flute_kernels / example.py
galqiwi's picture
Initial source: FLUTE kernel scaffold (vendored CUTLASS, split TUs)
67a5826 verified
"""Smoke example for galqiwi/flute_kernels."""
import torch
from kernels import get_kernel
flute = get_kernel("galqiwi/flute_kernels", version=1)
x = torch.randn(4, 4096, device="cuda", dtype=torch.float16)
y = flute.hadamard_transform(x, False)
print(y.shape, y.dtype)