ttnn.isclose

ttnn.isclose(input_tensor_a: ttnn.Tensor, input_tensor_b: ttnn.Tensor, *, rtol: float = 1e-05f, atol: float = 1e-08f, equal_nan: bool = False, memory_config: ttnn.MemoryConfig = None) → ttnn.Tensor

Computes isclose for input_tensor_a and input_tensor_b and returns the tensor with the same layout as input_tensor_a

\[\begin{split}\mathrm{output\_tensor} = \begin{cases} 1, & \text{if } |\mathrm{input\_tensor\_a} - \mathrm{input\_tensor\_b}| \leq (\mathrm{atol} + \mathrm{rtol} \times |\mathrm{input\_tensor\_b}|) \\ 0, & \text{otherwise} \end{cases}\end{split}\]
Parameters:

  • input_tensor_a (ttnn.Tensor) – the input tensor.

  • input_tensor_b (ttnn.Tensor) – the input tensor.

Keyword Arguments:

  • rtol (float) – relative tolerance. Defaults to 1e-05f.

  • atol (float) – absolute tolerance. Defaults to 1e-08f.

  • equal_nan (bool) – if NaN values should be treated as equal during comparison. Defaults to False.

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

Returns:

ttnn.Tensor – the output tensor.

Binary elementwise operations, C=op(A,B), support input tensors A and B in row major and tile layout, in interleaved or sharded format (height, width or block sharded), in DRAM or L1. A and B are completely independent, and can have different tensor specs.

Broadcast of A and B operands is supported up to dimension 5 (DNCHW). Any dimensions of size 1 in either A or B will be expanded to match the other input, and data will be duplicated along that dimension. For example, if the shape of A is [2,1,1,32] and B is [1,16,8,1], the output shape will be [2,16,8,32]. The size of dimensions higher than 5 must match between A and B.

The output C also supports row major and tile layout, interleaved or sharded format (height, width or block sharded), in DRAM or L1. The tensor spec of C is independent of A and B, and can be explicitly set using the optional output tensor input; if not provided, the operation will attempt a best decision at an appropriate tensor spec. The dimensions of C, or equivalently the optional output tensor, must match the broadcast-matched size of A and B.

Performance considerations: Elementwise operations operate natively in tile format, tiled tensors are preferred as an input, and row-major tensors are tilized and untilized during the operation. L1 sharded layout is preferred, with no broadcast and matching tensor specs for A, B and C.

Note

Supported dtypes, layouts, and ranks:

Dtypes

Layouts

Ranks

BFLOAT16

TILE

2, 3, 4

Example

>>> tensor1 = ttnn.from_torch(torch.tensor([[1, 2], [3, 4]], dtype=torch.bfloat16), layout=ttnn.TILE_LAYOUT, device=device)
>>> tensor2 = ttnn.from_torch(torch.tensor([[1, 2], [3, 4]], dtype=torch.bfloat16), layout=ttnn.TILE_LAYOUT, device=device)
>>> rtol = 1e-4
>>> atol = 1e-5
>>> equal_nan = False
>>> output = ttnn.isclose(tensor1, tensor2, rtol=rtol, atol=atol, equal_nan=equal_nan)