ArtificialRay
Kleidiai kernel changes:
d7a6eaa - 2.11 kB update kleidiai sme2 kernels
- 2.11 kB add baseline solution source at all neural network kernel definitions
- 2.11 kB add baseline solution source at all neural network kernel definitions
- 2.11 kB add baseline solution source at all neural network kernel definitions
- 2.11 kB add baseline solution source at all neural network kernel definitions
- 2.11 kB add baseline solution source at all neural network kernel definitions
- 1.72 kB Kleidiai kernel changes:
- 2.99 kB Update all w8a8ch definitions' correctness reference, aligns with what llama.cpp low-bit kernel correctness reference style: dequant input, then calculate in full precision
- 2.99 kB Update all w8a8ch definitions' correctness reference, aligns with what llama.cpp low-bit kernel correctness reference style: dequant input, then calculate in full precision
- 2.99 kB Update all w8a8ch definitions' correctness reference, aligns with what llama.cpp low-bit kernel correctness reference style: dequant input, then calculate in full precision
- 2.99 kB Update all w8a8ch definitions' correctness reference, aligns with what llama.cpp low-bit kernel correctness reference style: dequant input, then calculate in full precision
- 2.99 kB Update all w8a8ch definitions' correctness reference, aligns with what llama.cpp low-bit kernel correctness reference style: dequant input, then calculate in full precision
