ttnn.sqrt_bw

ttnn.sqrt_bw(grad_tensor: ttnn.Tensor, input_tensor: ttnn.Tensor, *, memory_config: ttnn.MemoryConfig = None, output_tensor: ttnn.Tensor = None) → List of ttnn.Tensor

Performs backward operations for square-root on input_tensor with given grad_tensor.

Parameters:

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

• input_tensor (ttnn.Tensor) – the input 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:

List of ttnn.Tensor – the output tensor.

Note

Supported dtypes, layouts, and ranks:

Dtypes	Layouts	Ranks
TILE	For more details about BFLOAT8_B, refer to the BFLOAT8_B limitations.	2, 3, 4

Example

# Create sample tensors for backward square root operation
grad_tensor = ttnn.from_torch(
    torch.tensor([[1, 2], [3, 4]], dtype=torch.bfloat16), layout=ttnn.TILE_LAYOUT, device=device
)
input_tensor = ttnn.from_torch(
    torch.tensor([[1, 2], [3, 4]], dtype=torch.bfloat16, requires_grad=True), layout=ttnn.TILE_LAYOUT, device=device
)

# Call the sqrt_bw function
output = ttnn.sqrt_bw(grad_tensor, input_tensor)
logger.info(f"Square Root Backward: {output}")