ray.rllib.policy.eager_tf_policy_v2.EagerTFPolicyV2.loss#

EagerTFPolicyV2.loss(model: ModelV2 | tf.keras.Model, dist_class: Type[TFActionDistribution], train_batch: SampleBatch) → numpy.array | jnp.ndarray | tf.Tensor | torch.Tensor | List[numpy.array | jnp.ndarray | tf.Tensor | torch.Tensor][source]#

Compute loss for this policy using model, dist_class and a train_batch.

Parameters:

  • model – The Model to calculate the loss for.

  • dist_class – The action distr. class.

  • train_batch – The training data.

Returns:

A single loss tensor or a list of loss tensors.