ochafik/llama.cpp

Pull Requests Commits

ochafik committed 1 year ago

7be2efe9

Merge remote-tracking branch 'origin/master' into grammar-fast

ochafik committed 1 year ago

2477d84c

`grammars`: fix resampling logic regression (#7424)

ochafik committed 1 year ago

Verified e402de36

CUDA: fix unused warning in mmq.cu (#7442)

JohannesGaessler committed 1 year ago

Verified fcf6538b

tests : test-tokenizer-0.sh print more info (#7402)

ggerganov committed 1 year ago

Verified c3f8d583

examples: cache hf model when --model not provided (#7353)

amirzia committed 1 year ago

Verified 11474e75

CUDA: deduplicate mmq code (#7397)

JohannesGaessler committed 1 year ago

Verified d8ee9022

Tokenizer SPM fixes for phi-3 and llama-spm (bugfix) (#7425)

jaime-m-p committed 1 year ago

Verified d7e852c1

Tokenizer SPM fixes for phi-3 and llama-spm (#7375)

jaime-m-p committed 1 year ago

Verified 917dc8cf

llama : remove Persimmon (#7408)

ggerganov committed 1 year ago

Verified fabf30b4

perplexity: update README FP16 results [no ci] (#7413)

JohannesGaessler committed 1 year ago

Verified 20385ceb

rpc : track allocated buffers (#7411)

rgerganov committed 1 year ago

Verified db10f013

server : fix temperature + disable some tests (#7409)

ggerganov committed 1 year ago

Verified 3bc10cb4

[SYCL] Update SYCL upscale operation (#7321)

AidanBeltonS committed 1 year ago

Verified 6bf9b66f

Update README.md (#7410)

binganao committed 1 year ago

Verified 26cd4237

ggml-opencl, llama: using reserve() if count already known (#7272)

GermanAizek committed 1 year ago

Verified 213e90ed

ggml : add loongarch lsx and lasx support (#6454)

junchao-loongson committed 1 year ago

Verified 65c58207

server : tuning tests (#7388)

ggerganov committed 1 year ago

Verified 1cc0155d

server : return error on too large embedding input (#7389)

ggerganov committed 1 year ago

Verified e932094d

tests : fix --keep_split -> --keep-split (#7374)

ggerganov committed 1 year ago

Verified 2789baf4

Add provisions for windows support for BF16 code including CMake provision for enabling AVX512_BF16 (#7258)

Srihari-mcw committed 1 year ago

Verified 33c8d50a

llama : remove MPI backend (#7395)

slaren committed 1 year ago

Verified d359f309

quantize : fix --keep-split check (#7374)

fredlas committed 1 year ago

Verified 1ea2a003

Vulkan Embedding Fix (#7360)

0cc4m committed 1 year ago

Verified f030ec1f

ggml : fix another case of quants nans (#7387)

slaren committed 1 year ago

Verified e4e6f67b

ggml: implement quantized KV cache for FA (#7372)

JohannesGaessler committed 1 year ago

Verified 5ca49cbe

server: add test for token probs (#7347)

JohannesGaessler committed 1 year ago

Verified 1b01f06d

server: fix seed being reported back (#7382)

JohannesGaessler committed 1 year ago

Verified 41858392

Add StableLM2 pre-tokenizer (#7349)

aahouzi committed 1 year ago

Verified 6aade19e

cuda : clear error after buffer allocation failure (#7376)

slaren committed 1 year ago

Verified ab33f7a3

Older