Google I/O returns May 18-20! Reserve space and build your schedule Register now

tfp.substrates.jax.math.log_combinations

Log multinomial coefficient.

tfp.substrates.jax.math.log_combinations(
    n, counts, name='log_combinations'
)

Given n and counts, where counts has last dimension k, we define the multinomial coefficient as:

n! / prod_i n_i!

where i runs over all k classes.

This function computes the natural logarithm of the multinomial coefficient.

n Floating-point Tensor broadcastable with counts. This represents n outcomes.
counts Floating-point Tensor broadcastable with n. This represents counts in k classes, where k is the last dimension of the tensor.
name A name for this operation (optional).

log_combinations Tensor representing the log of the multinomial coefficient between n and counts.