ttnn.hardtanh

ttnn.hardtanh(input_tensor: ttnn.Tensor, *, min_val: float = -1, max_val: float = 1, memory_config: ttnn.MemoryConfig = None) → ttnn.Tensor

Performs hardtanh function on input_tensor, min_val, max_val.

Parameters:

input_tensor (ttnn.Tensor) – the input tensor.

Keyword Arguments:

• min_val (float, optional) – min value. Defaults to -1.

• max_val (float, optional) – max value. Defaults to 1.

• 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
FLOAT32, BFLOAT16, BFLOAT8_B	TILE	2, 3, 4

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,
)

# Apply hard tanh activation function
output = ttnn.hardtanh(tensor, min_val=-1.0, max_val=1.0)
logger.info(f"Hard tanh: {output}")