ggerganov/llama.cpp

Pull Requests Commits

ngxson committed 42 days ago

c0b9903a

grammar: fix regression caused by #17381

ngxson committed 42 days ago

3b195d30

Improved file naming & structure for UI components (#17405)

allozaur committed 42 days ago

Verified 4c91f263

grammar : fix integer overflow (#17381)

pwilkin committed 42 days ago

Verified 92c0b387

ggerganov committed 42 days ago

2286a360

metal : fix compile on macos 11 (whisper/3533)

smilingpoplar committed 42 days ago

1d321e59

common : more accurate sampling timing (#17382)

ggerganov committed 42 days ago

Verified 196f5083

convert : fix TypeError when loading base model remotely in convert_lora_to_gguf (#17385)

o7si committed 42 days ago

Verified 5088b435

ggml : Fix transposed SOLVE_TRI result (#17323)

pwilkin committed 42 days ago

Verified 845f200b

DGX Spark: UMA support (#17368)

sfudally-nvidia committed 42 days ago

Verified a7784a8b

ggml : remove useless and error-prone variadic macros (#17399)

angt committed 42 days ago

Verified 79bb7435

kleidiai: fix zero-size array declaration (#17240)

sudhiarm committed 42 days ago

Verified 3ae282a0

ggml-cpu:add RISC-V RVV (Zvfh) optimization for FP16 vector scaling (#17314)

ixgbe committed 42 days ago

Verified 5be353ec

vulkan: implement ADD1, ARANGE, FILL, SOFTPLUS, STEP, ROUND, CEIL, FLOOR, TRUNC (#17319)

giuseppe committed 43 days ago

Verified 7d77f073

vulkan: support larger argsort (#17313)

jeffbolznv committed 43 days ago

Verified 1fa4551a

vulkan: Add copy_transpose shader (#17371)

jeffbolznv committed 43 days ago

Verified 2eba631b

webui: Add a "Continue" Action for Assistant Message (#16971)

allozaur committed 43 days ago

Verified 99c53d65

convert : use self.block_count everywhere instead of reading hparams (#17359)

CISC committed 43 days ago

Verified 07b0e7a5

cuda: fix rope fusion for gemma3 (#17378)

am17an committed 43 days ago

Verified fd7353d5

Fix too relaxed check on CUDA "fast copy" (can_be_transposed) condition (#17332)

pwilkin committed 43 days ago

Verified 6fd4f953

vulkan: force full subgroups for flash attention to fix intel subgroup crash (#17356)

0cc4m committed 43 days ago

Verified 980b7cd1

ggml-cpu: Don't pass -mpowerpc64 when -mcpu already implies it (#17308)

JeremyRand committed 43 days ago

Verified c49daff5

chat: fix int overflow, prevent size calculation in float/double (#17357)

ngxson committed 44 days ago

Verified 10e97801

vocab : call reserve() for building plamo-2-translate suffix (#17343)

haiyuewa committed 44 days ago

Verified a0454920

common : Generalized XML-style tool-call parsing with streaming support (GLM 4.5/4.6 + MiniMax M2 + SeedOSS + Kimi-K2 + Qwen3-Coder + Apriel-1.5 + Xiaomi-MiMo) (#16932)

hksdpc255 committed 44 days ago

Verified 1920345c

ci : change the openEuler-310p image to fix release (#17361)

xuedinge233 committed 44 days ago

Verified 561a3e27

gitignore : be more specific about ignored stuff (#17354)

ggerganov committed 44 days ago

Verified f40a2e5f

CANN: fix acl_tensor_ptr usage in ASCEND_310P ROPE (#17347)

noemotiovon committed 44 days ago

Verified bc4064cf

fix: resolve undefined variable 'svr' compilation error (#17348)

o7si committed 44 days ago

Verified 97cb3fd5

CANN: Add openEuler-cann in build and release (#17192)

xuedinge233 committed 44 days ago

Verified ffa277a5

Older