submission 66740
kaide_max · python · License unknown
Use it
Vendorable · source mirrored · license unknownView source →
No package. Vendor the mirrored source: 9 lines, June 9 Researcher Reciprocity License v1.0.
grayscale.py
curl "https://kernelindex.com/api/v1/implementations/kernelbot-grayscale-v2-66740?include=source"interfacepython
Compatibility
measured onNVIDIA B200
declared hardwareNVIDIA B200
architecturessm_100
dtypesfp32
Benchmark evidence
1 measurement across 1 GPU, fastest first.
Operation / workload
Hardware
Latency
Rank
Observed
Reported · How evidence levels are derived →
Source and license
sourceavailable
revision digestsha256:fbbcc93c4f45204b67301b745478d96a6019bbc2a68ea0ed4139724e509aecd7
license declaredunknown
license concludedunknown
authorskaide_max
imported2026-08-15
Kernel source
grayscale.py9 lines
import torch
def custom_kernel(input):
# Unpack if tuple (CLI spec)
input_tensor = input[0] if isinstance(input, tuple) else input
coeffs = torch.tensor([0.299, 0.587, 0.114], device=input_tensor.device, dtype=input_tensor.dtype)
gray = torch.sum(input_tensor * coeffs.view(1, 1, 3), dim=2) # Fused weighted sum
return gray
Source code from GPU Mode and the KernelBot dataset · June 9 Researcher Reciprocity License v1.0
Best evidence level for this revision: reported
JSON