix null head_dim AutoConfig regression (#14248)

CISC committed 1 year ago

Verified 3865cff4

ggerganov committed 1 year ago

d03172cc

ggml : disable warnings for tests when using MSVC (ggml/1273)

danbev committed 1 year ago

dd8e59f4

ggml : remove unused ggml_context_container (ggml/1272)

danbev committed 1 year ago

bbe98d27

Older