ggml-org/llama.cpp

Pull Requests Commits

Add tensor name to JSON output

pwilkin committed 15 days ago

22634e0e

tentative Metal support

pwilkin committed 16 days ago

2bfe4ff9

Add missing unrolls

pwilkin committed 16 days ago

28ef9417

Revert accidental change.

pwilkin committed 16 days ago

5ef996bd

pwilkin committed 16 days ago

1e47576c

Fix FATTN profiling

pwilkin committed 16 days ago

1b9b3e64

Converge implementation with export-graph-ops

pwilkin committed 16 days ago

56f349fd

Add missing op parameters to the profiler; add support for test-backend-ops to run performance tests with exactly the tensor shapes from the run

pwilkin committed 16 days ago

041605fd

docs, pass copy details

pwilkin committed 16 days ago

3f00bcd8

fix mul_mat_id stats, add throughput stat, add envvar trigger, add concurrent mode fix

pwilkin committed 16 days ago

61bb65d9

fix builds, integrate vulkan profiler, fix copy events, fix export

pwilkin committed 16 days ago

fbb7ceff

Fix more missing backend stuff (and Python errors)

pwilkin committed 16 days ago

28959252

add second dimension to reported tensors, fix Mac build, add missing initializer to all backends

pwilkin committed 16 days ago

4e927afd

feat: cool profiler thingy

pwilkin committed 16 days ago

893aa723

mtmd: support "frame merge" for qwen-vl-based models (#21858)

ngxson committed 16 days ago

Verified 31e82494

completion : remove useless statics (#24226)

angt committed 16 days ago

Verified 6b80c74f

completion : fix format specifier in LOG_INF (#24213)

angt committed 16 days ago

Verified 588f0dc2

mtmd, server: add "placeholder bitmap" for counting tokens , add */input_tokens API (#23913)

ngxson committed 16 days ago

Verified f5c6ae18

vulkan: check coopmat2 features before reporting support (#24186)

0cc4m committed 16 days ago

Verified 5a69c974

model : rename local n_layer_all variable (#24209)

CISC committed 16 days ago

Verified 5343f450

context : fix off-by-one comparisons to n_gpu_layers (#24208)

CISC committed 16 days ago

Verified 603300b0

opencl: improve get_rows, cpy, concat and q6_k flat gemv (#24160)

lhez committed 16 days ago

Verified 308f61c3

common/chat : unify and fix LFM2/LFM2.5 tool parser (#24178)

tdakhran committed 17 days ago

Verified da87e9b6

vulkan: add fwht support for Intel with shmem reduction (#23964)

0cc4m committed 17 days ago

Verified e82beaa6

model: fix build failed (#24193)

ngxson committed 17 days ago

Verified c4a278d6

model, mtmd: Granite4 Vision (#23545)

gabe-l-hart committed 17 days ago

Verified 64086f2b

TP: round up granularity to 128 (#24180)

JohannesGaessler committed 17 days ago

Verified 6effcecd

cli: fix model params not propagated (#23893)

therealkenc committed 17 days ago

Verified 86591c75

model : fix llama_model::n_gpu_layers() (#24188)

ggerganov committed 17 days ago

Verified 96fbe003

ui: run npm install when package-lock.json is newer than node_modules (#24171)

ServeurpersoCom committed 17 days ago

Verified 2016bf2b

Older