ochafik/llama.cpp

Pull Requests Commits

ochafik committed 1 year ago

e7fc8ed7

jinja: revamp Value

ochafik committed 1 year ago

9446d238

jinja: support string stripping

ochafik committed 1 year ago

17fd9292

parser improvements

ochafik committed 1 year ago

974027f4

Update jinja.hpp

ochafik committed 1 year ago

7a54e2c7

jinja: more parsing (broken)

ochafik committed 1 year ago

a305037a

jinja templates (broken)

ochafik committed 1 year ago

27be1faf

ochafik committed 1 year ago

32c6e07d

initial grammar trigger words code

ochafik committed 1 year ago

0b2e14fa

Update json_schema_to_grammar.py

ochafik committed 1 year ago

73d9307c

Update utils.hpp

ochafik committed 1 year ago

fe1892b6

ochafik committed 1 year ago

0b7531b4

c++ function calling

ochafik committed 1 year ago

a6c24258

Revert "rename openai/{server/proxy}.py"

ochafik committed 1 year ago

19069cd3

rename openai/{server/proxy}.py

ochafik committed 1 year ago

a72c47a1

revert tool_call_grammar

ochafik committed 1 year ago

3a22d8e6

start reverting changes

ochafik committed 1 year ago

6a9d82d0

Merge remote-tracking branch 'origin/master' into agent-example

ochafik committed 1 year ago

d75f629b

docker : build images only once (#9225)

slaren committed 1 year ago

Verified 9fe94cca

docker : update CUDA images (#9213)

slaren committed 1 year ago

Verified 66b039a5

vulkan : fix build (#0)

ggerganov committed 1 year ago

20f1789d

ggerganov committed 1 year ago

231cff5f

Fix minicpm example directory (#9111)

xyb committed 1 year ago

Verified 3246fe84

llama : fix qs.n_attention_wv for DeepSeek-V2 (#9156)

compilade committed 1 year ago

Verified 78eb487b

server : add some missing env variables (#9116)

ngxson committed 1 year ago

Verified a77feb5d

llama : fix ChatGLM4 wrong shape (#9194)

RealJosephus committed 1 year ago

Verified 2e59d61c

llama : fix llama3.1 rope_freqs not respecting custom head_dim (#9141)

nyxkrage committed 1 year ago

Verified 75e1dbba

common : Update stb_image.h to latest version (#9161)

arch-btw committed 1 year ago

Verified ad76569f

ggml : do not crash when quantizing q4_x_x with an imatrix (#9192)

slaren committed 1 year ago

Verified 7d787ed9

metal : separate scale and mask from QKT in FA kernel (#9189)

ggerganov committed 1 year ago

Verified 06658ad7

Older