espnet2.asr.transducer.rnnt_multi_blank.utils.rnnt_helper.exponential

source

espnet2.asr.transducer.rnnt_multi_blank.utils.rnnt_helper.exponential(x)

Compute the exponential of a given value.

This function calculates the exponential of the input value x using the mathematical constant e. It is designed to be used in a CUDA kernel, allowing for efficient computation on the GPU.

• Parameters:x (float) – The input value for which the exponential is to be computed.
• Returns: The exponential of the input value x, calculated as : e^x.
• Return type: float

Examples

>>> result = exponential(1.0)
>>> print(result)  # Output: 2.718281828459045

>>> result = exponential(0.0)
>>> print(result)  # Output: 1.0

>>> result = exponential(-1.0)
>>> print(result)  # Output: 0.36787944117144233

NOTE

This function should only be called within a CUDA kernel context.