testSimilarByWord('湖')
def testCi():
testSimilarByWord("道路")
testSimilarByWord("道家")
testSimilarByWord("道行")
def testVocab():
word = "月"
print("-5k-")
print(tools.getSimilarResult(model, word, vocab=5000))
print("-10k-")
print(tools.getSimilarResult(model, word, vocab=10000))
print("-20k-")
print(tools.getSimilarResult(model, word, vocab=20000))
if __name__ == "__main__":
model = tools.loadModel('sgns/sgns.literature.bigram-char.txt')
print("Lit model loaded")
# testCi();
# testZi()
testVocab()
else:
model = tools.loadModel('sgns/sgns.literature.bigram-char.txt')
print("LIT[word] Model loaded")
help='GPU ID (negative value indicates CPU)')
parser.add_argument('--input_num', '-in', type=int,default=30,
help='input num')
parser.add_argument('--next_day', '-nd', type=int,default=5,
help='predict next day')
parser.add_argument('--smooth_term', '-st', type=int,default=30,
help='timeperiod to smooth accuray')
parser.add_argument('--experiment_name', '-n', default='experiment', type=str,help='experiment name')
args = parser.parse_args()
input_num = args.input_num
#モデルの読み込み
model_1 = tools.loadModel('./train_result/20160523_2_vol2ema30class/final_model')
model_2 = tools.loadModel('./train_result/20160523_3_volrsistoch30class/final_model')
if args.gpu >= 0:
cuda.check_cuda_available()
print "use gpu"
model_1.to_gpu()
model_2.to_gpu()
xp = cuda.cupy if args.gpu >= 0 else np
START_TEST_DAY = 20090105
#START_TEST_DAY = 20100104
NEXT_DAY = args.next_day
batch = np.array(listbatch).astype(np.float32)
try:
x_batch = batch[:, :-output_num-2]
y_batch = batch[:, -output_num-2:-2]
except:
print (batch.shape)
print ("error!")
raw_input()
return x_batch, y_batch
model = tools.loadModel(args.modelpath)
if args.gpu >= 0:
cuda.check_cuda_available()
print "use gpu"
xp = cuda.cupy if args.gpu >= 0 else np
model.to_gpu()
N = sum(1 for line in open(trainfile))
print ('N = ', N)
N_test = sum(1 for line in open(testfile))
print ('N_test = ', N_test)
sum_loss = 0
for i in range(0,N,args.batchsize):
print 'checking train data... ', i,' / ',N
batch = read_batch2(trainfile,range(i,i+args.batchsize))
break
if newText not in history:
break
history += newText
if len(history) % 100 == 0:
print(len(history))
if len(history) < 4744:
chainVisualSimilarity(newText, history, limit=limit)
else:
print(len(history), history)
def testRelational():
result = model.most_similar('暖')
print(result)
result = model.most_similar(negative=['火'])
print(result)
result2 = model.most_similar(positive=['体', '本'], negative=['固'])
print(result2)
if __name__ == "__main__":
model = tools.loadModel('../embeddings/VC/v3.2/v3.2_embeddings_ep19.txt')
print(tools.getAnnoyIndex(model2, '情'))
else:
model = tools.loadModel('../embeddings/VC/v3.2/v3.2_embeddings_ep19.txt')
print("VISUAL Model loaded")
# -*- coding: utf-8 -*-
"""
@author: cqx931
2019
"""
import tools
def testVocab():
word = "情"
print(tools.getSimilarResult(model, word, vocab=5000))
print(tools.getSimilarResult(model, word, vocab=3000))
if __name__ == "__main__":
model = tools.loadModel('sgns/sgns.sikuquanshu.bigram.txt')
print(tools.testSimilarByWord(model, '品'))
# testVocab()
else:
model = tools.loadModel('zi/zi_embeddingsgns.literature.bigram-char.txt')
print("LIT Model loaded")
