Commits

History for

ondev-pad-reflect

Commits on Jun 26, 2024

ggml : reorganize source code + improve CMake (ggerganov#865 )
ggerganov
authoredJun 26, 2024

Commits on Jun 16, 2024

Vulkan Shader Refactor, Memory Debugging Option (llama/7947)

0cc4m
authored and
ggerganov
committedJun 16, 2024

Commits on Jun 15, 2024

move BLAS to a separate backend (llama/6210)

slaren
and
ggerganov
committedJun 15, 2024
tests : add non-cont unary tests (llama/7857)
ggerganov
committedJun 15, 2024
vulkan: select only one device for single gpu with multiple drivers (llama/7582)

Adriankhl
authored and
ggerganov
committedJun 15, 2024
Update Vulkan RoPE implementation (llama/7818)

authored and
ggerganov
committedJun 15, 2024
vulkan : reuse parent extra for views (llama/7806)

authored and
ggerganov
committedJun 15, 2024
ggml : refactor rope norm/neox (llama/7634)
ggerganov
committedJun 15, 2024
Vulkan Mixture of Experts (MoE) support (llama/7628)

0cc4m
authored and
ggerganov
committedJun 15, 2024

Commits on May 29, 2024

vulkan: properly initialize vulkan devices for LLAMA_SPLIT_MODE_NONE (llama/7552)

Adriankhl
authored and
ggerganov
committedMay 29, 2024

Commits on May 28, 2024

Update vulkan rope implementation to support frequency factors (llama/7475)

0cc4m
authored and
ggerganov
committedMay 28, 2024
llama : add phi3 128K model support (llama/7225)

liuwei-git
and
ggerganov
committedMay 28, 2024
Vulkan Embedding Fix (llama/7360)

0cc4m
authored and
ggerganov
committedMay 28, 2024
Update and fix Vulkan soft_max and argsort implementations (llama/7237)

0cc4m
authored and
ggerganov
committedMay 28, 2024

Commits on May 11, 2024

ggml : full ALiBi support (llama/7192)
ggerganov
committedMay 11, 2024
Vulkan Bugfixes and Improvements (llama/7084)

0cc4m
authored and
ggerganov
committedMay 11, 2024
ggml : add Flash Attention (llama/5021)

committedMay 11, 2024

Commits on Apr 6, 2024

Vulkan k-quant mmq and ggml-backend offload functionality (llama/6155)

0cc4m
authored and
ggerganov
committedApr 6, 2024

Commits on Mar 27, 2024

llama : greatly reduce output buffer memory usage (llama/6122)

compilade
and
ggerganov
committedMar 27, 2024
backend : offload large batches to GPU (llama/6083)

authored and
ggerganov
committedMar 27, 2024
ggml:fix finding transfer queue family index error (llama/6094)

authored and
ggerganov
committedMar 27, 2024

Commits on Mar 14, 2024

llama : add pipeline parallelism support (llama/6017)

committedMar 14, 2024
ggml : remove old quantization functions (llama/5942)
ggerganov
committedMar 14, 2024

Commits on Mar 8, 2024

Vulkan Improvements (llama/5835)

0cc4m
authored and
ggerganov
committedMar 8, 2024

Commits on Mar 4, 2024

ggml : introduce ggml_status (ggerganov#750 )

authoredMar 4, 2024
ggml-vulkan: fix VULKAN_CHECK_RESULTS flag, which was previously broken (llama/5813)

ddpasa
authored and
ggerganov
committedMar 4, 2024
make portability_enumeration_ext apple only (llama/5757)

netrunnereve
authored and
ggerganov
committedMar 4, 2024

Commits on Feb 25, 2024

code : normalize enum names (llama/5697)
ggerganov
committedFeb 25, 2024

Commits on Feb 24, 2024

Introduce backend GUIDs (ggerganov#743 )

uextm
and
slaren
authoredFeb 24, 2024

Commits on Feb 21, 2024

Refactor validation and enumeration platform checks into functions to clean up ggml_vk_instance_init()

0cc4m
authored and
ggerganov
committedFeb 21, 2024
Add check for VK_KHR_portability_enumeration for MoltenVK support

0cc4m
authored and
ggerganov
committedFeb 21, 2024
Add preprocessor checks for Apple devices.

dokterbob
authored and
ggerganov
committedFeb 21, 2024
Resolve ErrorIncompatibleDriver with Vulkan on MacOS.

dokterbob
authored and
ggerganov
committedFeb 21, 2024

Commits on Feb 19, 2024

cmake : fix VULKAN and ROCm builds (llama/5525)
ggerganov
committedFeb 19, 2024
vulkan: Find optimal memory type but with fallback (llama/5381)

luciferous
authored and
ggerganov
committedFeb 19, 2024