ngxson/llama.cpp

Pull Requests Commits

Update examples/llava/mtmd.cpp

ngxson committed 240 days ago

Verified 56bc6741

Merge branch 'master' into xsn/merge_llava_to_mtmd_cli

ngxson committed 242 days ago

6b2ea2c8

clip : refactor, add `image_manipulation` and `llava_uhd` classes (#13011)

ngxson committed 242 days ago

Verified 37b9f0d2

mtmd : add not supported msg for qwen2vl

ngxson committed 242 days ago

7360d43a

update hot topics

ngxson committed 242 days ago

ccca0fb6

remove cpp files of llava and minicpmv

ngxson committed 242 days ago

a1f65d1f

support for minicpmv

ngxson committed 242 days ago

da6b9eb2

main : Fix Ctrl+D/newline handling (#12951)

danielzgtg committed 242 days ago

Verified 64082100

gguf-py : GGUF Editor GUI - Python + Qt6 (#12930)

christopherthompson81 committed 243 days ago

Verified aff9d107

server : use std::move whenever possible (#12936)

ngxson committed 243 days ago

Verified 35370ba9

mtmd : merge `llava-cli` and `gemma3-cli` into single `mtmd-cli`

ngxson committed 243 days ago

4737bd01

SYCL: Refactor and enable FP16 in binary broadcast OPs (#12975)

qnixsynapse committed 243 days ago

Verified 8d660057

mtmd : add methods to access `mtmd_image_tokens` (#12906)

ngxson committed 243 days ago

Verified b9154ecf

rpc : add RPC_CMD_HELLO (#12955)

rgerganov committed 243 days ago

Verified 2db9ba14

graph : make FA compatible with MLA + add initial Metal kernels (#12953)

ggerganov committed 244 days ago

Verified 2f74c354

ggml: Re-enable CUDA graphs in presence of CONT and DUP nodes (#12970)

agray3 committed 244 days ago

Verified 207c22ec

CANN: Add support for async operator submission (#12864)

hipudding committed 244 days ago

Verified 7a395f67

llama : recognize IBM Granite 3.3 FIM tokens (#12988)

Noeda committed 244 days ago

Verified 971f245b

opencl: fix incorrect local_size index in profiling log (#12868)

kimminsu38oo committed 244 days ago

Verified 12b17501

vulkan: enable coopmat2 FA gqa and split_k optimizations more often (#12931)

jeffbolznv committed 245 days ago

Verified 015022bb

CANN: Add 310P operator support check (#12962)

noemotiovon committed 245 days ago

Verified b43d89e3

opencl: split `ggml-opencl.cl` into multiple files and cleanup (#12886)

lhez committed 245 days ago

Verified 80f19b41

metal : add FA-vec kernels for head size 96 (#12952)

ggerganov committed 246 days ago

Verified f8f820cc

CANN: Add x86 build ci (#12950)

hipudding committed 246 days ago

Verified 54a72720

CUDA/HIP: Share the same unified memory allocation logic. (#12934)

hjc4869 committed 246 days ago

Verified 84778e97

SYCL: Add ROPE vision kernel (#12887)

qnixsynapse committed 246 days ago

Verified 51067647

llama : DeepSeek V2/V3 MLA implementation (#12801)

jukofyork committed 246 days ago

Verified daa42288

ggml : Add AVX512 implementation of GEMM - Q4_Kx8 (#12829)

Srihari-mcw committed 246 days ago

Verified eccc7a16

CANN: Opt ROPE optimization (#12865)

noemotiovon committed 246 days ago

Verified 0019279b

CANN: Optimize CANN buffer pool memory management (#12875)

bachelor-dou committed 246 days ago

Verified b0c75ac9

Older