tf.contrib.crf.crf_log_norm

View source on GitHub

Computes the normalization for a CRF.

tf.contrib.crf.crf_log_norm(
    inputs,
    sequence_lengths,
    transition_params
)

Args:

• 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.

Returns:

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