ttnn.prod_bw

ttnn.prod_bw(grad_tensor: ttnn.Tensor, input_tensor: ttnn.Tensor, *, all_dimensions: bool | None = True, dim: int | None = 0, memory_config: ttnn.MemoryConfig | None = None) → List of ttnn.Tensor

Performs backward operations for prod on input_tensor with given grad_tensor along all_dimensions or a particular dim.

Parameters:

• grad_tensor (ttnn.Tensor) – the input gradient tensor.

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

Keyword Arguments:

• all_dimensions (bool, optional) – perform prod backward along all dimensions, ignores dim param. Defaults to True.

• dim (int, optional) – dimension to perform prod backward. Defaults to 0.

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

Returns:

List of ttnn.Tensor – the output tensor.

Note

Supported dtypes, layouts, and ranks:

Dtypes	Layouts	Ranks
BFLOAT16	TILE	4

For more details about BFLOAT8_B, refer to the BFLOAT8_B limitations.

Example

>>> grad_tensor = ttnn.from_torch(torch.rand([1, 1, 32, 32], dtype=torch.bfloat16), layout=ttnn.TILE_LAYOUT, device=device)
>>> input = ttnn.from_torch(torch.rand([1, 1, 32, 32], dtype=torch.bfloat16, requires_grad=True), layout=ttnn.TILE_LAYOUT, device=device)
>>> all_dimensions = True
>>> dim =0
>>> output = ttnn.prod_bw(grad_tensor, input, all_dimensions, dim)