Kernels:

HelionDSL
/

attention

Trusted publisher

Kernel card Files Files and versions

attention / build /torch-cuda

18.1 kB

Ctrl+K

Ctrl+K

1 contributor

History: 4 commits

Dunfan's picture

Publish kernel (build/ + card + benchmarks); drop source/build-tooling files

b43b93e verified 10 days ago

attention_helion
Upload Helion attention kernel (noarch triton, pre-tuned sm100+sm90) 11 days ago
__init__.py
2.48 kB
Publish kernel (build/ + card + benchmarks); drop source/build-tooling files 11 days ago
_helion_aot_attention_cuda_sm100.py
1.02 kB
Publish kernel (build/ + card + benchmarks); drop source/build-tooling files 11 days ago
_helion_aot_attention_cuda_sm90.py
5.15 kB
Publish kernel (build/ + card + benchmarks); drop source/build-tooling files 10 days ago
_ops.py
856 Bytes
Upload Helion attention kernel (noarch triton, pre-tuned sm100+sm90) 11 days ago
attention.py
6.21 kB
Publish kernel (build/ + card + benchmarks); drop source/build-tooling files 11 days ago
layers.py
474 Bytes
Upload Helion attention kernel (noarch triton, pre-tuned sm100+sm90) 11 days ago
metadata.json
196 Bytes
Upload Helion attention kernel (noarch triton, pre-tuned sm100+sm90) 11 days ago
op.py
894 Bytes
Publish kernel (build/ + card + benchmarks); drop source/build-tooling files 11 days ago