ggerganov/llama.cpp

Pull Requests Commits

readme : refresh (#10587)

ggerganov committed 281 days ago

Verified abadba05

vulkan: Dynamic subgroup size support for Q6_K mat_vec (#10536)

netrunnereve committed 281 days ago

Verified 0533e7fb

ggml : move AMX to the CPU backend (#10570)

slaren committed 281 days ago

Verified 7cc2d2c8

server : add more test cases (#10569)

ngxson committed 281 days ago

Verified b782e5c7

imatrix : support combine-only (#10492)

robbiemu committed 281 days ago

Verified 3a8e9af4

cleanup UI link list (#10577)

slaren committed 281 days ago

Verified a3a3048e

ggml : fix I8MM Q4_1 scaling factor conversion (#10562)

ggerganov committed 282 days ago

Verified f0678c5f

ggml-cpu: fix typo in gemv/gemm iq4_nl_4_4 (#10580)

FanShupei committed 282 days ago

Verified 4b3242bb

sycl : offload of get_rows set to 0 (#10432)

Alcpz committed 282 days ago

Verified 0f77aae5

sycl : Reroute permuted mul_mats through oneMKL (#10408)

Alcpz committed 282 days ago

Verified 266b8519

CANN: RoPE operator optimization (#10563)

noemotiovon committed 282 days ago

Verified 938f6087

vulkan: get the first command buffer submitted sooner (#10499)

jeffbolznv committed 282 days ago

Verified f095a649

llava: return false instead of exit (#10546)

tinglou committed 282 days ago

Verified 678d7994

ggml : remove redundant copyright notice + update authors

ggerganov committed 282 days ago

Verified dc223440

llama : add missing model types

ggerganov committed 282 days ago

Verified 4c0a95b1

server : (tests) don't use thread for capturing stdout/stderr, bump openai client library (#10568)

ngxson committed 282 days ago

Verified 6c595676

common: fix warning message when no GPU found (#10564)

JohannesGaessler committed 282 days ago

Verified 89071931

docs: fix outdated usage of llama-simple (#10565)

rand-fly committed 283 days ago

Verified 7281cf13

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

slaren committed 283 days ago

Verified e90688ed

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

ggerganov committed 283 days ago

Verified 76b27d29

cmake : fix ARM feature detection (#10543)

ggerganov committed 283 days ago

Verified eea986f2

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

FanShupei committed 283 days ago

Verified c202cef1

kompute : improve backend to pass test_backend_ops (#10542)

slp committed 283 days ago

Verified 2025fa67

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

HRXWEB committed 283 days ago

Verified c6bc7395

CANN: Fix SOC_TYPE compile bug (#10519)

leo-pony committed 283 days ago

Verified 605fa66c

CANN: ROPE operator optimization (#10540)

noemotiovon committed 283 days ago

Verified b7420131

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

ngxson committed 283 days ago

Verified 9f912511

Add some minimal optimizations for CDNA (#10498)

IMbackK committed 283 days ago

Verified 3ad5451f

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

slaren committed 284 days ago

Verified 46c69e0e

metal : fix group_norm support condition (#0)

ggerganov committed 284 days ago

Verified 9e2301f4

Newer Older