ggerganov/llama.cpp

Pull Requests Commits

metal: add GDN partial rollback

ggerganov committed 11 hours ago

Verified 6eb6d84e

vulkan: add GDN partial rollback

am17an committed 11 hours ago

8c059236

fix pending state

am17an committed 11 hours ago

9a3a4872

llama: allow partial seq_rm for GDN models for speculative decoding

am17an committed 14 hours ago

4e732e0a

add need_embd in speculative

am17an committed 1 day ago

e7b48481

remove unused llama_arch

am17an committed 1 day ago

19dd00b0

mtp -> draft-mtp

am17an committed 1 day ago

f2200a3a

MTP: clean-up (#9)

am17an committed 1 day ago

3c3aebaa

cont : simplify (#7)

ggerganov committed 1 day ago

5c58cc5b

am17an committed 1 day ago

180f8ff3

am17an committed 1 day ago

edcaf120

spec: support MTP

am17an committed 1 day ago

a55493bb

spec : update CLI arguments for better consistency (#22964)

ggerganov committed 1 day ago

Verified 634275fb

llama-eval : enable type check (#22988)

CISC committed 1 day ago

Verified bcfe63fc

ggml-zendnn : adaptive fallback to CPU backend for small batch sizes (#22681)

z-sachin committed 1 day ago

Verified 61af07c2

hexagon: eliminate scalar VTCM loads via HVX splat helpers (#22993)

trivikram-reddy1 committed 1 day ago

Verified 856c3ada

opencl: add opt-in Adreno xmem F16xF32 GEMM for prefill (#22755)

happyyzy committed 1 day ago

Verified a9883db8

convert : fix Pixtral 12B --mistral-format conversion (3 bugs) (#22981)

fredzillman committed 1 day ago

Verified cce09f0b

webui: Fix Chat Screen Form box disappearing + autoscroll issues on WebKit (#22977)

allozaur committed 1 day ago

Verified dded58b4

mtmd, server, common: expose modalities to /v1/models (#22952)

ngxson committed 2 days ago

Verified 7bfe120c

ggml-webgpu: Enables running gpt-oss-20b (#22906)

yomaytk committed 2 days ago

Verified 927dada6

ggml-webgpu: address precision issues for multimodal (#22808)

Constannnnnt committed 2 days ago

Verified 239a497e

model-conversion : add causal-convert-mmproj target [no ci] (#22969)

danbev committed 2 days ago

Verified 89730c8d

examples : add llama-eval (#21152)

ggerganov committed 2 days ago

Verified fde69a36

vulkan: Fix Windows performance regression on Intel GPU BF16 workloads for Xe2 and newer (#22461)

rillomas committed 2 days ago

Verified ef93e98d

vulkan: Check shared memory size for mmq shaders (#22693)

jeffbolznv committed 2 days ago

Verified 706fbd8a

ci : bump ty to 0.0.35 (#22961)

CISC committed 2 days ago

Verified fa62042a

mtmd: add MiMo v2.5 vision (#22883)

AesSedai committed 2 days ago

Verified 41782591

convert : add split() to LoraTorchTensor in LoRA converter (#22832)

jesus-talavera-ibm committed 2 days ago

Verified 78fbbc2c

metal : promote mul_mv/mul_mm batch divisors to function constants (#22711)

guyfischman committed 2 days ago

Verified da449533

Older