ggerganov/llama.cpp

Pull Requests Commits

conversion: refactor moe exp tensor handling

ngxson committed 26 days ago

051139f5

ci : fix dry-run reporting in make-release job [no ci] (#27167)

danbev committed 26 days ago

Verified 4df29be4

model : remove some ggml_concat (#27176)

fairydreaming committed 26 days ago

Verified 3cb7ffb1

chat: refactor handling supports_string_content / supports_typed_content (#27130)

ngxson committed 26 days ago

Verified b94041a9

llama : check LoRA tensor data is within file bounds (#27056)

oakkaya committed 27 days ago

Verified 10bf611e

ui: mask API Key field in settings and error splash to stop browser a… (#26562)

crowmoed committed 27 days ago

Verified ece963f4

ui: read structuredContent from MCP tool result when content is empty (#26691)

Gautam0507 committed 27 days ago

Verified 0d9ceae1

model: add Kimi-K3 text model (#26185)

pwilkin committed 27 days ago

Verified ad1de39e

server: re-design yield_to_queue thread model (#27133)

ngxson committed 27 days ago

Verified 22b8e310

vendor: update BoringSSL to 0.20260813.0 (#27099)

cabelo committed 27 days ago

Verified adb55e51

vendor : update cpp-httplib to 0.53.1 (#27103)

cabelo committed 27 days ago

Verified 77140d24

common: support --models-dir loading MTP assistant models (#24431)

EZForever committed 27 days ago

Verified 5f754ea0

fix: check gguf array type before reading (#27075)

ngxson committed 27 days ago

Verified 27df9199

vulkan: add SHMEM_STRIDE_PAD/APPLY_SLM_A_RESHAPE for coopmat1 on Intel Xe (#25380)

fish-jiang committed 28 days ago

Verified 9b0a2ce8

common: migrate the deprecated --mmap/--no-mmap to --load-mode (#26934)

fboudra committed 28 days ago

Verified 0177dcc7

rillomas committed 28 days ago

6b4344ec

Fixed gating logic for problematic Intel driver version

rillomas committed 28 days ago

7b38cb71

mtmd: fix Granite4 Vision image sequence assembly (#26653)

hbattu73 committed 28 days ago

Verified 9d57ce45

model : add support for MiniMaxText01ForCausalLM and MiniMaxM1ForCausalLM (#27018)

fairydreaming committed 28 days ago

Verified 16d222fc

mtmd, common: various fixes (#27071)

ngxson committed 28 days ago

Verified 6fed9f6f

jinja : fix quadratic cost in gather_string_parts (#27034)

Yunzez committed 28 days ago

Verified 9e40df63

chat : pass reasoning_effort to template

sobakasu committed 28 days ago

Verified 7e4c0a96

ggerganov committed 28 days ago

9b05354e

ggml : bump version to 0.20.0 (ggml/1584)

ggerganov committed 28 days ago

06ae2326

ggml : recurrent state rollback for ggml_ssm_scan (#26623)

lnigam committed 28 days ago

Verified 1692f9e5

llama : allow virtual igpu devices (#26953)

ggerganov committed 28 days ago

Verified 4c1a0af4

server: allow accessing /metrics and /slots during llama_decode() (#27041)

ngxson committed 28 days ago

Verified 77918caf

tests : replace personal home directory paths with generic placeholders (#27043)

jimw567 committed 29 days ago

Verified 885c5bbe

sycl: fuse mul_mat(gate) + mul_mat(up) + GLU for q4_K dense FFN (#26779)

Titaniumtown committed 29 days ago

Verified 65091386

ggml: force single thread on wasi (#25686)

MendyBerger committed 29 days ago

Verified c6f6a92c

Older