ttnn.addcdiv_bw

ttnn.addcdiv_bw(grad_tensor: ttnn.Tensor, input_tensor_a: ttnn.Tensor, input_tensor_b: ttnn.Tensor, input_tensor_c: ttnn.Tensor, alpha: float, *, memory_config: ttnn.MemoryConfig | None = None) → List of ttnn.Tensor

Performs backward operations for addcdiv of input_tensor_a, input_tensor_b and input_tensor_c with given grad_tensor.

Parameters:

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

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

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

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

  • alpha (float) – the alpha value.

Keyword Arguments:

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

TILE

2, 3, 4

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

Example

>>> value = 1.0
>>> grad_tensor = ttnn.from_torch(torch.tensor([[1, 2], [3, 4]], dtype=torch.bfloat16), layout=ttnn.TILE_LAYOUT, device=device)
>>> tensor1 = ttnn.from_torch(torch.tensor([[1, 2], [3, 4]], dtype=torch.bfloat16, requires_grad=True), layout=ttnn.TILE_LAYOUT, device=device)
>>> tensor2 = ttnn.from_torch(torch.tensor([[1, 2], [3, 4]], dtype=torch.bfloat16, requires_grad=True), layout=ttnn.TILE_LAYOUT, device=device)
>>> tensor3 = ttnn.from_torch(torch.tensor([[1, 2], [3, 4]], dtype=torch.bfloat16, requires_grad=True), layout=ttnn.TILE_LAYOUT, device=device)
>>> output = ttnn.addcdiv_bw(grad_tensor, tensor1, tensor2, tensor3, value)