Skip to content

Commit

Permalink
Addresses #2
Browse files Browse the repository at this point in the history
  • Loading branch information
@pavanchhatpar
pavanchhatpar committed Jun 18, 2020
1 parent d25cd97 commit 55f1175
Showing 1 changed file with 6 additions and 142 deletions.
148 changes: 6 additions & 142 deletions copynet_tf/metrics.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -29,12 +29,14 @@
class BLEU(Metric):
def __init__(
self, name='bleu', max_order=4, ignore_tokens=[],
ignore_all_tokens_after=-1, smooth=False, **kwargs):
ignore_all_tokens_after=-1, smooth=False,
weights=(0.25, 0.25, 0.25, 0.25), **kwargs):
super(BLEU, self).__init__(name=name, **kwargs)
self.max_order = max_order
self.ignore_tokens = ignore_tokens
self.ignore_all_tokens_after = ignore_all_tokens_after
self.smooth = smooth
self.weights = weights

self.matches_by_order = self.add_weight(
name='matches_by_order', initializer='zeros',
Expand Down Expand Up @@ -192,7 +194,9 @@ def result(self):
i, tf.constant(0, tf.float32))
precisions = precisions.stack()
if tf.reduce_min(precisions) > 0:
p_log_sum = tf.reduce_mean(tf.math.log(precisions))
p_log_sum = tf.nn.weighted_moments(
tf.math.log(precisions), axes=-1,
frequency_weights=self.weights)
geo_mean = tf.exp(p_log_sum)
else:
geo_mean = tf.constant(0, tf.float32)
Expand All @@ -213,143 +217,3 @@ def reset_states(self):
zeros = tf.zeros((self.max_order,), tf.float32)
self.matches_by_order.assign(zeros)
self.possible_matches_by_order.assign(zeros)


# def _get_ngrams(segment, max_order):
# """Extracts all n-grams upto a given maximum order from an input segment.
# Args:
# segment: text segment from which n-grams will be extracted.
# max_order: maximum length in tokens of the n-grams returned by this
# methods.
# Returns:
# The Counter containing all n-grams upto max_order in segment
# with a count of how many times each n-gram occurred.
# """
# ngram_counts = collections.Counter()
# for order in range(1, max_order + 1):
# for i in range(0, len(segment) - order + 1):
# ngram = tuple(segment[i:i+order])
# ngram_counts[ngram] += 1
# return ngram_counts


# def _preprocess(
# reference_corpus, translation_corpus,
# ignore_tokens, ignore_all_tokens_after):
# if ignore_all_tokens_after is not None:
# reference_corpus_n = []
# for references in reference_corpus:
# references_n = []
# for reference in references:
# i = 0
# while (i < len(reference)
# and reference[i] != ignore_all_tokens_after):
# i += 1
# i += 1
# references_n.append(reference[:i])
# reference_corpus_n.append(references_n)
# reference_corpus = reference_corpus_n

# translation_corpus_n = []
# for translation in translation_corpus:
# i = 0
# while (i < len(translation)
# and translation[i] != ignore_all_tokens_after):
# i += 1
# i += 1
# translation_corpus_n.append(translation[:i])
# translation_corpus = translation_corpus_n

# if ignore_tokens is not None:
# reference_corpus_n = []
# for references in reference_corpus:
# references_n = []
# for reference in references:
# ref = []
# for token in reference:
# if token not in ignore_tokens:
# ref.append(token)
# references_n.append(ref)
# reference_corpus_n.append(references_n)
# reference_corpus = reference_corpus_n

# translation_corpus_n = []
# for translation in translation_corpus:
# trans = []
# for token in translation:
# if token not in ignore_tokens:
# trans.append(token)
# translation_corpus_n.append(trans)
# translation_corpus = translation_corpus_n
# return reference_corpus, translation_corpus


# def compute_bleu(reference_corpus, translation_corpus, max_order=4,
# smooth=False, ignore_tokens=None,
# ignore_all_tokens_after=None):
# """Computes BLEU score of translated segments against one or more
# references.
# Args:
# reference_corpus: list of lists of references for each translation.
# Each reference should be tokenized into a list of tokens.
# translation_corpus: list of translations to score. Each translation
# should be tokenized into a list of tokens.
# max_order: Maximum n-gram order to use when computing BLEU score.
# smooth: Whether or not to apply Lin et al. 2004 smoothing.
# Returns:
# 3-Tuple with the BLEU score, n-gram precisions, geometric mean of
# n-gram precisions and brevity penalty.
# """
# reference_corpus, translation_corpus = _preprocess(
# reference_corpus, translation_corpus,
# ignore_tokens, ignore_all_tokens_after)
# matches_by_order = [0] * max_order
# possible_matches_by_order = [0] * max_order
# reference_length = 0
# translation_length = 0
# for (references, translation) in zip(reference_corpus,
# translation_corpus):
# reference_length += min(len(r) for r in references)
# translation_length += len(translation)

# merged_ref_ngram_counts = collections.Counter()
# for reference in references:
# merged_ref_ngram_counts |= _get_ngrams(reference, max_order)
# translation_ngram_counts = _get_ngrams(translation, max_order)
# overlap = translation_ngram_counts & merged_ref_ngram_counts
# for ngram in overlap:
# matches_by_order[len(ngram)-1] += overlap[ngram]
# for order in range(1, max_order+1):
# possible_matches = len(translation) - order + 1
# if possible_matches > 0:
# possible_matches_by_order[order-1] += possible_matches

# precisions = [0] * max_order
# for i in range(0, max_order):
# if smooth:
# precisions[i] = ((matches_by_order[i] + 1.) /
# (possible_matches_by_order[i] + 1.))
# else:
# if possible_matches_by_order[i] > 0:
# precisions[i] = (float(matches_by_order[i]) /
# possible_matches_by_order[i])
# else:
# precisions[i] = 0.0

# if min(precisions) > 0:
# p_log_sum = sum((1. / max_order) * math.log(p) for p in precisions)
# geo_mean = math.exp(p_log_sum)
# else:
# geo_mean = 0

# ratio = float(translation_length) / reference_length

# if ratio > 1.0:
# bp = 1.
# else:
# bp = math.exp(1 - 1. / ratio)

# bleu = geo_mean * bp

# return (
# bleu, precisions, bp, ratio, translation_length, reference_length)

0 comments on commit 55f1175

Please sign in to comment.