ttnn.fill

ttnn.fill(input_tensor: ttnn.Tensor, fill_value: float / int, *, memory_config: ttnn.MemoryConfig = None, output_tensor: ttnn.Tensor = None) → ttnn.Tensor

Applies fill to input_tensor element-wise with fill_value.

This will create a tensor of same shape and dtype as input reference tensor with fill_value.

\[\mathrm{output\_tensor}_i = \verb|fill|(\mathrm{input\_tensor}_i, \verb|fill_value|)\]

Parameters:

input_tensor (ttnn.Tensor) – the input tensor.
fill_value (float/int) – The value to be filled in the output 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.

Returns:

ttnn.Tensor – the output tensor.

Note

Supported dtypes, layouts, and ranks:

Dtypes	Layouts	Ranks
BFLOAT16, BFLOAT8_B, FLOAT32, INT32, UINT32	TILE	1, 2, 3, 4, 5, 6

Host memory is not supported.

Example

# Create a tensor with specific values
tensor = ttnn.from_torch(
    torch.tensor([[1, 2], [3, 4]], dtype=torch.bfloat16),
    dtype=ttnn.bfloat16,
    layout=ttnn.TILE_LAYOUT,
    device=device,
)
fill_value = 3

# Fill tensor with a specific value
output = ttnn.fill(tensor, fill_value)
logger.info(f"Fill: {output}")