hipudding 5aee53c40f CANN: Resolve soft_max precision issue (llama/15730) ...

Previously, the slope tensor was set to fp16 to improve efficiency.
While this worked correctly in FA, it caused precision issues in soft_max.
This change applies different data types for different operators
to balance both accuracy and performance.

2025-09-20 13:42:47 +03:00

..

cmake

ggml: Skip backend library linking code when GGML_BACKEND_DL=ON (llama/15094)

2025-08-18 20:30:45 +03:00

include

llama : separate compute buffer reserve from fattn check (llama/15696)

2025-09-20 13:42:45 +03:00

src

CANN: Resolve soft_max precision issue (llama/15730)

2025-09-20 13:42:47 +03:00

.gitignore

whisper : reorganize source code + improve CMake (#2256)

2024-06-26 19:34:09 +03:00

CMakeLists.txt

ggml: update kleidiai to v1.13.0 (llama/15663)

2025-09-20 13:42:44 +03:00