ggerganov/llama.cpp

Pull Requests Commits

ggerganov committed 24 days ago

72fef3d6

vulkan: replace std::aligned_alloc with posix_memalign for termux compatibility (#30828)

bill88t committed 24 days ago

Verified d7725699

nemotron : support non-standard GQA, BPE tokenizer and domyn-small pre-type (#23343)

martin-cimmino committed 24 days ago

Verified 0693dbfb

server : support typed content input for RERANK endpoint (#29625)

timothywang21 committed 24 days ago

Verified a5d22276

ci : disable slow WebGPU tests (#30963)

ggerganov committed 24 days ago

Verified 5f5acfe2

ci : be able to reconfigure HTTP port for server pytest workers (#30972)

ggerganov committed 24 days ago

Verified 1f4c2eea

sycl: fix CONV_2D wrong results for non-contiguous input (#30363)

mctylr-gh committed 24 days ago

Verified 7170cad4

vocab : add clean_spaces GGUF key (#30940)

tkran2 committed 24 days ago

Verified d88040c7

vocab : use Tekken pre-tokenizer for pixtral rather than the Llama 3 one (#30960)

tkran2 committed 24 days ago

Verified e197cef4

models : Record layer inputs for all granite models (#30957)

gabe-l-hart committed 24 days ago

Verified bf039c84

kleidiai : add Q8_0 SVE dotprod GEMV and SVE i8mm GEMM kernels (#30808)

chaxu01 committed 24 days ago

Verified dfeea520

sycl: merge the fp16 and fp32 dequant switches (#30740)

Titaniumtown committed 24 days ago

Verified e3797b18

webgpu : add Q2_0 support (#30914)

teddygood committed 24 days ago

Verified 0b1281ae

dflash : load optional FC input scale (#30945)

hthadicherla committed 24 days ago

Verified 4912ac24

ui: Add "Show tool call in progress" setting (#30681)

allozaur committed 24 days ago

Verified 84746c5c

lora : support prism.hadamard folded weights (#30910)

vieduy committed 24 days ago

Verified 0682a576

ui : key draft sidecar badges by model instead of repo (#30919)

allozaur committed 24 days ago

Verified a6b83d42

opencl: add F32-to-I32 ARGMAX (#30897)

kurquhar committed 24 days ago

Verified c7fc5094

models : add Flint model (FlintForCausalLM) (#30928)

alexey-rsi committed 24 days ago

Verified a37cc461

jinja : fix namespace attributes shadowed by dict methods (#30941)

tdakhran committed 24 days ago

Verified 99fb9971

model: Add support for SenseNova U1 model (#28919)

Maphist0 committed 24 days ago

Verified 7775ffb7

llama: create new fused op for QSA mask (#30640)

am17an committed 25 days ago

Verified d454a204

sycl: q8 fast dequant (#30775)

92es committed 25 days ago

Verified 0756e83f

jinja: support hexadecimal string escapes (#30929)

alexey-rsi committed 25 days ago

Verified 9febe63d

ggml-opencl : add IM2COL_3D support (#30895)

kurquhar committed 25 days ago

Verified 9b56ccf0

mtmd: add state object for mtmd_gen (#30902)

ngxson committed 25 days ago

Verified 0a1ef309

server : forward llama.cpp sampling params on /v1/messages (#30856)

Dgarc359 committed 25 days ago

Verified 261700b8

vendor: update BoringSSL to 0.20261020.0 (#30889)

cabelo committed 25 days ago

Verified f8490953

llama : optimize vocabulary loading (#30679)

mykhailen committed 25 days ago

Verified 824eaecf

llama : place d2t by its set_rows use so it can stay on the GPU (#30912)

am17an committed 25 days ago

Verified eb182d85

Older