ggerganov/llama.cpp

Pull Requests Commits

preset: allow named remote preset (#18728)

ngxson committed 21 days ago

Verified 23f82f24

docs(ggml): update backend ops (#18734)

taronaeo committed 21 days ago

Verified 2656c0d2

Corrected: changed s13 = src1->nb[3] instead of nb[2] (#18724)

michaelw9999 committed 21 days ago

Verified 600a3664

common : add --license to display embedded licenses (#18696)

angt committed 21 days ago

Verified ea23c159

server: fix n_cmpl not skipping processing prompt (#18663)

ngxson committed 22 days ago

Verified 9ac2693a

mtmd: Add Gemma3n multimodal support with MobileNetV5 vision encoder (#18256)

simrnsingh committed 22 days ago

Verified a61c8bc3

opencl: add EXPM1 op (#18704)

shaofeiqi committed 22 days ago

Verified 593da7fa

Updates to webgpu get_memory (#18707)

reeselevine committed 22 days ago

Verified 9e41884d

Webui/file upload (#18694)

ServeurpersoCom committed 22 days ago

Verified ec8fd787

cmake: only build cli when server is enabled (#18670)

AsbjornOlling committed 22 days ago

Verified a180ba78

server : fix timing of prompt/generation (#18713)

ggerganov committed 22 days ago

Verified 53eb9435

scripts : pr2wt.sh reset to remote head (#18695)

ggerganov committed 22 days ago

Verified d3435efc

server : use different seeds for child completions (#18700)

ggerganov committed 22 days ago

Verified f5f8812f

common: support remote preset (#18520)

ngxson committed 23 days ago

Verified 8ece3836

llama: use host memory if device reports 0 memory (#18587)

taronaeo committed 23 days ago

Verified 046d5fd4

ggml-webgpu: Fix GGML_MEM_ALIGN to 8 for emscripten. (#18628)

yomaytk committed 23 days ago

Verified 480160d4

ggml webgpu: initial flashattention implementation (#18610)

reeselevine committed 23 days ago

Verified 15bff84b

vulkan: fix push constant size for quantize_q8_1 (#18687)

jeffbolznv committed 23 days ago

Verified 2524c261

vulkan: optimize ssm_scan (#18630)

jeffbolznv committed 23 days ago

Verified cb14b069

vendor : update cpp-httplib to 0.30.0 (#18660)

angt committed 23 days ago

Verified 55abc393

scripts : support chaining commands in pr2wt.sh (#18671)

ggerganov committed 23 days ago

Verified f2f6c880

metal : add MoE kernel specialization for ne20=5 (#18667)

dororodoroddo committed 23 days ago

Verified 945bf106

llama-fit-params: free memory target per device (#18679)

JohannesGaessler committed 23 days ago

Verified 64848deb

ggml: add env var GGML_OP_OFFLOAD_MIN_BATCH (#18535)

DocShotgun committed 23 days ago

Verified 9a5724de

model-conversion : add warn about transformers mismatch (#18691)

danbev committed 23 days ago

Verified 9c142e3a

model-conversion : remove -st targets for converted model (#18689)

danbev committed 23 days ago

Verified df7fb921

llama : add `use_direct_io` flag for model loading (#18166)

JTischbein committed 24 days ago

Verified 2038101b

opencl: add FILL op support (#18682)

shaofeiqi committed 24 days ago

Verified 568371a7

scripts : fix repos cloned with .git extension (#18669)

CISC committed 24 days ago

Verified 5b8844ae

convert : more variants of rope_theta config entries (#18668)

CISC committed 24 days ago

Verified 7e16fef0

Newer Older