ngxson/llama.cpp

Pull Requests Commits

bump openai to 1.55.3

ngxson committed 1 year ago

8a349b31

test: bump openai to 1.55.2

ngxson committed 1 year ago

e59bed91

server : (tests) don't use thread for capturing stdout/stderr

ngxson committed 1 year ago

7439ba7b

docs: fix outdated usage of llama-simple (#10565)

rand-fly committed 1 year ago

Verified 7281cf13

ci : fix tag name in cuda and hip releases (#10566)

slaren committed 1 year ago

Verified e90688ed

ggml : fix row condition for i8mm kernels (#10561)

ggerganov committed 1 year ago

Verified 76b27d29

cmake : fix ARM feature detection (#10543)

ggerganov committed 1 year ago

Verified eea986f2

ggml-cpu: support IQ4_NL_4_4 by runtime repack (#10541)

FanShupei committed 1 year ago

Verified c202cef1

kompute : improve backend to pass test_backend_ops (#10542)

slp committed 1 year ago

Verified 2025fa67

CANN: Update cann.md to display correctly in CLion (#10538)

HRXWEB committed 1 year ago

Verified c6bc7395

CANN: Fix SOC_TYPE compile bug (#10519)

leo-pony committed 1 year ago

Verified 605fa66c

CANN: ROPE operator optimization (#10540)

noemotiovon committed 1 year ago

Verified b7420131

common : fix duplicated file name with hf_repo and hf_file (#10550)

ngxson committed 1 year ago

Verified 9f912511

Add some minimal optimizations for CDNA (#10498)

IMbackK committed 1 year ago

Verified 3ad5451f

ci : faster CUDA toolkit installation method and use ccache (#10537)

slaren committed 1 year ago

Verified 46c69e0e

metal : fix group_norm support condition (#0)

ggerganov committed 1 year ago

Verified 9e2301f4

ggerganov committed 1 year ago

Verified fee824a1

Do not include arm_neon.h when compiling CUDA code (ggml/1028)

frankier committed 1 year ago

Verified 9150f8fe

vulkan: define all quant data structures in types.comp (#10440)

jeffbolznv committed 1 year ago

Verified c31ed2ab

vulkan: Handle GPUs with less shared memory (#10468)

jeffbolznv committed 1 year ago

Verified 5b3466be

vulkan: further optimize q5_k mul_mat_vec (#10479)

jeffbolznv committed 1 year ago

Verified 249a7902

vulkan: skip integer div/mod in get_offsets for batch_idx==0 (#10506)

jeffbolznv committed 1 year ago

Verified 71a64989

vulkan: optimize Q2_K and Q3_K mul_mat_vec (#10459)

jeffbolznv committed 1 year ago

Verified 4a57d362

ci : fix cuda releases (#10532)

slaren committed 1 year ago

Verified c9b00a70

Add OLMo 2 model in docs (#10530)

2015aroras committed 1 year ago

Verified de509735

ci : remove nix workflows (#10526)

slaren committed 1 year ago

Verified 5a349f28

llama : disable warnings for 3rd party sha1 dependency (#10527)

slaren committed 1 year ago

Verified 30ec3983

Fix HIP flag inconsistency & build docs (#10524)

tristandruyen committed 1 year ago

Verified be0e350c

mtgpu: Add MUSA_DOCKER_ARCH in Dockerfiles && update cmake and make (#10516)

yeahdongcn committed 1 year ago

Verified 249cd93d

vulkan: fix group_norm (#10496)

jeffbolznv committed 1 year ago

Verified 904109ed

Older