torch_to_nnef.op.aten

PyTorch Aten::* operators translation.

Modules:

Name	Description
`activation`
`attn`	Attention mechanisms.
`axes_change`
`complex`
`concat`
`expand`
`fft`	Aten FFT/STFT/ISTFT handlers mapping to NNEF ops.
`loss`	ATen loss-family op emitters (mse_loss, nll_loss, cross_entropy_loss, ...).
`math`
`matmul`
`norm`
`other`
`pad`
`pool`
`qops`
`reducer`
`rnn`	Aten-level RNN op handlers, adapters, and shared orchestration.
`selector`
`split`
`tensor_build`
`unary`

Functions:

Name	Description
`aten_to_nnef_tensor_and_ops`	Main primitive dispatcher.

aten_to_nnef_tensor_and_ops

aten_to_nnef_tensor_and_ops(g, node, name_to_tensor, null_ref, torch_graph, inference_target: InferenceTarget) -> T.Optional[T.List[str]]

Main primitive dispatcher.

Allow to write in graph any not Quantized Operation from pytorch defined in node attribute.