ggml-org/llama.cpp

Pull Requests Commits

ggml : fix batched tensor 2D detection

ggerganov committed 2 days ago

43144a4c

ggml : remove async tensor_n APIs and add null buffer hooks

ggerganov committed 2 days ago

55118177

ggml : add CUDA support for batched tensor APIs

ggerganov committed 2 days ago

32e23059

llama : batch device io tensor copies

ggerganov committed 2 days ago

8eb79de9

ggml : add batched tensor copy API

ggerganov committed 2 days ago

926b097a

llama : use batched tensor copy API in context io

ggerganov committed 2 days ago

860596cf

ggml : add generic batched tensor copy API

ggerganov committed 2 days ago

1ee48d8e

vocab : add \p{M} to lookaheads in Tekken and GPT-4o pre-tokenizer regexes (#30845)

tkran2 committed 2 days ago

Verified d8d92973

tests : make MoE router selection independent of the backend noise (#30818)

ggerganov committed 2 days ago

Verified 7879491c

hexagon: update to snapdragon-toolchain v0.8 and enable hexagon v85 (#30854)

max-krasnyansky committed 2 days ago

Verified 84663953

sycl: fix compiler warnings - missing prototypes, dead code, extra semicolons (#30271)

mctylr-gh committed 2 days ago

Verified dd002b91

ui: hide the native clear button in SearchInput (#30719)

aiqubits committed 2 days ago

Verified 1d88cb05

sycl : fix get_rows on reordered quantized weights (#30430)

skipzx-lab committed 2 days ago

Verified 3bb63cbc

mtmd : match the transformers Qwen-VL resize (#30736)

boshjerns committed 2 days ago

Verified f711c8f3

ci, docs: add WASI CI and docs (#30754)

MendyBerger committed 2 days ago

Verified 921257f5

ci : re-enable type/lint-check for skills (#30713)

CISC committed 2 days ago

Verified 3ffb6293

server: add --connect, allow p2p access to server outside LAN (#28277)

ngxson committed 2 days ago

Verified 7adf160c

ui : fix attachment text preview clipped when a line is longer than the container (#30806)

A-Raphie committed 2 days ago

Verified ab7ed0d4

hybrid dna tokenizer fixes (#30825)

kashif committed 2 days ago

Verified 4c3c5810

opencl: add cooperative-k dp4a gemm for q4_0, q4_k and q6_k (#30741)

lhez committed 2 days ago

Verified 1b29621b

opencl: add get_rows for i32, cpy for i32 and f32 (#30756)

lhez committed 2 days ago

Verified 6fa4f6cd

jinja : implement `unique` filter (#30819)

aldehir committed 3 days ago

Verified 0ce8cd11

server : fix swa checkpoint invalidation logic (#30666)

tdakhran committed 3 days ago

Verified f174553e

vendor : update cpp-httplib to 0.62.0 (#30785)

cabelo committed 3 days ago

Verified 7951f601

moe-cache : replace MoE LRU cache with a 50/50 SLRU cache (#30637)

eapache committed 3 days ago

Verified 33d6135c

metal : fix moe_reduce fusion race when dst aliases its inputs (#30795)

ggerganov committed 3 days ago

Verified 098ce526

cann : refactor `#ifdef` blocks to avoid unreachable code after `return` (#25273)

rauletorresc committed 3 days ago

Verified f3fb2fcd

tests: add big-endian conversion tests (#30705)

taronaeo committed 3 days ago

Verified fda00f18

CUDA: fix warning for nonexistent extra semicolon (#30774)

JohannesGaessler committed 3 days ago

Verified 4866db35

ggml-cpu: vectorize ggml_v_silu and ggml_v_expf (#30791)

taronaeo committed 3 days ago

Verified 42b70032

Older