ggerganov/llama.cpp

Pull Requests Commits

llama : add Jina Embeddings architecture (#6826)

JoanFM committed 1 year ago

Verified b83cc3f5

ggml : full ALiBi support (#7192)

ggerganov committed 1 year ago

Verified 9cb317f7

llama-bench : add pp+tg test type (#7199)

slaren committed 1 year ago

Verified e8496488

metal : fix flash attention kernel requirements (#7169)

ggerganov committed 1 year ago

Verified 18e43766

convert : print "ignore_merges" field

ggerganov committed 1 year ago

Verified 8c660242

llama : use n_vocab to differentiate between mistral 7B and llama3 8B (#7200)

slaren committed 1 year ago

Verified 25c6e82e

Fix memory bug in grammar parser (#7194)

jart committed 1 year ago

Verified 4e388097

Main+: optionally allow special tokens from user in interactive mode (#7097)

hanishkvc committed 1 year ago

Verified f89fe273

llava : fix moondream support (#7163)

abetlen committed 1 year ago

Verified d11afd66

Minor arithmetic improvement to mmvq wrapper kernel (#7172)

OuadiElfarouki committed 1 year ago

Verified 8c570c94

eval-callback : fix conversion to float (#7184)

slaren committed 1 year ago

Verified eaf4bd8b

Vulkan Bugfixes and Improvements (#7084)

0cc4m committed 1 year ago

Verified befddd0f

readme : add scheduled server workflow status badge

ggerganov committed 1 year ago

Verified d46dbc76

readme : add app (#6371)

l3utterfly committed 1 year ago

Verified 0961d866

llama3 custom regex split (#6965)

jaime-m-p committed 1 year ago

Verified 43248e55

CUDA: generalize FP16 fattn vec kernel (#7061)

JohannesGaessler committed 1 year ago

Verified a743d76a

Add warning if token is invalid (#7173)

Galunid committed 1 year ago

Verified f31ec120

llama : update llama_timings.n_p_eval setting (#7160)

danbev committed 1 year ago

Verified fd9f92b1

gguf-py : add special token modification capability (#7166)

CISC committed 1 year ago

Verified 22842164

opencl : alignment size converted from bits to bytes (#7090)

albertjin committed 1 year ago

Verified 47345248

TypoFix (#7162)

AhmedZeer committed 1 year ago

Verified 07cd41d0

cmake : fix typo (#7151)

cebtenzzre committed 1 year ago

Verified 4426e298

convert-hf : save memory with lazy evaluation (#7075)

compilade committed 1 year ago

Verified f98eb31c

Introduction of CUDA Graphs to LLama.cpp (#6766)

agray3 committed 1 year ago

Verified bc4bba36

JSON: [key] -> .at(key), assert() -> GGML_ASSERT (#7143)

JohannesGaessler committed 1 year ago

Verified c12452c7

Revert "llava : add support for moondream vision language model (#6899)"

ggerganov committed 1 year ago

Verified 9da243b3

server : add themes + favicon (#6848)

jboero committed 1 year ago

Verified bd1871fa

metal : use `vm_allocate` instead of `posix_memalign` on macOS (#7078)

giladgd committed 1 year ago

Verified 26458af1

main : add --conversation / -cnv flag (#7108)

Dawid Potocki committed 1 year ago

Verified 83330d8c

sgemm : AVX Q4_0 and Q8_0 (#6891)

netrunnereve committed 1 year ago

Verified 465263d0

Newer Older