228f5545 · Enable QD8 16x16c4 AMX GEMM microkernel · Mar 26, 2024

7ea191cb · Remove no longer needed `xnn_compute_pad_qd8` function and the... · Mar 25, 2024

8c0427b2 · Fix loading of quantization parameters in `NEON` `QS8` `GEMM` microkernels. · Mar 25, 2024

f3a76d67 · Remove check for AMXTILE. AMXTILE is required for amx kernels. · Mar 24, 2024

fa3685cf · Set weights stride in dynamic fully connected · Mar 19, 2024

5be0339f · Add breaks to prevent fallthough · Mar 19, 2024

1a2c0004 · Remove unrecognized build flag. · Mar 18, 2024

680703fd · Added static_transpose for slinky via copy · Mar 14, 2024

e814b27d · Generate QD8 AMX microkernels · Mar 12, 2024

7d0d9628 · Refactor automatically generated tests for `GEMM`. · Mar 12, 2024

fa2e1ee7 · Binary elementwise functions handle scalars correctly. · Mar 12, 2024

0f848f77 · Relax a tolerance that failed flakily here:... · Mar 11, 2024

23f33648 · Use correct flags to match static_mean test expected behavior. · Mar 11, 2024

885ed47f · Remove outdated `--cpu=ios_*`, `--cpu=darwin_*` settings. · Mar 08, 2024

f168f433 · XNNPack's qd8-f[16|32]-qc4w microkernels can accept either 4 signed 4 bit... · Mar 05, 2024

5c744aa5 · Values set as external inputs or outputs have their allocation type as... · Mar 05, 2024

59f437c5 · Add QS8 AVX512 AMX microkernel · Mar 05, 2024

72e5f3c0 · xnn_init_hardware_config: init cpuinfo on Windows Arm64 · Mar 04, 2024

bff32b1e · Add broadcasting to the Batch Matrix Multiply op. · Mar 04, 2024

598c1b24 · Remove shape checks for split - XNNpack can now handle this. · Mar 04, 2024