ngxson/llama.cpp

Pull Requests Commits

vulkan: implement more backpropagation operators (#11914)

remyoudompheng committed 348 days ago

Verified 61d4f39d

server: support add_generation_prompt query param (#12062)

ochafik committed 348 days ago

Verified 0b527456

Add Doc for Converting Granite Vision -> GGUF (#12006)

alex-jw-brooks committed 348 days ago

Verified 4d1051a4

llama : expose llama_model_n_head_kv in the API (#11997)

vlovich committed 348 days ago

Verified 3e9a2860

metal : copy kernels for quant to F32/F16 conversions (#12017)

gcp committed 348 days ago

Verified 58d07a80

opencl: fix for small models (#11950)

lhez committed 349 days ago

Verified 34a846b5

llava : Add Granite Vision Support (#11794)

alex-jw-brooks committed 349 days ago

Verified 7a2c913e

[SYCL] Optimize mul_mat for Q4_0 on Intel GPU (#12035)

NeoZhangJianyu committed 349 days ago

Verified 08d59862

gguf_convert_endian.py: implement byteswapping for q4_k and q6_k (#11349)

AlekseiNikiforovIBM committed 349 days ago

Verified 651adf4b

SYCL: Fix GGML_SYCL_DEBUG macro (#11995)

qnixsynapse committed 349 days ago

Verified 8303e8b0

run: allow to customize prompt by env var LLAMA_PROMPT_PREFIX (#12041)

benoitf committed 350 days ago

Verified 7ad0779f

Some llama-run cleanups (#11973)

ericcurtin committed 350 days ago

Verified f777a73e

ggml-cpu: Support s390x SIMD Instruction Set (#12019)

taronaeo committed 351 days ago

Verified af7747c9

CUDA: app option to compile without FlashAttention (#12025)

JohannesGaessler committed 351 days ago

Verified a28e0d5e

llava: build clip image from pixels (#11999)

tinglou committed 351 days ago

Verified 36c258ee

ci : fix arm upload artifacts (#12024)

ggerganov committed 351 days ago

Verified f3e64859

CUDA: optimize FA for GQA + large batches (#12014)

JohannesGaessler committed 351 days ago

Verified 5fa07c2f

ci : Build on Github-hosted arm64 runners (#12009)

Rohanjames1997 committed 351 days ago

Verified 335eb04a

server : disable Nagle's algorithm (#12020)

ggerganov committed 351 days ago

Verified cf756d6e

cuda: Add Q5_1, Q5_0, Q4_1 and Q4_0 to F32 conversion support. (#12000)

gcp committed 351 days ago

Verified d7090842

llama.swiftui : add "Done" dismiss button to help view (#11998)

danbev committed 351 days ago

Verified de8b5a36

llama : skip loading unused tensors (#12004)

ggerganov committed 352 days ago

Verified 51f311e0

doc: update contributing guidelines [no ci] (#11969)

JohannesGaessler committed 352 days ago

Verified 586d5fe6

CUDA: correct the lowest Maxwell supported by CUDA 12 (#11984)

PureJourney committed 352 days ago

Verified ecc8e3ae

MUSA: support ARM64 and enable dp4a .etc (#11843)

BodhiHu committed 352 days ago

Verified 0b3863ff

clip : fix visual encoders with no CLS (#11982)

alex-jw-brooks committed 352 days ago

Verified ee02ad02

server (webui): Fix Premature Submission During IME Conversion (#11971)

mmngays committed 353 days ago

Verified c392e509

ggml-cpu: Add CPU backend support for KleidiAI library (#11390)

chaxu01 committed 353 days ago

Verified c5d91a74

ggml: aarch64: implement SVE kernels for q3_K_q8_K vector dot (#11917)

Vithulep committed 353 days ago

Verified 4806498b

run : add --chat-template-file (#11961)

engelmi committed 353 days ago

Verified 0d559580

Newer Older