sync : ggml#2528

Merged

ggerganov merged 31 commits intomasterfrom

sync

Nov 1, 2024

Commits on Oct 31, 2024

ggml-alloc : remove buffer_id from leaf_alloc (ggml/987)

danbev
authored and
ggerganov
committed
metal : single allocation of encode_async block (llama/9747)

ptsochantaris
and
ggerganov
committed
ggml : add metal backend registry / device (llama/9713)

ggerganov
and
slaren
committed
Update building for Android (llama/9672)

amqdn
authored and
ggerganov
committed
ggml : add backend registry / device interfaces to BLAS backend (llama/9752)

slaren
authored and
ggerganov
committed
ggml : fix BLAS with unsupported types (llama/9775)

slaren
authored and
ggerganov
committed
musa: add docker image support (llama/9685)

yeahdongcn
authored and
ggerganov
committed
rpc : add backend registry / device interfaces (llama/9812)

slaren
authored and
ggerganov
committed
ggml : move more prints to the ggml log system (llama/9839)

slaren
authored and
ggerganov
committed
Vectorize load instructions in dmmv f16 CUDA kernel (llama/9816)

authored and
ggerganov
committed
Fix cann compilation error (llama/9891)

leo-pony
authored and
ggerganov
committed
CUDA: fix 1D im2col, add tests (ggml/993)

JohannesGaessler
authored and
ggerganov
committed
fix: use `vm_allocate` to allocate CPU backend buffer on macOS (llama/9875)

giladgd
authored and
ggerganov
committed
fix: allocating CPU buffer with size `0` (llama/9917)

giladgd
authored and
ggerganov
committed
vulkan : add backend registry / device interfaces (llama/9721)

slaren
authored and
ggerganov
committed
add amx kernel for gemm (llama/8998)

mingfeima
authored and
ggerganov
committed
Add SYCL Backend registry, device and Event Interfaces (llama/9705)

OuadiElfarouki
authored and
ggerganov
committed
rpc : backend refactoring (llama/9912)

rgerganov
authored and
ggerganov
committed
fix mul_mat_vec_q and *_vec_q error (llama/9939)

authored and
ggerganov
committed
rpc : pack only RPC structs (llama/9959)

rgerganov
authored and
ggerganov
committed
ggml : add asserts for type conversion in fattn kernels (llama/9971)
ggerganov
committed
Adapt to dynamically loadable backends mechanism (llama/9970)

leo-pony
authored and
ggerganov
committed
metal : add POOL2D and fix IM2COL (llama/9943)

junhee-yoo
authored and
ggerganov
committed
increase cuda_cpy block size (ggml/996)

authored and
ggerganov
committed
CUDA: fix MMQ for non-contiguous src0, add tests (llama/10021)

JohannesGaessler
authored and
ggerganov
committed
CUDA: fix insufficient buffer clearing for MMQ (llama/10032)

JohannesGaessler
authored and
ggerganov
committed
metal : support permuted matrix multiplicaions (llama/10033)
ggerganov
committed
ggml : add AMX backend (llama/8998)

mingfeima
authored and
ggerganov
committed
sync : ggml
ggerganov
committed
talk-llama : sync llama.cpp
ggerganov
committed
whisper : backend registry init before model load
ggerganov
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

sync : ggml#2528

sync : ggml#2528
ggerganov merged 31 commits intomasterfrom
sync

Commits on Oct 31, 2024

ggml-alloc : remove buffer_id from leaf_alloc (ggml/987)

metal : single allocation of encode_async block (llama/9747)

ggml : add metal backend registry / device (llama/9713)

Update building for Android (llama/9672)

ggml : add backend registry / device interfaces to BLAS backend (llama/9752)

ggml : fix BLAS with unsupported types (llama/9775)

musa: add docker image support (llama/9685)

rpc : add backend registry / device interfaces (llama/9812)

ggml : move more prints to the ggml log system (llama/9839)

Vectorize load instructions in dmmv f16 CUDA kernel (llama/9816)

Fix cann compilation error (llama/9891)

CUDA: fix 1D im2col, add tests (ggml/993)

fix: use `vm_allocate` to allocate CPU backend buffer on macOS (llama/9875)

fix: allocating CPU buffer with size `0` (llama/9917)

vulkan : add backend registry / device interfaces (llama/9721)

add amx kernel for gemm (llama/8998)

Add SYCL Backend registry, device and Event Interfaces (llama/9705)

rpc : backend refactoring (llama/9912)

fix mul_mat_vec_q and *_vec_q error (llama/9939)

rpc : pack only RPC structs (llama/9959)

ggml : add asserts for type conversion in fattn kernels (llama/9971)

Adapt to dynamically loadable backends mechanism (llama/9970)

metal : add POOL2D and fix IM2COL (llama/9943)

increase cuda_cpy block size (ggml/996)

CUDA: fix MMQ for non-contiguous src0, add tests (llama/10021)

CUDA: fix insufficient buffer clearing for MMQ (llama/10032)

metal : support permuted matrix multiplicaions (llama/10033)

ggml : add AMX backend (llama/8998)

sync : ggml

talk-llama : sync llama.cpp

whisper : backend registry init before model load

sync : ggml#2528ggerganov merged 31 commits intomasterggml-org/whisper.cpp:masterfrom syncggml-org/whisper.cpp:syncCopy head branch name to clipboard

Commits

Commits on Oct 31, 2024

sync : ggml#2528
ggerganov merged 31 commits intomasterfrom
sync