def gen_embedding(word):
'''
Generates embedding of the word from the model trained.
'''
try:
with open('./embeddings.pickle', 'rb') as f:
embeddings = pickle.load(f)
return embeddings[word]
except Exception as e:
print ("Exception: Model file not found, please train the model first by runing train")
def gen_embedding(word):
'''
Generates embedding of the word from the model trained.
'''
try:
with open('./embeddings.pickle', 'rb') as f:
embeddings = pickle.load(f)
return embeddings[word]
except Exception as e:
print(
"Exception: Model file not found, please train the model first by runing train"
)
def closest_words(word,topn=10):
'''
Returns top 10 closest words to word provided by user.
'''
try:
with open("./embeddings.pickle", "rb") as f:
embeddings = pickle.load(f)
words = embeddings.keys()
closest = []
# Embedding of word provided by user
vec = embeddings[word]
for word in words:
heapq.heappush(closest, (1 - cosine(vec, embeddings[word]), word))
closest_words = heapq.nlargest(topn, closest)
return closest_words
except Exception as e:
print ("Exception: Check if model file is present, if not them train the model first, if present, then vocabulary issue, queried word not present in vocab")
def closest_words(word,topn=10):
'''
Returns top 10 closest words to word provided by user.
'''
try:
with open("./embeddings.pickle", "rb") as f:
embeddings = pickle.load(f)
words = embeddings.keys()
closest = []
# Embedding of word provided by user
vec = embeddings[word]
for word in words:
heapq.heappush(closest, (1 - cosine(vec, embeddings[word]), word))
closest_words = heapq.nlargest(topn, closest)
return closest_words
except Exception as e:
print ("Exception: Model file not found, please train the model by running train function")
def closest_words(word, topn=10):
'''
Returns top 10 closest words to word provided by user.
'''
try:
with open("./embeddings.pickle", "rb") as f:
embeddings = pickle.load(f)
words = embeddings.keys()
closest = []
# Embedding of word provided by user
vec = embeddings[word]
for word in words:
heapq.heappush(closest, (1 - cosine(vec, embeddings[word]), word))
closest_words = heapq.nlargest(topn, closest)
return closest_words
except Exception as e:
print(
"Exception: Check if model file is present, if not them train the model first, if present, then vocabulary issue, queried word not present in vocab"
)
def closest_words(word, topn=10):
'''
Returns top 10 closest words to word provided by user.
'''
try:
with open("./embeddings.pickle", "rb") as f:
embeddings = pickle.load(f)
words = embeddings.keys()
closest = []
# Embedding of word provided by user
vec = embeddings[word]
for word in words:
heapq.heappush(closest, (1 - cosine(vec, embeddings[word]), word))
closest_words = heapq.nlargest(topn, closest)
return closest_words
except Exception as e:
print(
"Exception: Model file not found, please train the model by running train function"
)