Function skip_annotation

backends/qualcomm/utils/utils.py:604–791

r""" Exclude speific operators from quantizer annotation. Skipped operators will defaultly stay in CPU, set 'fallback_to_cpu' to False for trying to delegate them with FP16 precision. e.g.: consider following graph: bias_1 weight_1 input_1 bias_2 weight_2 input_2 | (plac

(
    nn_module: torch.nn.Module,
    quantizer,
    compiler_specs,
    sample_input: Tuple[torch.Tensor, ...],
    calibration_cb: Callable[[torch.fx.GraphModule], None],
    fp_node_id_set: set = None,
    fp_node_op_set: set = None,
    fallback_to_cpu: bool = True,
)

def skip_annotation(
    nn_module: torch.nn.Module,
    quantizer,
    compiler_specs,
    sample_input: Tuple[torch.Tensor, ...],
    calibration_cb: Callable[[torch.fx.GraphModule], None],
    fp_node_id_set: set = None,
    fp_node_op_set: set = None,
    fallback_to_cpu: bool = True,
):
    r"""
    Exclude speific operators from quantizer annotation.
    Skipped operators will defaultly stay in CPU, set 'fallback_to_cpu'
    to False for trying to delegate them with FP16 precision.

    e.g.: consider following graph:
    bias_1 weight_1 input_1   bias_2 weight_2 input_2
      | (placeholder) |         | (placeholder) |
       \      |      /           \      |      /
        \     |     /             \     |     /
         \    |    /               \    |    /
           conv2d_1                 conv2d_2
           (torch.ops.aten.conv2d.default)
               \                       /
                \                     /
                 \_______     _______/
                         add_1
             (torch.ops.aten.add.default)
                           |
                         output

    If user wants to skip convolution op by names with
    'skip_node_id_set' = {"conv2d_1"}
    "bias_1 / weight_1 / input_1 / input_2 / conv2d_1"
    will be partitioned out and not annotated / lowered with QNN.

    [Generated graph]
    bias_1 weight_1 input_1   input_2
      | (placeholder) |          |
       \      |      /           |
        \     |     /            |
         \    |    /             |
           conv2d_1              |
              \                 /
               \               /
                \             /
               lowered_module_1
            (QNN fixed precision)
                      |
                    output

    If user wants to skip convolution op by target with
    'skip_node_op_set' = {torch.ops.aten.conv2d.default}
    "bias_1 / weight_1 / input_1 / conv2d_1,
     bias_2 / weight_2 / input_2 / conv2d_2"
    will be partitioned out and not annotated / lowered with QNN.

    [Generated graph]