Note that a call to tfd_sample() without arguments will generate a single sample.

tfd_sample(distribution, sample_shape = list(), ...)

Arguments

distribution

The distribution being used.

sample_shape

0D or 1D int32 Tensor. Shape of the generated samples.

...

Additional parameters passed to Python.

Value

a Tensor with prepended dimensions sample_shape.

See also

Examples

# \donttest{ d <- tfd_normal(loc = c(1, 2), scale = c(1, 0.5)) d %>% tfd_sample()
#> tf.Tensor([1.2293875 1.8353437], shape=(2,), dtype=float32)
# }