ggml-org/llama.cpp

Pull Requests Commits

fix: Post-review fixes

allozaur committed 5 days ago

18749134

server: read the GGUF metadata once per model

ServeurpersoCom committed 5 days ago

831d7e74

feat: Remove sorting from Status column in Models Table

allozaur committed 5 days ago

a872e623

fix: Favorite models grouping in models table

allozaur committed 5 days ago

5d44c8f3

allozaur committed 5 days ago

b1459be5

fix: UI text nit

allozaur committed 5 days ago

75080882

ui : mark favorite models with a heart

allozaur committed 5 days ago

9ec41545

ui : take the trained context from the models listing

allozaur committed 5 days ago

78e83c26

server : report the trained context in the models listing

allozaur committed 5 days ago

6a7da90b

common : read a GGUF's trained context from its metadata

allozaur committed 5 days ago

659b11c9

ci : set run-name for publish release (#30214)

CISC committed 6 days ago

Verified 5e1d7404

server: use port 9931 by default (#30159)

ngxson committed 6 days ago

Verified b42b7e6d

ui : add the models manager (#29583)

allozaur committed 6 days ago

Verified b013e56a

ci: fix HIP quality check by ignoring the 320/256 FA kernel spill (#30212)

ServeurpersoCom committed 6 days ago

Verified 2c564f43

metal: add the 128/96 flash attention kernels, accept every compiled pair (#30209)

ServeurpersoCom committed 6 days ago

Verified 1f8fa523

CUDA: pass src1 precision to host MMQ config helpers (#30168)

ynankani committed 6 days ago

Verified 8a1a9b51

hexagon: fix IM2COL patch-embed DMA ring overflow (#30189)

marktodorovich committed 6 days ago

Verified d4d82d67

sycl : Q5_K reorder-layout MMVQ and fused GLU (#29375)

NickM-27 committed 6 days ago

Verified 3d65c90d

Musa FWHT fix (#30167)

bri-prism committed 6 days ago

Verified de7fa0a3

CUDA: improve top-k algorithm selection (#28713)

praneshgo committed 6 days ago

Verified 71ad0590

model : fix DFlash output head sharing (#30111)

hthadicherla committed 6 days ago

Verified a11f57ba

CUDA: fix MMQ out-of-bounds reads (#29953)

JohannesGaessler committed 6 days ago

Verified fc9ce6b9

CUDA : looped PAD kernel for more than 65535 rows or slices (#30147)

pskrunner14 committed 6 days ago

Verified c35b6674

CUDA: fix CCCL version guard breaking on major version rollover (#29453)

hey-gm committed 6 days ago

Verified 1167d3f4

ui: apply ui_settings on first visit in router mode (#29668)

simonether committed 6 days ago

Verified 4f92965a

llama: support MoE cache over multiple GPUs (#30112)

am17an committed 7 days ago

Verified c811cb8f

server : preserve context checkpoints across slot save/restore (#26004)

Tough-Respawn committed 7 days ago

Verified 033df86b

vulkan : fix TOP_K for +inf/NaN inputs and k = 1 on negative values (#30107)

gianni-cor committed 7 days ago

Verified ff5888f9

vulkan: extend sparse FA support to coopmat2 (#30003)

jeffbolznv committed 7 days ago

Verified dac30873

vendor : update cpp-httplib to 0.60.1 (#30134)

angt committed 7 days ago

Verified 03aa006a

Older