ngxson/llama.cpp

Pull Requests Commits

metal: use shared buffers on eGPU (#17866)

jdemeule committed 23 days ago

Verified 165caaf5

mtmd: refactor audio preprocessing (#17978)

ngxson committed 24 days ago

Verified 96a181a9

cli: fixed dead links to tools/main for cli and completion, fixed code owners (#17993)

andrew-aladev committed 24 days ago

Verified 4a4f7e65

webui: add "delete all conversations" button to import/export tab (#17444)

thomasjfox committed 24 days ago

Verified e73d5486

llama: automatically set parameters not set by the user in such a way that maximizes GPU utilization (#16653)

JohannesGaessler committed 24 days ago

Verified b1f3a6e5

[SYCL] Support gpt-oss by OPs add-id, mul_mat for mxfp4, swiglu_oai (#17826)

NeoZhangJianyu committed 24 days ago

Verified 4aced7a6

model : add glm-asr support (#17901)

piDack committed 24 days ago

Verified 745fa0e7

preset: handle negated arg, reverse the meaning if needed (#18041)

ngxson committed 24 days ago

Verified 52392291

convert : refactor rope scaling handling (#18013)

CISC committed 24 days ago

Verified 5c8a7171

mtmd: enhance image resizing in llava_uhd (#18014)

bluebread committed 24 days ago

Verified 37f5a109

vulkan: fix mul_mat_vec_iq1_s formatting (#18026)

0cc4m committed 24 days ago

Verified 9e6649ec

graph: add f_attn_temp_offset (#18025)

ngxson committed 25 days ago

Verified 0759b09c

common : refactor common_sampler + grammar logic changes (#17937)

ggerganov committed 25 days ago

Verified 254098a2

vulkan: Fix data race/hang in scalar/cm1 flash attention (#17887)

jeffbolznv committed 25 days ago

Verified 3238b140

vulkan: improve mul_mat_vec_iq1_s speed (#17874)

lovedheart committed 25 days ago

Verified 47226716

vulkan: faster q6_k matmul (#17813)

netrunnereve committed 25 days ago

Verified d15d177f

model-conversion : cast logits to float32 (#18009)

ggerganov committed 25 days ago

Verified 77ad8542

models : fix YaRN regression + consolidate logic (#18006)

ggerganov committed 25 days ago

Verified 609a2d02

ggml : arm repack fix build

ggerganov committed 25 days ago

a63cbafb

ggerganov committed 25 days ago

0e592249

ggml : arm repack fix build (whisper/0)

ggerganov committed 25 days ago

71fdcf06

cmake : set `CMAKE_RUNTIME_OUTPUT_DIRECTORY` for non standalone build (ggml/1394)

HerrCai0907 committed 25 days ago

615655aa

scripts: add script to compare logprobs of llama.cpp against other frameworks (#17947)

ngxson committed 25 days ago

Verified c00ff929

server-models.cpp: add missing <filesystem> (#18000)

barracuda156 committed 25 days ago

Verified 4ed2bae5

llama_context: synchronize before reallocating output buffer (#17974)

jeffbolznv committed 25 days ago

Verified 5266379b

arg: fix common_params_parse not accepting negated arg (#17991)

ngxson committed 26 days ago

Verified 4d5ae24c

cmake: correct scope - link ws2_32 for MinGW/w64devkit builds in cpp-httplib (#17972)

gustrd committed 26 days ago

Verified 66ba5125

vulkan: support get_rows for i32 (#17941)

jeffbolznv committed 26 days ago

Verified 36255a22

vulkan: support GGML_OP_DIAG (#17893)

jeffbolznv committed 26 days ago

Verified 3229a23f

vulkan: Multi-pass softmax for large number of cols (#17892)

jeffbolznv committed 26 days ago

Verified 303f8615

Newer Older