ggml-org/llama.cpp

Pull Requests Commits

py : fix empty bytes arg

ggerganov committed 1 year ago

Verified adcf16fd

py : handle byte tokens in `get_token_type`

ggerganov committed 1 year ago

Verified ded2ad5b

ggml : avoid duplicating function calls using MIN/MAX macros (#5325)

tom7 committed 1 year ago

Verified abb61944

iq3_xxs: quards for the no-imatrix situation (#5334)

ikawrakow committed 1 year ago

Verified 89503dcb

py : fix internlm2-hf convert to gguf (#5305)

SolenoidWGT committed 1 year ago

Verified 7e1ae372

iq2_xxs: tune quantization (#5320)

ikawrakow committed 1 year ago

Verified 6fdfa2ec

server : allow to get default generation settings for completion (#5307)

z80maniac committed 1 year ago

Verified a2d60c91

common : add dynamic temperature parameters to main example cli (#5295)

l3utterfly committed 1 year ago

Verified e6f81775

scripts : fix typos, cleanup (#5303)

ggerganov committed 1 year ago

Verified 30679d43

scripts : add non-interactive server-llm.sh (#5303)

garrnizon committed 1 year ago

Verified 4be04c89

readme : add CodeShell models to the supported models list (#5330)

chiranko committed 1 year ago

Verified 5d55b0cd

[SYCL] Fix cpy with dims of 3 (#5289)

AidanBeltonS committed 1 year ago

Verified 4833ac20

flake.lock: Update

github-actions[bot] committed 1 year ago

9392ebd4

Adding some imatrix tools (#5302)

ikawrakow committed 1 year ago

Verified 5ed26e1f

cmake : use set() for LLAMA_WIN_VER (#5298)

crimsonmagick committed 1 year ago

Verified 277fad30

make: add nvcc info print (#5310)

JohannesGaessler committed 1 year ago

Verified 3c0d25c4

make: fix nvcc optimization flags for host code (#5309)

JohannesGaessler committed 1 year ago

Verified 3cc5ed35

add Vulkan support to Nix flake

mschwaig committed 1 year ago

60ecf099

Vulkan Intel Fixes, Optimizations and Debugging Flags (#5301)

0cc4m committed 1 year ago

Verified e920ed39

refactor : switch to emplace_back to avoid extra object (#5291)

MKlimenko committed 1 year ago

Verified 52bb63c7

YaRN : store rope scaling type as int32_t in memory (#5285)

cebtenzzre committed 1 year ago

Verified 1ec3332a

readme : add tenere in the ui tools list (#5284)

pythops committed 1 year ago

Verified 6a66c507

Fix im2col with 32fp (#5286)

AidanBeltonS committed 1 year ago

Verified a305dba8

perplexity : fix KL divergence calculations on Windows (#5273)

kalomaze committed 1 year ago

Verified 19122117

scripts : parse wtype in server-llm.sh (#5167)

ggerganov committed 1 year ago

Verified e437b37f

py : add check for '.attn.masked_bias' layers to GPT2model (#5281)

MirrorAzure committed 1 year ago

Verified 2d40085c

Tidy ggml-sycl (#5261)

AidanBeltonS committed 1 year ago

Verified b05102fe

docker : add build for SYCL, Vulkan + update readme (#5228)

ngxson committed 1 year ago

Verified 6b91b1e0

[SYCL] get MAX_MEM_ALLOC from device property (#5270)

airMeng committed 1 year ago

Verified e805f0fa

[SYCL] update guide of SYCL backend (#5254)

NeoZhangJianyu committed 1 year ago

Verified af3ba5d9

Older