liangsu9988's picture
Promote latest kernel artifacts to main
02c100f verified
|
Raw History Blame Contribute Delete
328 Bytes

benchmarks

Package-level microbenchmarks for adaptive norm APIs.

Run:

python flashrt-adaptive-norms/benchmarks/benchmark.py --backend source --shapes all

The benchmark compares the fused FlashRT Tensor API against a PyTorch eager reference chain that performs the same adaptive norm and FP8 quantization math.