|View source on GitHub|
Weighted cross-entropy loss for a sequence of logits.
tf.contrib.seq2seq.sequence_loss( logits, targets, weights, average_across_timesteps=True, average_across_batch=True, sum_over_timesteps=False, sum_over_batch=False, softmax_loss_function=None, name=None )
Depending on the values of
sum_over_batch, the return Tensor will have
rank 0, 1, or 2 as these arguments reduce the cross-entropy at each target,
which has shape
[batch_size, sequence_length], over their respective
dimensions. For example, if
False, then the return Tensor will have shape
sum_over_timesteps cannot be True
at same time. Same for
The recommended loss reduction in tf 2.0 has been changed to sum_over, instead
of weighted average. User are recommend to use
sum_over_batch for reduction.
logits: A Tensor of shape
[batch_size, sequence_length, num_decoder_symbols]and dtype float. The logits correspond to the prediction across all classes at each timestep.
targets: A Tensor of shape
[batch_size, sequence_length]and dtype int. The target represents the true class at each timestep.
weights: A Tensor of shape
[batch_size, sequence_length]and dtype float.
weightsconstitutes the weighting of each prediction in the sequence. When using
weightsas masking, set all valid timesteps to 1 and all padded timesteps to 0, e.g. a mask returned by
average_across_timesteps: If set, sum the cost across the sequence dimension and divide the cost by the total label weight across timesteps.
average_across_batch: If set, sum the cost across the batch dimension and divide the returned cost by the batch size.
sum_over_timesteps: If set, sum the cost across the sequence dimension and divide the size of the sequence. Note that any element with 0 weights will be excluded from size calculation.
sum_over_batch: if set, sum the cost across the batch dimension and divide the total cost by the batch size. Not that any element with 0 weights will be excluded from size calculation.
softmax_loss_function: Function (labels, logits) -> loss-batch to be used instead of the standard softmax (the default if this is None). Note that to avoid confusion, it is required for the function to accept named arguments.
name: Optional name for this operation, defaults to "sequence_loss".
A float Tensor of rank 0, 1, or 2 depending on the
average_across_batch arguments. By default,
it has rank 0 (scalar) and is the weighted average cross-entropy
(log-perplexity) per symbol.
ValueError: logits does not have 3 dimensions or targets does not have 2 dimensions or weights does not have 2 dimensions.