Skip to content

operator

Classes

TransposeConvOperator 

TransposeConvOperator(op: AirOperator, model: AirModel, platform: SocPlatform, prefix: str = 'aot', attributes: dict[str, str] = {})

TRANSPOSE_CONV operator.

This operator performs a transposed convolution operation on the input tensor with the given weight and bias tensors.

Parameters:

  • op

    (AirOperator) –

    The AIR operator to wrap.

  • model

    (AirModel) –

    The AIR model.

  • platform

    (SocPlatform) –

    The target platform for code generation.

  • prefix

    (str, default: 'aot' ) –

    Prefix for generated code files. Defaults to "aot".

  • attributes

    (dict[str, str], default: {} ) –

    Attributes for template values. Defaults to {}.

Functions

validate 
validate()

Validate the operator.

ctx_buf_required 
ctx_buf_required() -> bool

Whether the dispatched transpose-conv kernel dereferences ctx->buf.

Only the int8 kernel takes a buffer; the float lowering calls arm_transpose_conv_f32 / _f16, which discard the context entirely. Stated against the output dtype -- the same discriminator the template uses to choose the kernel -- rather than against the scratch byte count, so the guard compares two independent derivations.

Returns:

  • bool –

    True for the int8 lowering.

on_resolve 
on_resolve()

Resolve the operator.

compute_values 
compute_values() -> dict[str, str]

Compute the values for the operator.

Returns:

  • dict[str, str] –

    dict[str, str]: Computed values for the operator.

emit 
emit(save_path: Path)

Generate the source code for the operator.

Parameters:

  • save_path
    (Path) –

    Path to save the generated code.

Functions