LossStrategy

Computes a DQN loss for a sampled batch, without performing the optimizer step.

Source

Methods

compute

compute(batch: LearnerBatch) -> LossResult

Source