def _ctc_loss_with_beam_search(logits,
sparse_labels,
seq_length,
top_path=1,
merge_repeated=False):
ctc_loss = math_ops.reduce_mean(
ctc_ops.ctc_loss(sparse_labels, logits, seq_length))
pre_label_tensors, log_prob = tf.nn.ctc_beam_search_decoder(
logits, seq_length, merge_repeated=merge_repeated, top_paths=top_path)
top1_label_tensor = math_ops.cast(pre_label_tensors[0], dtypes.int32)
top1_ed = math_ops.reduce_mean(
array_ops.edit_distance(top1_label_tensor, sparse_labels))
return ctc_loss, top1_ed, pre_label_tensors, log_prob
def _testEditDistanceST(self,
hypothesis_st,
truth_st,
normalize,
expected_output,
expected_shape,
expected_err_re=None):
edit_distance = array_ops.edit_distance(
hypothesis=hypothesis_st, truth=truth_st, normalize=normalize)
if expected_err_re is None:
self.assertEqual(edit_distance.get_shape(), expected_shape)
output = edit_distance.eval()
self.assertAllClose(output, expected_output)
else:
with self.assertRaisesOpError(expected_err_re):
edit_distance.eval()
def _testEditDistanceST(self,
hypothesis_st,
truth_st,
normalize,
expected_output,
expected_shape,
expected_err_re=None):
edit_distance = array_ops.edit_distance(
hypothesis=hypothesis_st, truth=truth_st, normalize=normalize)
if expected_err_re is None:
self.assertEqual(edit_distance.get_shape(), expected_shape)
output = edit_distance.eval()
self.assertAllClose(output, expected_output)
else:
with self.assertRaisesOpError(expected_err_re):
edit_distance.eval()