ngxson/llama.cpp

Pull Requests Commits

arg: fix ASAN error on sampler_type_names empty

ngxson committed 33 days ago

5fb4235e

remove i_major_dual (#18157)

zhang-hui-yulo committed 33 days ago

Verified 54189c0d

webui: Fix selecting generated output issues during active streaming (#18091)

allozaur committed 33 days ago

Verified 9ce64aed

webui: fix chat screen shadow width (#18010)

polydecay committed 33 days ago

Verified 900316da

llama: offload output layer to GPU first (#18148)

JohannesGaessler committed 33 days ago

Verified 57c1e056

convert : sort and use file parts from model index if present (#18043)

CISC committed 33 days ago

Verified 9cff4cc5

llama : Async DirectIO model loading on Linux (#18012)

JTischbein committed 33 days ago

Verified 4d4f4cac

ggml-hexagon: swiglu_oai operation (#18114)

joeldushouyu committed 33 days ago

Verified 0a0bba05

convert : force patch_merger tensors to f16/f32 (#18124)

CISC committed 33 days ago

Verified 5166aaf8

server: (webui) add --webui-config (#18028)

ServeurpersoCom committed 33 days ago

Verified 6ce3d857

server: (router) disable SSL on child process (#18141)

ngxson committed 33 days ago

Verified e85e9d76

llama-fit-params: fix memory print (#18136)

JohannesGaessler committed 33 days ago

Verified 8dcc3662

webui: fix chat header width when sidebar is closed (#17981)

polydecay committed 33 days ago

Verified d37fc935

ggml-hexagon: gelu operation (#17921)

joeldushouyu committed 33 days ago

Verified 4470a076

common : restore grammar-based rejection sampling (#18137)

ggerganov committed 34 days ago

Verified 4301e273

common: clarify instructions for bug reports (#18134)

JohannesGaessler committed 34 days ago

Verified a2c199e4

model: fix GLM-ASR-Nano-2512 load error (#18130) (#18142)

HonestQiao committed 34 days ago

Verified 15dd67d8

server: (router) allow child process to report status via stdout (#18110)

ngxson committed 34 days ago

Verified bde461de

Extend run-org-model.py, add (a) batching (b) loading prompt from file (c) multimodal capacity (#18034)

pwilkin committed 34 days ago

Verified 8faa87db

Github: ask for -v logs for params_fit [no ci] (#18128)

JohannesGaessler committed 34 days ago

Verified 6f1f6a96

ggml-cpu: ARM64: repack version of q8_0 (dotprod and i8mm) (#18096)

Alcpz committed 34 days ago

Verified 669696e0

model: fix LFM2_MOE missing tensors (#18132)

tdakhran committed 34 days ago

Verified 982060fa

ci : clean up webui jobs (#18116)

CISC committed 34 days ago

Verified 6853bee6

common: fix --override-kv to support comma-separated values (#18056)

ServeurpersoCom committed 34 days ago

Verified 487674fb

HIP: Refactor mma for RDNA and CDNA (#17990)

zhang-hui-yulo committed 34 days ago

Verified acec774e

llama.android : Rewrite Android binding (w/o cpu_features dep) (#17413)

naco-siren committed 34 days ago

Verified 5c0d1888

arg: allow -kvu flag for llama-perplexity (#18117)

TrevorS committed 34 days ago

Verified 4b2a4778

ggml : use WARP_SIZE/2 for argmax reduction offset (#18092)

Aadeshveer committed 34 days ago

Verified 58062860

gguf-py : allow converting multi-tensor models from read-only locations (#18100)

ykhrustalev committed 34 days ago

Verified 2973a65e

llama-fit-params: force disable mlock (#18103)

JohannesGaessler committed 34 days ago

Verified d0794e89

Older