-
Notifications
You must be signed in to change notification settings - Fork 130
/
eval_translation.py
172 lines (154 loc) · 7.63 KB
/
eval_translation.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
# Copyright (C) 2016-2018 Mikel Artetxe <artetxem@gmail.com>
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
import embeddings
from cupy_utils import *
import argparse
import collections
import numpy as np
import sys
BATCH_SIZE = 500
def topk_mean(m, k, inplace=False): # TODO Assuming that axis is 1
xp = get_array_module(m)
n = m.shape[0]
ans = xp.zeros(n, dtype=m.dtype)
if k <= 0:
return ans
if not inplace:
m = xp.array(m)
ind0 = xp.arange(n)
ind1 = xp.empty(n, dtype=int)
minimum = m.min()
for i in range(k):
m.argmax(axis=1, out=ind1)
ans += m[ind0, ind1]
m[ind0, ind1] = minimum
return ans / k
def main():
# Parse command line arguments
parser = argparse.ArgumentParser(description='Evaluate embeddings of two languages in a shared space in word translation induction')
parser.add_argument('src_embeddings', help='the source language embeddings')
parser.add_argument('trg_embeddings', help='the target language embeddings')
parser.add_argument('-d', '--dictionary', default=sys.stdin.fileno(), help='the test dictionary file (defaults to stdin)')
parser.add_argument('--retrieval', default='nn', choices=['nn', 'invnn', 'invsoftmax', 'csls'], help='the retrieval method (nn: standard nearest neighbor; invnn: inverted nearest neighbor; invsoftmax: inverted softmax; csls: cross-domain similarity local scaling)')
parser.add_argument('--inv_temperature', default=1, type=float, help='the inverse temperature (only compatible with inverted softmax)')
parser.add_argument('--inv_sample', default=None, type=int, help='use a random subset of the source vocabulary for the inverse computations (only compatible with inverted softmax)')
parser.add_argument('-k', '--neighborhood', default=10, type=int, help='the neighborhood size (only compatible with csls)')
parser.add_argument('--dot', action='store_true', help='use the dot product in the similarity computations instead of the cosine')
parser.add_argument('--encoding', default='utf-8', help='the character encoding for input/output (defaults to utf-8)')
parser.add_argument('--seed', type=int, default=0, help='the random seed')
parser.add_argument('--precision', choices=['fp16', 'fp32', 'fp64'], default='fp32', help='the floating-point precision (defaults to fp32)')
parser.add_argument('--cuda', action='store_true', help='use cuda (requires cupy)')
args = parser.parse_args()
# Choose the right dtype for the desired precision
if args.precision == 'fp16':
dtype = 'float16'
elif args.precision == 'fp32':
dtype = 'float32'
elif args.precision == 'fp64':
dtype = 'float64'
# Read input embeddings
srcfile = open(args.src_embeddings, encoding=args.encoding, errors='surrogateescape')
trgfile = open(args.trg_embeddings, encoding=args.encoding, errors='surrogateescape')
src_words, x = embeddings.read(srcfile, dtype=dtype)
trg_words, z = embeddings.read(trgfile, dtype=dtype)
# NumPy/CuPy management
if args.cuda:
if not supports_cupy():
print('ERROR: Install CuPy for CUDA support', file=sys.stderr)
sys.exit(-1)
xp = get_cupy()
x = xp.asarray(x)
z = xp.asarray(z)
else:
xp = np
xp.random.seed(args.seed)
# Length normalize embeddings so their dot product effectively computes the cosine similarity
if not args.dot:
embeddings.length_normalize(x)
embeddings.length_normalize(z)
# Build word to index map
src_word2ind = {word: i for i, word in enumerate(src_words)}
trg_word2ind = {word: i for i, word in enumerate(trg_words)}
# Read dictionary and compute coverage
f = open(args.dictionary, encoding=args.encoding, errors='surrogateescape')
src2trg = collections.defaultdict(set)
oov = set()
vocab = set()
for line in f:
src, trg = line.split()
try:
src_ind = src_word2ind[src]
trg_ind = trg_word2ind[trg]
src2trg[src_ind].add(trg_ind)
vocab.add(src)
except KeyError:
oov.add(src)
src = list(src2trg.keys())
oov -= vocab # If one of the translation options is in the vocabulary, then the entry is not an oov
coverage = len(src2trg) / (len(src2trg) + len(oov))
# Find translations
translation = collections.defaultdict(int)
if args.retrieval == 'nn': # Standard nearest neighbor
for i in range(0, len(src), BATCH_SIZE):
j = min(i + BATCH_SIZE, len(src))
similarities = x[src[i:j]].dot(z.T)
nn = similarities.argmax(axis=1).tolist()
for k in range(j-i):
translation[src[i+k]] = nn[k]
elif args.retrieval == 'invnn': # Inverted nearest neighbor
best_rank = np.full(len(src), x.shape[0], dtype=int)
best_sim = np.full(len(src), -100, dtype=dtype)
for i in range(0, z.shape[0], BATCH_SIZE):
j = min(i + BATCH_SIZE, z.shape[0])
similarities = z[i:j].dot(x.T)
ind = (-similarities).argsort(axis=1)
ranks = asnumpy(ind.argsort(axis=1)[:, src])
sims = asnumpy(similarities[:, src])
for k in range(i, j):
for l in range(len(src)):
rank = ranks[k-i, l]
sim = sims[k-i, l]
if rank < best_rank[l] or (rank == best_rank[l] and sim > best_sim[l]):
best_rank[l] = rank
best_sim[l] = sim
translation[src[l]] = k
elif args.retrieval == 'invsoftmax': # Inverted softmax
sample = xp.arange(x.shape[0]) if args.inv_sample is None else xp.random.randint(0, x.shape[0], args.inv_sample)
partition = xp.zeros(z.shape[0])
for i in range(0, len(sample), BATCH_SIZE):
j = min(i + BATCH_SIZE, len(sample))
partition += xp.exp(args.inv_temperature*z.dot(x[sample[i:j]].T)).sum(axis=1)
for i in range(0, len(src), BATCH_SIZE):
j = min(i + BATCH_SIZE, len(src))
p = xp.exp(args.inv_temperature*x[src[i:j]].dot(z.T)) / partition
nn = p.argmax(axis=1).tolist()
for k in range(j-i):
translation[src[i+k]] = nn[k]
elif args.retrieval == 'csls': # Cross-domain similarity local scaling
knn_sim_bwd = xp.zeros(z.shape[0])
for i in range(0, z.shape[0], BATCH_SIZE):
j = min(i + BATCH_SIZE, z.shape[0])
knn_sim_bwd[i:j] = topk_mean(z[i:j].dot(x.T), k=args.neighborhood, inplace=True)
for i in range(0, len(src), BATCH_SIZE):
j = min(i + BATCH_SIZE, len(src))
similarities = 2*x[src[i:j]].dot(z.T) - knn_sim_bwd # Equivalent to the real CSLS scores for NN
nn = similarities.argmax(axis=1).tolist()
for k in range(j-i):
translation[src[i+k]] = nn[k]
# Compute accuracy
accuracy = np.mean([1 if translation[i] in src2trg[i] else 0 for i in src])
print('Coverage:{0:7.2%} Accuracy:{1:7.2%}'.format(coverage, accuracy))
if __name__ == '__main__':
main()