Spaces:
Sleeping
Sleeping
Commit History
move BLAS to a separate backend (llama/6210)
c773aa9
tests : add non-cont unary tests (llama/7857)
6dc2887
vulkan: select only one device for single gpu with multiple drivers (llama/7582)
ee56a37
Update Vulkan RoPE implementation (llama/7818)
71850e7
vulkan : reuse parent extra for views (llama/7806)
b9b60de
ggml : refactor rope norm/neox (llama/7634)
ded0c68
Vulkan Mixture of Experts (MoE) support (llama/7628)
ad9ee26
vulkan: properly initialize vulkan devices for LLAMA_SPLIT_MODE_NONE (llama/7552)
da90a1e
Update vulkan rope implementation to support frequency factors (llama/7475)
be0ec58
llama : add phi3 128K model support (llama/7225)
ef68527
Vulkan Embedding Fix (llama/7360)
2bfeba3
Update and fix Vulkan soft_max and argsort implementations (llama/7237)
a0218a3
ggml : full ALiBi support (llama/7192)
192bda4
Vulkan Bugfixes and Improvements (llama/7084)
8dade62
ggml : add Flash Attention (llama/5021)
34d3b03
Vulkan k-quant mmq and ggml-backend offload functionality (llama/6155)
1ff7b08
unverified
sync : ggml (#2001)
cbbfa9e
unverified
ggml : remove old quantization functions (llama/5942)
11a2545
unverified
Vulkan Improvements (llama/5835)
ea2da45
unverified
ggml : introduce ggml_status (ggml/750)
151c676
unverified
ggml-vulkan: fix VULKAN_CHECK_RESULTS flag, which was previously broken (llama/5813)
472195f
unverified
ddpasa
commited on
make portability_enumeration_ext apple only (llama/5757)
c164918
unverified
Eve
commited on
code : normalize enum names (llama/5697)
93e0830
unverified
Introduce backend GUIDs (ggml/743)
a7eb9f6
unverified
UEXTM.com
slaren
commited on
Refactor validation and enumeration platform checks into functions to clean up ggml_vk_instance_init()
8637c17
unverified
Add check for VK_KHR_portability_enumeration for MoltenVK support
85caa3f
unverified
Add preprocessor checks for Apple devices.
b8e3b87
unverified
Resolve ErrorIncompatibleDriver with Vulkan on MacOS.
0bc3433
unverified
cmake : fix VULKAN and ROCm builds (llama/5525)
ae570e4
unverified
vulkan: Find optimal memory type but with fallback (llama/5381)
24e2319
unverified
vulkan: only use M-sized matmul on Apple GPUs (llama/5412)
350284e
unverified
Sergio López
commited on