ochafik/llama.cpp

Pull Requests Commits

Merge branch 'fix-gen-prompt' into enable-thinking

ochafik committed 1 year ago

91623806

switch to --reasoning-budget flag

ochafik committed 1 year ago

9cdeebe8

const char* return for chat enum name helpers

ochafik committed 1 year ago

8547fccf

Merge branch 'enable-thinking' of github.com:ochafik/llama.cpp into enable-thinking

ochafik committed 1 year ago

fe6022f4

Update common/chat.cpp

ochafik committed 1 year ago

Verified 355b38c9

rewrite docs as list as suggested

ochafik committed 1 year ago

6b9efe7b

fix command r7b's nothink w/ official template

ochafik committed 1 year ago

473c01e5

update --reasoning-format={disabled -> nothink} as suggested

ochafik committed 1 year ago

cdea6a9b

Add models/templates/Qwen-Qwen3-0.6B.jinja

ochafik committed 1 year ago

b6eb0a5c

Update README.md

ochafik committed 1 year ago

df25e6bb

server: add --reasoning-format=disabled to disable thinking (incl. qwen3 w/ enable_thinking:false)

ochafik committed 1 year ago

b457f89e

tools: enable hermes2/qwen chat logic even w/o tools

ochafik committed 1 year ago

43b5626e

server: fix/test add_generation_prompt

ochafik committed 1 year ago

8a25f796

`server`: streaming of tool calls and thoughts when `--jinja` is on (#12379)

ochafik committed 1 year ago

Verified f5cd27b7

releases : bundle llvm omp library in windows release (#13763)

slaren committed 1 year ago

Verified a2d02d57

releases : enable openmp in windows cpu backend build (#13756)

slaren committed 1 year ago

Verified 17fc817b

ggml-cpu : set openmp wait time if not set (#13758)

slaren committed 1 year ago

Verified 2bd1b30f

Move GLM4 f32 attention fix to the correct function (#13750)

0cc4m committed 1 year ago

Verified 259469c4

ggml : add ggml_gelu_erf() CUDA kernel (#13719)

ngxson committed 1 year ago

Verified 4c32832c

vocab : fix ugm tokenizer precision (#13743)

CISC committed 1 year ago

Verified c3a26243

CUDA: fix race condition in FA vector kernels (#13742)

JohannesGaessler committed 1 year ago

Verified ffd0eae6

ci : enable winget package updates (#13734)

slaren committed 1 year ago

Verified b775345d

ci : add winget package updater (#13732)

slaren committed 1 year ago

Verified a70a8a69

hparams : initialize arrays (#13728)

ggerganov committed 1 year ago

Verified d13d0f61

llama : allow custom list of swa_layers (#13726)

ngxson committed 1 year ago

Verified 8a2afb75

server : support audio input (#13714)

ngxson committed 1 year ago

Verified 9ecf3e66

CANN: Support MUL_MAT_ID for q8_0 and q4_0 (#13705)

noemotiovon committed 1 year ago

Verified faaaff5f

ggml : fix the order of ggml_unary_op (#13718)

ngxson committed 1 year ago

Verified e16c4731

vulkan: support CPY from any type to itself (#13695)

jeffbolznv committed 1 year ago

Verified 1dcd0196

vulkan: Disable coopmat/coopmat2/bfloat extensions if glslc doesn't support it (#13696)

jeffbolznv committed 1 year ago

Verified c10ed6cb

Older