ggerganov/llama.cpp

Pull Requests Commits

mtmd, llama: shared backend sched

ngxson committed 20 days ago

b6cf9cd8

vulkan: reduce iq1 shared memory usage for mul_mm (#24287)

jeffbolznv committed 20 days ago

Verified d6d0ce82

vulkan: add `v_dot2_f32_f16` support in matrix-matrix multiplication and Flash Attention (#24123)

0cc4m committed 20 days ago

Verified b4e3dc61

ui: Fix excessive style recalculation on hover (#24243)

ntowle committed 20 days ago

Verified ae735b13

mtmd: refactor video subproc handling (#24316)

ngxson committed 20 days ago

Verified 9682e351

server: log prompts to directory (#22031)

jacekpoplawski committed 20 days ago

Verified 1e912561

ui: fix mobile chat form overflow and bust stale bundle cache (#24158)

ServeurpersoCom committed 20 days ago

Verified efbacf8d

ggml : add GGML_OP_COL2IM_1D (#24206)

ServeurpersoCom committed 20 days ago

Verified 26021699

server : do not clear slots without unified KV cache (#24190)

fiesh committed 20 days ago

Verified 961e9a3e

models : fix plamo2 attention_key/value_length regression (#24317)

CISC committed 20 days ago

Verified f0152efe

ggml-cpu : fix rms_norm_back wrong output under in-place aliasing (#24305)

devYRPauli committed 20 days ago

Verified fd3271e0

Remove case for GGML_TYPE_Q4_K in mvvq.cu (#23528)

ravel7524 committed 21 days ago

Verified e3471b3e

ggml-webgpu: Add clang-format job (#24308)

reeselevine committed 21 days ago

Verified 3ac3c20c

ggml-webgpu: Improve prefill speeds for k-quants + refactor matmul for Q4/Q5/Q8 and k-quants (#24225)

yomaytk committed 21 days ago

Verified 1e1aca09

mtp: support for gemma-4 E2B and E4B assistants (#24282)

max-krasnyansky committed 21 days ago

Verified 7d2b45b4

server : do not parse when flushing http headers (#24281)

aldehir committed 21 days ago

Verified 42a0afd5

graph: guard iswa kq_mask on its own buffer (#24294)

ServeurpersoCom committed 21 days ago

Verified a66d5058

[ggml-webgpu] Handle buffer overlap / buffer aliasing for concat operator (#24000)

nikhilJain17 committed 21 days ago

Verified 1705d434

[ggml-webgpu] Implement 2D workgroups for scale, binary, and unary ops (#24044)

nikhilJain17 committed 21 days ago

Verified 3b3da01d

docker: install ffmpeg in the released image (#24302)

ngxson committed 21 days ago

Verified 3ebe862b

mtmd : add video input support (#24269)

ngxson committed 21 days ago

Verified 8f83d6c2

ggerganov committed 21 days ago

c2b1518f

ggml : bump version to 0.14.0 (ggml/1533)

ggerganov committed 21 days ago

6a1de6fb

cli: fix spinner not show during prompt processing (#24283)

ngxson committed 21 days ago

Verified 715b86a3

vulkan: Use cm2 decode_vector for mul_mat_id B matrix loads (#23991)

jeffbolznv committed 21 days ago

Verified c74759a2

cuda: reset cuda context after reading memory size (#23935)

0cc4m committed 21 days ago

Verified 0f7fada5

HIP: add gfx1152 and gfx1153 to RDNA3.5 (#24129)

harkgill-amd committed 21 days ago

Verified 19bba67c

metal : fix im2col 1D case (audio models) (#24220)

ngxson committed 22 days ago

Verified daf6bc9f

[SYCL] Update compute runtime version to 26.x in docker (#24070)

arthw committed 22 days ago

Verified d403f00e

common : relax sampler name matching (#23744)

ddh0 committed 22 days ago

Verified 9e3b928f

Older