0ffcaad3 · XNNPack will add the number of dims to axis/split-dims once known if required. · Mar 04, 2024

5cd847ab · Rolling back bad changelist · Mar 04, 2024

16247291 · Explicitly set num output dims for FC · Mar 03, 2024

afa3eb82 · Rename xnn_reshape_external_value to xnn_reshape_external_input as it only accepts inputs · Feb 28, 2024

5901f42b · Remove incorrect weights cache finalization check · Feb 27, 2024

1e6e268f · Refactor the unary NC operator tests to use a common base class, avoid code duplication. · Feb 27, 2024

29d78ba9 · Add iterative `vrsqrt` microkernels for SSE, AVX, FMA3, and AVX512f. · Feb 27, 2024

51899c06 · Remove some C4 AVXVNNI microkernels · Feb 26, 2024

7f615181 · XNNPack delegate support models with dynamic tensors and subgraph reshaping. · Feb 26, 2024

2d75c42f · Always allocate memory after a delegate has been applied. A node's init... · Feb 23, 2024

5da9a9ca · Add missing GFNI to template for future 4 bit optimized kernel · Feb 19, 2024

f3af4e7c · Enable QD8_F16_QC8W MRx8C8 AVXVNNI IGEMM microkernels · Feb 19, 2024

57c2c989 · QD8_F16_QC8W MRx8C8 AVXVNNI IGEMM microkernels · Feb 19, 2024

70385b39 · Add XNN_LIKELY branch hint for full stores on avx microkernels · Feb 19, 2024

c96e4573 · Fix missing generator for neon dot product kernel · Feb 19, 2024

7f492823 · Add xnn_setup_runtime_v2 which does not allocate memory · Feb 19, 2024

28830972 · xnn_reshape_runtime allocates memory if required. · Feb 19, 2024

1986e420 · Remove source that is not generated · Feb 18, 2024

0ffad834 · Enable QU8 GEMM/IGEMM templates · Feb 18, 2024

fa63020f · Clean up `vunary.h` a bit. · Feb 16, 2024