ggerganov/llama.cpp

Pull Requests Commits

server: Adding spec-decode counters to /metrics endpoint (#26389)

BLSharda committed 15 days ago

Verified a035a888

convert: Add endianness conversion for Q1 and TQ2 quantizations (#26618)

Andreas-Krebbel committed 15 days ago

Verified 020760ad

vendor : apply patches for subprocess.h (#26606)

ngxson committed 15 days ago

Verified 61881b1f

ui: show generation statistics by default in chat settings (#26624)

allozaur committed 15 days ago

Verified 3e3a7a41

build : remove GGML_METAL_USE_BF16 from all build scripts (#26604)

nikwen committed 15 days ago

Verified d52ec04a

ui: Update vulnerable packages + cleanup Storybook config (#26607)

allozaur committed 15 days ago

Verified e031d956

Prefer npm ci over install for security (#26601)

eapache committed 15 days ago

Verified 6ea215d1

server: decode Windows OEM output to UTF-8 in built-in tools (#26597)

ServeurpersoCom committed 15 days ago

Verified 4308a4f0

mtmd: correcting duplicate empty audio chunks for short inputs (#26536)

Abhinaykrishna2 committed 15 days ago

Verified 474c92e7

sampler : remove "full-context windows" from history-based samplers (#26524)

ggerganov committed 15 days ago

Verified a6aa6f54

gguf-split: Add option to delete split parts during merge (#26538)

guiq83 committed 15 days ago

Verified 76c956c1

ui: CWD for agent (#26518)

allozaur committed 15 days ago

Verified 2f56fc34

mtmd: support Qwen3-TTS (note: breaking change to llama-tts binary) (#26254)

ngxson committed 16 days ago

Verified 07132750

models : fix dflash wo_a reshape on load (#26577)

ggerganov committed 16 days ago

Verified 1c3c9674

ci: fix pre-built binaries no longer working on macOS 15 and below (#26375)

nikwen committed 16 days ago

Verified 6b5224cf

speculative : refactor enabled configs common_speculative_init (#26510)

danbev committed 16 days ago

Verified 7bd8282c

gguf-py: validate n_dims and guard against uint64 overflow in reader (#25401)

hclsys committed 16 days ago

Verified 5788b510

ggerganov committed 16 days ago

2e17f69e

ggml : bump version to 0.18.1 (ggml/1578)

ggerganov committed 16 days ago

15831f57

convert : add missing return after setting tekken vocab (#25947)

AngelGalindo7 committed 16 days ago

Verified b5746d28

vulkan backend ops: implemented GATED_LINEAR_ATTN (#25601)

PranavUttarkar committed 16 days ago

Verified f26efa02

vocab : validate plamo2 byte tokens (#26511)

CISC committed 16 days ago

Verified cf06ad7d

convert : import bytes_to_unicode from convert_slow_tokenizer (#26217)

SolshineCode committed 16 days ago

Verified b06fbc96

model : allow reshape of tensors during load (#26531)

ggerganov committed 16 days ago

Verified 1269cb1f

llama : move n_vocab from llama_sampler_data to penalty_sampler (#26520)

ggerganov committed 16 days ago

Verified 935cad64

ci: fix vulkan llvmpipe runs (#26533)

netrunnereve committed 16 days ago

Verified 22dc605c

sycl: parallelize the non-contiguous concat kernel (#25852)

Titaniumtown committed 16 days ago

Verified 6c8dcaa7

Extended SYCL oneDNN SDPA to non-FP16 KV caches (Q4_0–Q8_0 and FP32) (#25874)

johnkarlhill committed 16 days ago

Verified 66fa168a

chat : add new template for DeepSeek V4 Flash 0731 (#26398)

tarruda committed 16 days ago

Verified 0ef6e55e

vendor : update cpp-httplib to 0.52.0 (#26485)

cabelo committed 16 days ago

Verified 94bc47f2

Newer Older