operator

Classes

DepthwiseConv2DOperator

DepthwiseConv2DOperator(op: AirOperator, model: AirModel, platform: SocPlatform, prefix: str = 'aot', attributes: dict[str, str] = {})

DEPTHWISE_CONV_2D operator.

This operator performs depthwise convolution on the input tensor using the specified filter tensor.

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

ctx_buf_required

ctx_buf_required() -> bool

Whether the selected depthwise kernel dereferences ctx->buf.

The float lowering is deliberately excluded: when :meth:_resolve_float_scratch declines to allocate, the shape is one the float kernel serves from a route that never reads ctx->buf, so an unbacked context is a documented, supported state there, not a bug.

For the integer lowering the requirement is the selected kernel's, narrowed by the platform capability that makes upstream's get_buffer_size non-zero (see :data:_CTX_BUF_KERNELS).

Returns:

• bool –

  True if the selected kernel needs a non-NULL ctx.buf.

validate

validate()

Validate the operator.

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