ggerganov/llama.cpp

Pull Requests Commits

ui : models discover dialog

allozaur committed 31 days ago

5f2d7ed2

ui : render shared model row hints as native titles

allozaur committed 31 days ago

516afbb9

ui : remember hub avatars that failed to load

allozaur committed 31 days ago

3870f158

ui : shared model display primitives

allozaur committed 31 days ago

7f1c9236

ui : optional sanitized raw HTML in markdown

allozaur committed 31 days ago

f0127f0c

ui : model download pipeline

allozaur committed 31 days ago

85ecd8ce

ui : model memory-fit estimation

allozaur committed 31 days ago

31fc7a0f

ui : strip provider tilde prefix from hub avatar urls

allozaur committed 31 days ago

61ccf5b8

ui : Hugging Face Hub data layer

allozaur committed 31 days ago

8e37618a

ui : model id grammar for sidecars, quants and capability parsing

allozaur committed 31 days ago

774b215c

ui : type-safe API types, fetch helpers and download-ready models store plumbing

allozaur committed 31 days ago

b5f7cbeb

server : fix deadlock when removing a finished download

allozaur committed 31 days ago

d1cfcee6

common : resolve <quant>-<sidecar> download tags and list cached sidecars

allozaur committed 31 days ago

95cf88dd

context : do not re-reserve the scheduler when toggling causal_attn (#28751)

sihanyu03 committed 31 days ago

Verified ed7ac35e

Enables Windows ARM64 build with MSVC cl.exe (#28362)

sarahwu185 committed 31 days ago

Verified 0c6a6a7c

tests : fix ggml init (#29554)

ggerganov committed 31 days ago

Verified 81ef10ea

vulkan: fix wrong results when a mul_mat reads a slice of a larger cache (#28956)

ServeurpersoCom committed 31 days ago

Verified 52624716

server : allow RANK pooling batch splitting for causal LLM rerankers (ie. Qwen3 and Qwen3-VL) (#28876)

timothywang21 committed 31 days ago

Verified 4da63377

common : avoid side effects around params parsing (#29537)

ggerganov committed 31 days ago

Verified a97cce86

common : make string_split<T> throw on invalid values (#29518)

angt committed 31 days ago

Verified 136887b6

convert : export YaRN scaling parameters for PLaMo-3 (#29528)

tokinasin committed 32 days ago

Verified 9adc7f42

ci : bump ty to 0.0.84 (#29529)

CISC committed 32 days ago

Verified 6fd50a40

jinja : add support for dict builtin (#29477)

CISC committed 32 days ago

Verified 33c923db

opencl: refine bin kernel loading condition (#29503)

lhez committed 32 days ago

Verified c9064dde

sycl: FWHT kernels for block widths above 512 (#29243)

bri-prism committed 32 days ago

Verified c8296709

CUDA: tune fp16 tile FlashAttention configs for head sizes 40-112 (#26289)

animeshsri14 committed 32 days ago

Verified 36d7b083

HIP: Enable fattn-mma kernel on cdna for dkq > 256 for large batch sizes (#28907)

IMbackK committed 32 days ago

Verified 2ebd9ae6

vulkan: fix argsort kernel selection for Adreno (#29469)

0cc4m committed 32 days ago

Verified cea74625

common : throw instead of abort on grammar without llguidance (#29516)

angt committed 32 days ago

Verified da6c28eb

RPC: use RDMA completion channel to not spin (#29440)

am17an committed 32 days ago

Verified d7fb90e8

Older