ttnn.mac

ttnn.mac(input_tensor_a: ttnn.Tensor, input_tensor_b: ttnn.Tensor or Number, input_tensor_c: ttnn.Tensor or Number, *, memory_config: ttnn.MemoryConfig = None) → ttnn.Tensor

Computes Mac on input_tensor_a, input_tensor_b and input_tensor_c and returns the tensor with the same layout as input_tensor_a

Parameters:

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

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

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

Keyword Arguments:

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

Returns:

ttnn.Tensor – the output tensor.

Note

Supported dtypes, layouts, and ranks:

Dtypes

Layouts

Ranks

BFLOAT16, BFLOAT8_B

TILE

2, 3, 4

bfloat8_b/bfloat4_b supports only on TILE_LAYOUT

Example

# Create three tensors
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([[5, 6], [7, 8]], dtype=torch.bfloat16), layout=ttnn.TILE_LAYOUT, device=device
)
tensor3 = ttnn.from_torch(
    torch.tensor([[9, 10], [11, 12]], dtype=torch.bfloat16), layout=ttnn.TILE_LAYOUT, device=device
)

# Perform the mac operation
output = ttnn.mac(tensor1, tensor2, tensor3)
logger.info(f"MAC result: {output}")