|View source on GitHub|
log(1 / mean(1 / exp(input_tensor))).
tfp.substrates.numpy.math.reduce_log_harmonic_mean_exp( input_tensor, axis=None, keepdims=False, experimental_named_axis=None, name=None )
input_tensor along the dimensions given in
keepdims is true, the rank of the tensor is reduced by 1 for each entry in
keepdims is true, the reduced dimensions are retained with length
axis has no entries, all dimensions are reduced, and a tensor with a
single element is returned.
This function is more numerically stable than
log(1 / mean(1 - exp(input))).
It avoids overflows caused by taking the exp of large inputs and underflows
caused by taking the log of small inputs.
||The tensor to reduce. Should have numeric type.|
The dimensions to reduce. If
Boolean. Whether to keep the axis as singleton dimensions.
||The reduced tensor.|