''

qq.py

-import torch
-import torch.nn as nn
-import torch.nn.functional as F
-import math
-import numpy as np
-
-
-# x = np.load('./THUCNews/data/vocab.embedding.sougou.npz')["embeddings"]
-# print(x[0])
-# print(len(x))
-# print(len(x[0]))
-
-# lis = []
-# f = open('./THUCNews/data/sgns.sogou.char', encoding='UTF-8')
-# for i, line in enumerate(f.readlines()):
-#     if i == 0:
-#         continue
-#     lin = line.split(' ')[0]
-#     lis.append(lin)
-# f.close()
-# print(lis)
-# if '<PAD>' in lis:
-#     print('yes')
-
-# if '<UNK>' in lis:
-#     print('yyyes')
-# import argparse
-# parser = argparse.ArgumentParser(description='TextCNN text classifier')
-# # learning
-# parser.add_argument('-lr', type=str, default='ppp', help='fff')
-# parser.add_argument('-oo', type=str, default=lr+'ooo', help='fff')
-# args = parser.parse_args()
-# print(args.oo)
-
-
-# dataset = 'THUCNews'
-# embedding = 'embedding_SougouNews.npz'  # random
-# embedding_pretrained = torch.tensor(np.load(dataset + '/data/' + embedding)["embeddings"].astype('float32'))
-# print(embedding_pretrained.size(1))
-# mode_name = 'models.TextCNN'
-# from models.+ mode_name import Config, TextCNN
-
-
-# w  = torch.randn(256)
-# H = torch.randn(128, 32, 256)
-# M = F.tanh(H)
-# out = torch.matmul(M, w)
-# print(out.size())
-
-
-x = torch.randn(128, 32, 256)
-y = torch.randn(256)
-z = torch.matmul(x, y)
-print(z.size())
-
-xx = torch.tensor([[2,3],[3,4]])
-yy = xx.unsqueeze(-1)
-print(xx)
-print(yy)
\ No newline at end of file