ngxson/llama.cpp

Pull Requests Commits
llama: Support MiniCPM-1B (with & w/o longrope) (#10559)
JFLFY2255 committed 1 year ago
Verified 8d0cfd55
vulkan: Implement "fast divide" (mul+shift) for unary ops like copy (#10642)
jeffbolznv committed 1 year ago
Verified 2759916d
SYCL : Move to compile time oneMKL interface backend selection for NVIDIA backend (#10584)
s-Nick committed 1 year ago
Verified 40c6d79f
fix typo of README.md (#10605)
WrRan committed 1 year ago
Verified 98036d56
Avoid using __fp16 on ARM with old nvcc (#10616)
frankier committed 1 year ago
Verified cd2f37b3
Add docs for creating a static build (#10268) (#10630)
mostlygeek committed 1 year ago
Verified da6aac91
clip : add sycl support (#10574)
piDack committed 1 year ago
Verified 01e6d9bb
vulkan: optimize and reenable split_k (#10637)
jeffbolznv committed 1 year ago
Verified cc98896d
server : (web ui) Various improvements, now use vite as bundler (#10599)
ngxson committed 1 year ago
Verified 91c36c26
scripts : remove amx sync
ggerganov committed 1 year ago
1cd3df46
sync : ggml
ggerganov committed 1 year ago
c5054718
CUDA: remove unnecessary warp reduce in FA (ggml/1032)
mahorozte committed 1 year ago
e9e661bd
feat: add `GGML_UNARY_OP_ARGMAX` Metal kernel (ggml/1019)
PABannier committed 1 year ago
efb6ae96
metal : add `GGML_OP_CONV_TRANSPOSE_1D` kernels (ggml/1026)
PABannier committed 1 year ago
667d70d1
llama : add missing LLAMA_API for llama_chat_builtin_templates (#10636)
ngxson committed 1 year ago
Verified 3b4f2e33
readme : add option, update default value, fix formatting (#10271)
pothitos committed 1 year ago
Verified 82bca225
metal : small-batch mat-mul kernels (#10581)
ggerganov committed 1 year ago
Verified 0115df2f
github : minify link [no ci] (revert)
ggerganov committed 1 year ago
Verified 515d4e53
github : minify link [no ci]
ggerganov committed 1 year ago
Verified 844e2e1f
server : fix default draft model parameters (#10586)
ggerganov committed 1 year ago
Verified 70b98fad
llama : add enum for built-in chat templates (#10623)
ngxson committed 1 year ago
Verified 642330ac
make : deprecate (#10514)
ggerganov committed 1 year ago
Verified 8648c521
server: Add "tokens per second" information in the backend (#10548)
lhpqaq committed 1 year ago
Verified 64ed2091
SYCL: Fix and switch to GGML_LOG system instead of fprintf (#10579)
qnixsynapse committed 1 year ago
Verified 991f8aab
contrib : refresh (#10593)
ggerganov committed 1 year ago
Verified 4cb003dd
Add `mistral-v1`, `mistral-v3`, `mistral-v3-tekken` and `mistral-v7` chat template types (#10572)
jukofyork committed 1 year ago
Verified 917786f4
grammars : add English-only grammar (#10612)
ggerganov committed 1 year ago
Verified 5e1ed955
ci: add error handling for Python venv creation in run.sh (#10608)
wangqin0 committed 1 year ago
Verified 5c7a5aa0
ggml : automatic selection of best CPU backend (#10606)
slaren committed 1 year ago
Verified 3420909d
server : bind to any port when specified (#10590)
alek3y committed 1 year ago
Verified 86dc11c5
Terms Of Use Privacy Policy Refund Policy Impressum