ngxson/llama.cpp

Pull Requests Commits

CANN: update docker images to 8.5.0 and improve CANN.md (#20801)

KokerZhou committed 12 days ago

Verified 6861f650

mtmd: fix "v.patch_embd" quant and unsupported im2col ops on Metal for deepseek-ocr (#21027)

sfallah committed 12 days ago

Verified 1743d980

hip: use fnuz fp8 for conversion on CDNA3 (#21040)

IMbackK committed 12 days ago

Verified 7ca0c9cc

ci: pin external actions to exact commit SHA (#21033)

ngxson committed 12 days ago

Verified 8c60b8a2

common : add getpwuid fallback for HF cache when HOME is not set (#21035)

angt committed 12 days ago

Verified 287b5b1e

mtmd: refactor image preprocessing (#21031)

ngxson committed 12 days ago

Verified a73bbd5d

opencl: allow large buffer for adreno (#20997)

lhez committed 12 days ago

Verified ded446b3

convert : support Qwen3.5/Qwen3.5 Moe NVFP4 and add input scales (#20505)

michaelw9999 committed 12 days ago

Verified f8d4abae

convert : add RuGPT3XL (RuGPT3XLForCausalLM) support (#21011)

EvilFreelancer committed 12 days ago

Verified 3d5acab3

common : filter out imatrix when finding models (#21023)

angt committed 12 days ago

Verified 9900b29c

fix(ggml): correct RISC-V ISA string canonical ordering for RVV in CMake (#20888)

ihb2032 committed 13 days ago

Verified dc8d14c5

common : make LLAMA_CACHE the one cache for everything (#21009)

angt committed 13 days ago

Verified 93dfbc12

common : fix split model migration (#21019)

angt committed 13 days ago

Verified 3cba8bba

ggml-cuda: Add NVFP4 dp4a kernel (#20644)

michaelw9999 committed 13 days ago

Verified 112c7815

imatrix : fix crash when using --show-statistics with zero counts (#19532)

ssam18 committed 13 days ago

Verified 0fac87b1

CUDA & CPU: support F32 kernel type for `CONV_TRANSPOSE_2D` (#17094)

AgainstEntropy committed 13 days ago

Verified 0a524f24

common : do not delete old files from the old cache when updating (#21000)

angt committed 13 days ago

Verified c0159f9c

mtmd: Add DeepSeekOCR Support (#17400)

sfallah committed 13 days ago

Verified a970515b

common : fix verbosity setup (#20989)

angt committed 13 days ago

Verified 056b50c3

common : fix gguf selection in common_list_cached_models (#20996)

angt committed 13 days ago

Verified f2c72b8f

ci : fix parsing of vgpr counts in hip-quality-check (#20987)

IMbackK committed 13 days ago

Verified ec54ac13

model: codefuse-ai/F2LLM-v2 support

sfallah committed 13 days ago

Verified 80322ebd

model : allow causal_attn and pooling_type on all architectures (#20973)

Bing-su committed 13 days ago

Verified 44c51e52

snapdragon: add missing features to WoS scripts to achieve parity with ADB scripts (#20884)

aparmp-quic committed 13 days ago

Verified 1922f87c

Use docker in build-android.yml (#20928)

shreyajn committed 13 days ago

Verified 345de3cd

llama-bench: print `-n-cpu-moe` when offloaded layers > 1 (#20984)

am17an committed 14 days ago

Verified 9c600bcd

ci: Allow ninja to be used during unit test (#20742)

rillomas committed 14 days ago

Verified b2704f90

ci : disable self-hosted mac jobs (#20985)

ggerganov committed 14 days ago

Verified 3fab96cd

jinja: fix macro with kwargs (#20960)

ngxson committed 14 days ago

Verified 914eb5ff

gguf-split : clarify operation of gguf-split (#19749)

alosslessdev committed 14 days ago

Verified 8fc17493

Newer Older