View source on GitHub |
Converts a set of Q-value logits into Q-values using the provided support.
tf_agents.utils.common.convert_q_logits_to_values(
logits, support
)
Args | |
---|---|
logits
|
A Tensor representing the Q-value logits. |
support
|
The support of the underlying distribution. |
Returns | |
---|---|
A Tensor containing the expected Q-values. |