ggml-org/llama.cpp

Pull Requests Commits

docs: Update documentation with Granite 4.0/4.1 (#23404)

jesus-talavera-ibm committed 134 days ago

Verified 95feeab5

ggml-zendnn : add Q8_0 quantization support (#23414)

z-sachin committed 134 days ago

Verified 99d4026b

cmake : build router app only during standalone builds (#23521)

fairydreaming committed 134 days ago

Verified 9c92e96a

vocab : fix HybridDNA tokenizer (#23466)

kashif committed 134 days ago

Verified afcda09d

cmake : add install() for impl libraries + fix apple builds (#23511)

ggerganov committed 134 days ago

Verified bbce619a

CUDA: fix PDL CC check for JIT compilation (#23471)

JohannesGaessler committed 134 days ago

Verified 4f0e43da

cmake : remove STATIC from impl libraries, enable LLAMA_BUILD_APP by default (#23462)

ggerganov committed 134 days ago

Verified bb28c1fe

Update WebGPU support and add link to blog/demo (#23483)

reeselevine committed 134 days ago

Verified ee7c3057

vulkan: fuse snake activation (mul, sin, sqr, mul, add) (#22855)

ServeurpersoCom committed 134 days ago

Verified 47c0eda9

fix(flash-attn): replace f32 with kv_type and q_type (#23372)

Constannnnnt committed 135 days ago

Verified 5306f4b3

tests : move save-load-state from examples to tests (#23336)

ggerganov committed 135 days ago

Verified 40d5358d

server: expose prompt token counts in /slots endpoint (#23454)

ScrewTSW committed 135 days ago

Verified b65bb4ba

metal : optimize concat kernel and fix set kernel threads (#23411)

ggerganov committed 135 days ago

Verified a1a69f77

server : free draft/MTP resources on sleep to fix VRAM leak (#23461)

am17an committed 135 days ago

Verified 52fb93a2

server: re-inject subcommand when router spawns children under unified binary (#23442)

ServeurpersoCom committed 135 days ago

Verified c9021714

app : add batched-bench, fit-params, quantize & perplexity (#23459)

angt committed 135 days ago

Verified 1d7ab2b9

mtp: use inp_out_ids for skipping logit computation (#23433)

am17an committed 135 days ago

Verified 12e5d990

vocab : add Carbon-3B (HybridDNATokenizer) support (#23410)

kashif committed 135 days ago

Verified 7ea23ddf

doc: fix spec mtp typo (#23435)

ruixiang63 committed 135 days ago

Verified 2fc8d185

ui: Improve Git Hooks for UI development (#23403)

allozaur committed 135 days ago

Verified 5e932a1c

ggml : Check the right iface method before using the fallback 2d get (#23306)

TheBlueMatt committed 135 days ago

Verified 2754ce1b

llama-graph: fix null-buffer crash in llm_graph_input_attn_kv_iswa for SWA-only models (#23131)

ssfdre38 committed 135 days ago

Verified eeeaf618

hexagon: ssm-conv fix for large prompts (#23307)

tboinovski1 committed 135 days ago

Verified 0be84685

app : show version (#23426)

angt committed 135 days ago

Verified ce02093f

mtmd, model : merge HunyuanOCR into HunyuanVL and fix OCR vision precision (#23329)

wendadawen committed 135 days ago

Verified 6a257d44

ui: Add max image size option (#22849)

stduhpf committed 135 days ago

Verified 3a479c91

Move to backend sampling for MTP draft path (#23287)

gaugarg-nv committed 135 days ago

Verified ad277572

opencl: refactor backend initilization (#23318)

lhez committed 135 days ago

Verified 3a6db741

common/speculative : fix nullptr crash in get_devices_str (#23386)

ggerganov committed 135 days ago

Verified 510b5c2a

mtmd : DeepSeek-OCR image processing fixes, img_tool::resize padding refactor (#23345)

sfallah committed 135 days ago

Verified a8681a0e

Newer Older