Reshape

Reshape

Changes the shape of a tensor without changing its data. Total element count must remain the same. Use -1 for one dimension to infer its size.

Parameters:

• target_shape: Tuple specifying the new shape (can include -1 for inference)

Shape Contract:

• Input: [*shape_in] tensor with original shape
• Output: [*shape_out] tensor with new shape (same total elements)

Notes:

• No learnable parameters (pure structural operation)
• Total elements must be preserved: prod(shape_in) == prod(shape_out)
• Use -1 for at most one dimension to auto-calculate
• Common use: flattening for FC layers, splitting/merging attention heads
• Data is not copied, only the view changes (memory efficient)

Signature

neuron Reshape(target_shape)

Ports

Inputs:

• default: [*shape_in]

Outputs:

• default: [*shape_out]

Implementation

Source { source: "core", path: "structural/Reshape" }