ttnn.identity

ttnn.identity = Operation(python_fully_qualified_name='ttnn.identity', function=<ttnn._ttnn.operations.unary.identity_t object>, preprocess_golden_function_inputs=<function default_preprocess_golden_function_inputs>, golden_function=<function register_ttnn_cpp_unary_function.<locals>._golden_function>, postprocess_golden_function_outputs=<function default_postprocess_golden_function_outputs>, is_cpp_operation=True, is_experimental=False)

Returns a copy of the input_tensor; useful for profiling the SFPU. This shouldn’t normally be used. Users should normally use clone operation instead for the same functionality since this results in lower performance.

\[\mathrm{output\_tensor}_i = \verb|identity|(\mathrm{input\_tensor}_i)\]
Parameters:

input_tensor (ttnn.Tensor) – the input tensor.

Keyword Arguments:

  • memory_config (ttnn.MemoryConfig, optional) – memory configuration for the operation. Defaults to None.

  • output_tensor (ttnn.Tensor, optional) – preallocated output tensor. Defaults to None.

  • queue_id (int, optional) – command queue id. Defaults to 0.

Returns:

ttnn.Tensor – the output tensor.

Note

Supported dtypes, layouts, and ranks:

Dtypes

Layouts

Ranks

BFLOAT16, BFLOAT8_B, FLOAT32, UINT32, UINT16, UINT8

TILE

2, 3, 4

Example

>>> tensor = ttnn.from_torch(torch.tensor([[1, 2], [3, 4]], dtype=torch.float16), dtype=ttnn.bfloat16, layout=ttnn.TILE_LAYOUT, device=device)
>>> output = ttnn.identity(tensor)