metal : separate scale and mask from QKT in FA kernel (#9189) #14822
build.yml
on: push
Matrix: windows-latest-cmake-cuda
Matrix: windows-latest-cmake
macOS-latest-cmake-arm64
3m 30s
macOS-latest-cmake-x64
10m 31s
ubuntu-focal-make
3m 57s
ubuntu-latest-cmake
2m 56s
macOS-latest-make
2m 2s
macOS-latest-cmake
2m 6s
ubuntu-focal-make-curl
3m 8s
ubuntu-latest-cmake-rpc
2m 24s
ubuntu-22-cmake-vulkan
3m 10s
ubuntu-22-cmake-hip
18m 47s
ubuntu-22-cmake-sycl
10m 21s
ubuntu-22-cmake-sycl-fp16
10m 29s
macOS-latest-cmake-ios
1m 55s
macOS-latest-cmake-tvos
2m 1s
windows-latest-cmake-sycl
12m 38s
windows-latest-cmake-hip
12m 9s
ios-xcode-build
2m 13s
android-build
16m 24s
Matrix: macOS-latest-swift
Matrix: ubuntu-latest-cmake-sanitizer
Matrix: windows-msys2
release
52s
Annotations
1 error and 9 warnings
Artifacts
Produced during runtime
Name | Size | |
---|---|---|
cudart-llama-bin-win-cu11.7.1-x64.zip
Expired
|
293 MB |
|
cudart-llama-bin-win-cu12.2.0-x64.zip
Expired
|
412 MB |
|
llama-bin-macos-arm64.zip
Expired
|
48.5 MB |
|
llama-bin-macos-x64.zip
Expired
|
50 MB |
|
llama-bin-ubuntu-x64.zip
Expired
|
54 MB |
|
llama-bin-win-avx-x64.zip
Expired
|
7.65 MB |
|
llama-bin-win-avx2-x64.zip
Expired
|
7.65 MB |
|
llama-bin-win-avx512-x64.zip
Expired
|
7.65 MB |
|
llama-bin-win-cu11.7.1-x64.zip
Expired
|
143 MB |
|
llama-bin-win-cu12.2.0-x64.zip
Expired
|
142 MB |
|
llama-bin-win-kompute-x64.zip
Expired
|
7.92 MB |
|
llama-bin-win-llvm-arm64.zip
Expired
|
11.2 MB |
|
llama-bin-win-msvc-arm64.zip
Expired
|
13.2 MB |
|
llama-bin-win-noavx-x64.zip
Expired
|
7.64 MB |
|
llama-bin-win-openblas-x64.zip
Expired
|
18.6 MB |
|
llama-bin-win-sycl-x64.zip
Expired
|
68.6 MB |
|
llama-bin-win-vulkan-x64.zip
Expired
|
8.24 MB |
|