ggerganov/llama.cpp

Pull Requests Commits

Fix FlashAttention debug test, FP32 assert (#7684)

JohannesGaessler committed 1 year ago

Verified e141ce62

server : new UI (#7633)

mounta11n committed 1 year ago

Verified 2e666832

SimpleChat: Simple histogram/repeatMatching driven garbageTrimming, Settings UI, Streaming mode, OpenAi Compat (Model, Authorization Bearer), Save/Restore session, Auto Settings UI (#7548)

hanishkvc committed 1 year ago

Verified 2ac95c9d

CUDA: fix Pascal FA, deq. KV to FP16 for batch > 8 (#7681)

JohannesGaessler committed 1 year ago

Verified 750f60c0

CUDA: quantized KV support for FA vec (#7527)

JohannesGaessler committed 1 year ago

Verified 9b596417

server : update js (#7670)

ggerganov committed 1 year ago

Verified a323ec60

convert-hf : Handle NotImplementedError in convert-hf-to-gguf (#7660)

Galunid committed 1 year ago

Verified 0515ad93

scripts: update compare_llama_bench.py [no ci] (#7673)

JohannesGaessler committed 1 year ago

Verified c8047d53

Improve HIP compatibility (#7672)

daniandtheweb committed 1 year ago

Verified 30e238b2

readme : link homebrew discussion

ggerganov committed 1 year ago

Verified 16926dff

ggml : fix loongson compile warnings (#7537)

ggerganov committed 1 year ago

Verified 0c27e6f6

Somehow '**' got lost (#7663)

Galunid committed 1 year ago

Verified 2e32f874

Add convert.py removal to hot topics (#7662)

Galunid committed 1 year ago

Verified 1af511fc

[no ci] docs: add aikit to readme (#7650)

sozercan committed 1 year ago

Verified 0541f062

Fixed painfully slow single process builds. (#7326)

jboero committed 1 year ago

Verified 9022c336

llama : cache llama_token_to_piece (#7587)

ggerganov committed 1 year ago

Verified 5921b8f0

Fix conan badge display [no ci] (#7645)

MartinDelille committed 1 year ago

Verified 5dcdf946

Add brew installation instruction to README [no ci] (#7616)

makuche committed 1 year ago

Verified 2e2340de

readme : add Conan badge (#7638)

MartinDelille committed 1 year ago

Verified 7846540b

github: add contact links to issues and convert question into research [no ci] (#7612)

mofosyne committed 1 year ago

Verified e6157f94

Move convert.py to examples/convert-legacy-llama.py (#7430)

Galunid committed 1 year ago

Verified 9c4c9cc8

faster avx512 exp implementation (#7551)

chriselrod committed 1 year ago

Verified 59b0d077

ggml : fix loongarch build (O2 issue) (#7636)

junchao-loongson committed 1 year ago

Verified d5c05821

README: explain parallel build [no ci] (#7618)

JohannesGaessler committed 1 year ago

Verified 972b555a

[SYCL] fix intel docker (#7630)

airMeng committed 1 year ago

Verified 3854c9d0

gguf-py : Add tokenizer.ggml.pre to gguf-new-metadata.py (#7627)

Galunid committed 1 year ago

Verified eb57fee5

metal : remove invalid asserts (#7617)

ggerganov committed 1 year ago

Verified 55d62262

metal : add missing asserts (#7617)

ggerganov committed 1 year ago

Verified 975ec63f

ggml : fix YARN + add tests + add asserts (#7617)

ggerganov committed 1 year ago

Verified fb76ec31

cuda : non-cont concat support (#7610)

ggerganov committed 1 year ago

Verified cce3dcff

Newer Older