def checkpoint(self, epoch, sess):
"""
Computes intrinsic scores for embeddings and dumps the embeddings embeddings
Parameters
----------
epoch: Current epoch number
sess: Tensorflow session object
Returns
-------
"""
embed_matrix, \
context_matrix = sess.run([self.embed_matrix, self.context_matrix])
voc2vec = {wrd: embed_matrix[wid] for wrd, wid in self.voc2id.items()}
embedding = Embedding.from_dict(voc2vec)
results = evaluate_on_all(embedding)
results = {key: round(val[0], 4) for key, val in results.items()}
curr_int = np.mean(list(results.values()))
self.logger.info('Current Score: {}'.format(curr_int))
if curr_int > self.best_int_avg:
self.logger.info("Saving embedding matrix")
f = open('{}/{}'.format(self.p.emb_dir, self.p.name), 'w')
for id, wrd in self.id2voc.items():
f.write('{} {}\n'.format(wrd, ' '.join([str(round(v, 6)) for v in embed_matrix[id].tolist()])))
self.saver.save(sess=sess, save_path=self.save_path)
self.best_int_avg = curr_int
def main():
embeddings, name = get_embeddings()
results = evaluate_on_all(embeddings)
out_fname = "{}.csv".format(name)
logger.info("Saving results...")
print(results)
results.to_csv(out_fname)
def run_job(j):
fn, kwargs = j
outf = path.join(opts.output_dir, fn + "_" + "_".join(str(k) + "=" + str(v) for k, v in iteritems(kwargs))) + ".csv"
logger.info("Processing " + outf)
if not path.exists(outf):
w = getattr(embeddings, fn)(**kwargs)
res = evaluate_on_all(w)
res.to_csv(outf)
def run_job(j):
fn, kwargs = j
outf = path.join(opts.output_dir, fn + "_" + "_".join(str(k) + "=" + str(v) for k,v in iteritems(kwargs))) + ".csv"
logger.info("Processing " + outf)
if not path.exists(outf):
w = getattr(embeddings, fn)(**kwargs)
res = evaluate_on_all(w)
res.to_csv(outf)
def benchmark(filepath, savepath, save_file_name):
from web.embeddings import load_embedding
w = load_embedding(filepath, format='word2vec')
# out_fname = os.path.join(savepath) + save_file_name + "_results.csv"
results = evaluate.evaluate_on_all(w)
# print(results['AP'].item())
for ind in results.keys():
print(ind,results[ind].item())
def evaluate(embed_matrix: dict, voc2id: dict) -> np.float:
"""
Computes intrinsic scores for embeddings and dumps the embeddings embeddings
Parameters
----------
epoch: Current epoch number
sess: Tensorflow session object
Returns
-------
"""
voc2vec = {wrd: embed_matrix[wid] for wrd, wid in voc2id.items()}
embedding = Embedding.from_dict(voc2vec)
results = evaluate_on_all(embedding)
results = {key: round(val[0], 4) for key, val in results.items()}
curr_int = np.mean(list(results.values()))
return curr_int
_, ext = os.path.splitext(fname)
if ext == ".bin":
format = "word2vec_bin"
elif ext == ".txt":
format = "word2vec"
elif ext == ".pkl":
format = "dict"
assert format in ['word2vec_bin', 'word2vec', 'glove',
'bin'], "Unrecognized format"
load_kwargs = {}
if format == "glove":
load_kwargs['vocab_size'] = sum(1 for line in open(fname))
load_kwargs['dim'] = len(next(open(fname)).split()) - 1
w = load_embedding(fname,
format=format,
normalize=True,
lower=False,
clean_words=options.clean_words,
load_kwargs=load_kwargs)
out_fname = options.output if options.output else "results.csv"
results = evaluate_on_all(w)
logger.info("Saving results...")
print(results)
results.to_csv(out_fname)
elif ext == ".txt":
format = "word2vec"
elif ext == ".pkl":
format = "dict"
assert format in ['word2vec_bin', 'word2vec',
'glove', 'bin'], "Unrecognized format"
load_kwargs = {}
if format == "glove":
load_kwargs['vocab_size'] = sum(1 for line in open(fname))
load_kwargs['dim'] = len(next(open(fname)).split()) - 1
print ("loading embedding...")
start = time.time()
w = load_embedding(fname, format=format, normalize=True, lower=True, clean_words=options.clean_words,
load_kwargs=load_kwargs)
elapsed_time = time.time() - start
print ("elapsed_time:{0}".format(elapsed_time) + "[sec]")
out_fname = options.output if options.output else "results.csv"
print ("Evaluating on embedding...")
start = time.time()
results = evaluate_on_all(w, options.entity, options.fastText_ML)
elapsed_time = time.time() - start
print ("elapsed_time:{0}".format(elapsed_time) + "[sec]")
logger.info("Saving results...")
results.to_csv(out_fname)
for c in set(compositional_dict[w]):
q_c += word_vectors[c]
q_c = q_c / len(word_vectors[c])
word_vectors[w] += q_c
word_vectors[w] = word_vectors[w] / 2
except KeyError:
pass
#print("before",w, Y_prev[w])
Y[w] -= word_vectors[w]
#print("after",w, Y[w])
return Y
Y = enhance_euclid(word_vectors,syn_dict,ant_dict,compositional_dict)
results_retrofitted = evaluate_on_all(Y)
print(results_retrofitted)
'''
{'AP': {0: 0.63681592039800994},
'BLESS': {0: 0.82000000000000006},
'Battig': {0: 0.41693748805199771},
'ESSLI_1a': {0: 0.77272727272727271},
'ESSLI_2b': {0: 0.82500000000000007},
'ESSLI_2c': {0: 0.64444444444444449},
'Google': {0: 0.10284486287351617},
'MEN': {0: 0.73746469698055173},
'MSR': {0: 0.086999999999999994},
'MTurk': {0: 0.63318199788472018},
'RG65': {0: 0.76952497886121318},
'RW': {0: 0.36701861264054064},
# assert format in ['word2vec_bin', 'word2vec', 'glove', 'bin'], "Unrecognized format"
# load_kwargs = {}
# if format == "glove":
# load_kwargs['vocab_size'] = sum(1 for line in open(fname))
# load_kwargs['dim'] = len(next(open(fname)).split()) - 1
# w = load_embedding(fname, format=format, normalize=True, lower=True, clean_words=options.clean_words,
# load_kwargs=load_kwargs)
if os.path.exists("results.csv"):
os.remove("results.csv")
out_fname = options.output if options.output else "results.csv"
results_sum = pd.DataFrame()
temporary_file = open("results.txt", 'w')
for word_embedding_name, w in iteritems(pretrained_word_embeddings):
results = evaluate.evaluate_on_all(w, word_embedding_name)
logger.info("Saving results... {}".format(word_embedding_name))
print("results:", results)
temporary_file.write(str(results) + '\n')
results_sum = results_sum.append(results)
results_sum.to_csv(out_fname)
temporary_file.close()
# results = evaluate_on_all(w)
# print(results)
# results.to_csv(out_fname)
assert format in [
'word2vec_bin', 'word2vec', 'glove', 'bin', 'dict', 'dict_poly'
], "Unrecognized format"
load_kwargs = {}
if format == "glove":
vocab_size = sum(1 for line in open(fname))
dim = len(next(open(fname)).split()) - 1
load_kwargs = {'dim': dim, 'vocab_size': vocab_size}
w = load_embedding(fname,
format=format,
normalize=options.normalize,
clean_words=options.clean_words,
lower=options.lowercase,
lowercase_if_OOV=options.lower_or_lemma,
lemmatize_if_OOV=options.lower_or_lemma,
load_kwargs=load_kwargs)
out_fname = options.output if options.output else "results.csv"
if options.multi_prototype:
results = evaluate_on_all_multi(w, options.model)
else:
results = evaluate_on_all(w, only_sim_rel=options.only_sim_rel)
logger.info("Saving results...")
print(results)
results.to_csv(out_fname)
if ext == ".bin":
format = "word2vec_bin"
elif ext == ".txt":
format = "word2vec"
elif ext == ".pkl":
format = "dict"
# assert format in ['word2vec_bin', 'word2vec', 'glove', 'bin'], "Unrecognized format"
load_kwargs = {}
if format == "glove":
vocab_size = sum(1 for line in open(fname))
dim = len(next(open(fname)).split()) - 1
load_kwargs = {"vocab_size": vocab_size, "dim": dim}
w = load_embedding(fname,
format=format,
normalize=True,
lower=True,
clean_words=options.clean_words,
load_kwargs=load_kwargs,
nonorm=options.nonorm,
nocheat=options.nocheat)
out_fname = options.output if options.output else "results.csv"
results = evaluate_on_all(w, options.nocheat)
logger.info("Saving results...")
print(results)
results.to_csv(out_fname)
fname = os.path.join(_get_dataset_dir(), fname)
format = options.format
if not format:
_, ext = os.path.splitext(fname)
if ext == ".bin":
format = "word2vec_bin"
elif ext == ".txt":
format = "word2vec"
elif ext == ".pkl":
format = "dict"
assert format in ['word2vec_bin', 'word2vec', 'glove', 'bin'], "Unrecognized format"
load_kwargs = {}
if format == "glove":
vocab_size = sum(1 for line in open(fname))
dim = len(next(open(fname)).split()) - 1
w = load_embedding(fname, format=format, normalize=True, lower=True, clean_words=options.clean_words,
load_kwargs=load_kwargs)
out_fname = options.output if options.output else "results.csv"
results = evaluate_on_all(w)
logger.info("Saving results...")
print(results)
results.to_csv(out_fname)
