TileOPs-Metax/tileops/kernels/__init__.py

139 lines
4.0 KiB
Python

from .attention import (
FlashAttnBwdPostprocessKernel,
FlashAttnBwdPreprocessKernel,
GqaBwdKernel,
GqaBwdWgmmaPipelinedKernel,
GqaDecodeKernel,
GqaDecodePagedKernel,
GqaFwdKernel,
GqaFwdWgmmaPipelinedKernel,
GqaSlidingWindowFwdKernel,
GqaSlidingWindowFwdWgmmaPipelinedKernel,
GqaSlidingWindowVarlenFwdKernel,
GqaSlidingWindowVarlenFwdWgmmaPipelinedKernel,
MeanPoolingFwdKernel,
MhaBwdKernel,
MhaBwdWgmmaPipelinedKernel,
MhaDecodeKernel,
MhaDecodePagedKernel,
MhaFwdKernel,
MhaFwdWgmmaPipelinedKernel,
MlaDecodeKernel,
MlaDecodeWsKernel,
NSACmpFwdVarlenKernel,
NSAFwdVarlenKernel,
NSATopkVarlenKernel,
SparseMlaKernel,
)
from .convolution import Conv1dKernel, Conv2d1x1Kernel, Conv2dKernel, Conv3dKernel
from .deltanet import DeltaNetBwdKernel, DeltaNetFwdKernel
from .deltanet_recurrence import DeltaNetDecodeFP32Kernel, DeltaNetDecodeKernel
from .dropout import DropoutKernel
from .elementwise import BinaryKernel, FusedGatedKernel, UnaryKernel
from .engram import EngramDecodeKernel, EngramGateConvBwdKernel, EngramGateConvFwdKernel
from .fft import FFTC2CKernel
from .fp8_lighting_indexer import FP8LightingIndexerKernel
from .fp8_quant import FP8QuantKernel
from .gated_deltanet import GatedDeltaNetBwdKernel, GatedDeltaNetFwdKernel
from .gated_deltanet_recurrence import GatedDeltaNetDecodeFP32Kernel, GatedDeltaNetDecodeKernel
from .gemm import GemmKernel, GemvKernel
from .gla import GLABwdKernel, GLAFwdKernel
from .gla_recurrence import GLADecodeFP32Kernel, GLADecodeKernel
from .grouped_gemm import GroupedGemmKernel
from .kernel import Kernel
from .mhc import MHCPostKernel, MHCPreKernel
from .moe import MoePermuteAlignKernel
from .norm import (
BatchNormBwdKernel,
BatchNormFwdInferKernel,
BatchNormFwdTrainKernel,
GroupNormKernel,
LayerNormKernel,
RmsNormKernel,
)
from .pool import AvgPool1dKernel, AvgPool2dKernel, AvgPool3dKernel
from .rope import (
RopeLlama31Kernel,
RopeLongRopeKernel,
RopeNeoxKernel,
RopeNonNeoxKernel,
RopeYarnKernel,
)
from .topk_selector import TopkSelectorKernel
__all__ = [
"BinaryKernel",
"AvgPool1dKernel",
"AvgPool2dKernel",
"AvgPool3dKernel",
"BatchNormBwdKernel",
"BatchNormFwdInferKernel",
"BatchNormFwdTrainKernel",
"DropoutKernel",
"Conv1dKernel",
"Conv2d1x1Kernel",
"Conv2dKernel",
"Conv3dKernel",
"EngramDecodeKernel",
"EngramGateConvBwdKernel",
"EngramGateConvFwdKernel",
"FFTC2CKernel",
"FusedGatedKernel",
"FlashAttnBwdPostprocessKernel",
"FlashAttnBwdPreprocessKernel",
"FP8LightingIndexerKernel",
"FP8QuantKernel",
"GqaSlidingWindowFwdKernel",
"GqaSlidingWindowFwdWgmmaPipelinedKernel",
"GqaSlidingWindowVarlenFwdKernel",
"GqaSlidingWindowVarlenFwdWgmmaPipelinedKernel",
"DeltaNetBwdKernel",
"DeltaNetDecodeFP32Kernel",
"DeltaNetDecodeKernel",
"DeltaNetFwdKernel",
"GatedDeltaNetBwdKernel",
"GatedDeltaNetDecodeFP32Kernel",
"GatedDeltaNetDecodeKernel",
"GatedDeltaNetFwdKernel",
"GLABwdKernel",
"GLADecodeFP32Kernel",
"GLADecodeKernel",
"GLAFwdKernel",
"GemmKernel",
"GemvKernel",
"GqaBwdKernel",
"GqaBwdWgmmaPipelinedKernel",
"GqaFwdKernel",
"GqaFwdWgmmaPipelinedKernel",
"GroupNormKernel",
"Kernel",
"LayerNormKernel",
"MeanPoolingFwdKernel",
"MhaBwdKernel",
"MhaBwdWgmmaPipelinedKernel",
"MhaFwdKernel",
"MhaFwdWgmmaPipelinedKernel",
"NSACmpFwdVarlenKernel",
"NSAFwdVarlenKernel",
"NSATopkVarlenKernel",
"RmsNormKernel",
"RopeLlama31Kernel",
"RopeLongRopeKernel",
"RopeNeoxKernel",
"RopeNonNeoxKernel",
"RopeYarnKernel",
"SparseMlaKernel",
"TopkSelectorKernel",
"UnaryKernel",
"GqaDecodeKernel",
"GqaDecodePagedKernel",
"GroupedGemmKernel",
"MoePermuteAlignKernel",
"MHCPostKernel",
"MHCPreKernel",
"MhaDecodeKernel",
"MhaDecodePagedKernel",
"MlaDecodeKernel",
"MlaDecodeWsKernel",
]