tfa.text.crf_log_norm

Stay organized with collections Save and categorize content based on your preferences.

Computes the normalization for a CRF.

tfa.text.crf_log_norm(
    inputs: tfa.types.TensorLike,
    sequence_lengths: tfa.types.TensorLike,
    transition_params: tfa.types.TensorLike
) -> tf.Tensor

inputs A [batch_size, max_seq_len, num_tags] tensor of unary potentials to use as input to the CRF layer.
sequence_lengths A [batch_size] vector of true sequence lengths.
transition_params A [num_tags, num_tags] transition matrix.

log_norm A [batch_size] vector of normalizers for a CRF.