Function clip

tensorrt_llm/functional.py:798–819

Add a CLIP operation that sets the range to [alpha, beta]. Parameters: input : Tensor The input tensor on which the activation function is applied. alpha : float The lower bound of the CLIP function. beta : float The upper bound

(input: Tensor, alpha: float, beta: float)

def clip(input: Tensor, alpha: float, beta: float) -> Tensor:
    '''
    Add a CLIP operation that sets the range to [alpha, beta].

    Parameters:
        input : Tensor
            The input tensor on which the activation function is applied.

        alpha : float
            The lower bound of the CLIP function.

        beta : float
            The upper bound of the CLIP function.

    Returns:
        The tensor produced by the activation layer.
    '''
    layer = default_trtnet().add_activation(input.trt_tensor,
                                            trt.ActivationType.CLIP)
    layer.alpha = alpha
    layer.beta = beta
    return _create_tensor(layer.get_output(0), layer)


relu = partial(activation, act_type=trt.ActivationType.RELU)