ngxson/llama.cpp

Pull Requests Commits

server : add cache reuse card link to help (#13230)

ggerganov committed 1 year ago

Verified fab647e8

convert : explicitly disable trust_remote_code for AutoConfig (#13246)

ngxson committed 1 year ago

Verified dcf88600

ci: fix cross-compile sync issues (#12804)

mtmcp committed 1 year ago

Verified d24d5928

rpc : avoid uninitialized memory in serialize_tensor (#13210)

justinsb committed 1 year ago

Verified 8efbdadc

ggml: Don't assert fail when tensor data changes (#13222)

jessegross committed 1 year ago

Verified f057808f

build : fix build info on windows (#13239)

slaren committed 1 year ago

Verified d7a14c42

clip : (minicpmv) Re-enable upscaling of images smaller than the CLIP image size (#13237)

lcarrere committed 1 year ago

Verified b6e4ff69

llama-chat : update GLM4 chat template (#13238)

matteoserva committed 1 year ago

Verified e0f572c8

vulkan: Add bfloat16 support (#12554)

jeffbolznv committed 1 year ago

Verified 79f26e9e

vulkan: Handle src1 batch dimension in non-contiguous mat-vec-mul shader (#13191)

jeffbolznv committed 1 year ago

Verified fc727bcd

test: non-cont. b in test-backend-ops -o MUL_MAT (#13187)

JohannesGaessler committed 1 year ago

Verified b0ecbd43

ggerganov committed 1 year ago

b1dd4d08

whisper : add check that target name exists (whisper/3103)

danbev committed 1 year ago

99881f77

ggml : suppress Windows compiler warnings (whisper/3075)

danbev committed 1 year ago

b5769d92

mtmd : add **vision** support for Mistral Small 3.1 (#13231)

ngxson committed 1 year ago

Verified 8936784f

arg : remove CURLINFO_EFFECTIVE_METHOD (#13228)

ngxson committed 1 year ago

Verified 13c9a331

llama-model : fix the reported size class for nomic-embed-text-v2-moe (#13223)

cebtenzzre committed 1 year ago

Verified a70183eb

ggerganov committed 1 year ago

8d33d740

ggml : fix ggml_gallocr_ptr type (ggml/1205)

slaren committed 1 year ago

4254bb49

cuda : fix unused variable compile warning (whisper/0)

ggerganov committed 1 year ago

99985401

CUDA: batched+noncont MMQ, refactor bs>1 MoE code (#13199)

JohannesGaessler committed 1 year ago

Verified e1e8e099

arg : -hf do not fail if url mismatch (#13219)

ngxson committed 1 year ago

Verified 6f67cf1f

fix typo: `n_ctx_pre_seq` -> `n_ctx_per_seq` (#13221)

ddh0 committed 1 year ago

Verified 16a457fa

convert : improve model arch handling (#13122)

ngxson committed 1 year ago

Verified 3e168bed

llava : remove duplicate include (#13207)

tattn committed 1 year ago

Verified ceda28ef

common : add -jf / --json-schema-file flag (#12011)

ochafik committed 1 year ago

Verified 3b127c73

vulkan: use uint array index to avoid glslang bug (#13193)

jeffbolznv committed 1 year ago

Verified e5007a5e

ggml : fix ppc64le build (#13176)

shalinib-ibm committed 1 year ago

Verified 41631377

convert : correct typo image_mean --> image_std (#13208)

ngxson committed 1 year ago

Verified 07c2e2f7

feat(ggml-cpu): enable z17 compile (#13182)

taronaeo committed 1 year ago

Verified 44cd8d91

Newer Older