ggml-org/llama.cpp

Pull Requests Commits

ggml : fix fallback to CPU for ununsupported ops (#15118)

slaren committed 186 days ago

Verified 0d883154

chat : fix yandex chat template (#15116)

CISC committed 186 days ago

Verified 65c797c4

chat : fix hunyuan auto-detection (#15114)

stevenkuang-tencent committed 186 days ago

Verified 25726898

CANN: add support for ACL Graph (#15065)

noemotiovon committed 186 days ago

Verified 22414532

ggml: WebGPU disable SET_ROWS for now (#15078)

reeselevine committed 186 days ago

Verified 9515c613

llama : add gpt-oss (#15091)

ggerganov committed 186 days ago

Verified fd1234cb

chat : only remove double bos/eos if added (#15086)

CISC committed 186 days ago

Verified f324a3b7

readme : update hot topics (#15097)

ggerganov committed 186 days ago

Verified be426425

sycl: fix mul_mat selection (#15092)

Rbiessy committed 187 days ago

Verified 3306ceab

Fix `glm4moe` bug (#15088)

jukofyork committed 187 days ago

Verified c81de6e1

webui: fix markdown table (#15081)

dindinw committed 187 days ago

Verified 22f060c9

context : fix index overflow on huge outputs (#15080)

compilade committed 187 days ago

Verified ee3a9fcf

llama : add --n-cpu-moe option (#15077)

slaren committed 187 days ago

Verified ec428b02

imatrix : warn when GGUF imatrix is saved without .gguf suffix (#15076)

compilade committed 187 days ago

Verified 19f68fa5

cmake: Add GGML_BACKEND_DIR option (#15074)

ckastner committed 187 days ago

Verified 41613437

gguf-py : add --chat-template-file to gguf_new_metadata (#15075)

CISC committed 187 days ago

Verified e5bebe52

model: support GLM 4.5 family of models (#14939)

sammcj committed 187 days ago

Verified ef0144c0

quantize : fix confusing error message if ftype is invalid (#15071)

CISC committed 188 days ago

Verified 2721257e

ggml: WebGPU backend host improvements and style fixing (#14978)

reeselevine committed 188 days ago

Verified 587d0118

vulkan: fix build when using glslang that does not support coopmat2 (#15062)

jeffbolznv committed 188 days ago

Verified 5aa1105d

imatrix : use GGUF by default (#14842)

compilade committed 188 days ago

Verified d31192b4

imatrix : fix 3d activation handling for hybrid and recurrent models (#14994)

compilade committed 188 days ago

Verified 0a2f5496

memory : handle kv_unified for hybrid models (#15050)

compilade committed 188 days ago

Verified 11a38111

vocab : JetBrains Mellum pre-tokenizer (#15045)

csabakecskemeti committed 188 days ago

Verified 97366dc6

model : add text-only support for Kimi-VL (and find special tokens in text_config) (#15051)

gabriellarson committed 189 days ago

Verified 83bc2f28

vulkan: Use coopmat2 for conv2d (#14982)

jeffbolznv committed 189 days ago

Verified 6c7a4411

opencl: fix adreno compiler detection logic (#15029)

lhez committed 189 days ago

Verified 5c0eb5ef

CUDA: use mma FA kernel for gqa > 4 on RTX 4000 (#15035)

JohannesGaessler committed 190 days ago

Verified 03d46982

cuda: make im2col a little faster (#15025)

leejet committed 190 days ago

Verified 3303c19b

kv-cache : skip alignment of n_stream in kv-cache log msg [no ci] (#15040)

danbev committed 190 days ago

Verified 4fdea540

Newer Older