print("Note: You are using GPU for training")
torch.cuda.set_device(args.gpu)
torch.cuda.manual_seed(args.seed)
if torch.cuda.is_available() and not args.cuda:
print(
"Warning: You have Cuda but do not use it. You are using CPU for training"
)
# load word vocab for questions, relation vocab for relations
word_vocab = torch.load(args.vocab_file)
print('load word vocab, size: %s' % len(word_vocab))
rel_vocab = torch.load(args.rel_vocab_file)
print('load relation vocab, size: %s' % len(rel_vocab))
# load data
train_loader = SeqRankingLoader(args.train_file, args.gpu)
print('load train data, batch_num: %d\tbatch_size: %d' %
(train_loader.batch_num, train_loader.batch_size))
valid_loader = SeqRankingLoader(args.valid_file, args.gpu)
print('load valid data, batch_num: %d\tbatch_size: %d' %
(valid_loader.batch_num, valid_loader.batch_size))
os.makedirs(args.save_path, exist_ok=True)
# define models
config = args
config.n_cells = config.n_layers
if config.birnn:
config.n_cells *= 2
print(config)
#-*- coding: utf-8 -*- # Author: QuYingqi # mail: [email protected] # Created Time: 2017-11-09 import sys import torch from seqRankingLoader import SeqRankingLoader import numpy as np sys.path.append('../vocab') rel_vocab = torch.load('../vocab/vocab.rel.pt') neg_range = len(rel_vocab) print(neg_range) word_vocab = torch.load('../vocab/vocab.word.pt') loader = SeqRankingLoader('data/test.relation_ranking.pt', 5, neg_range, 0) batch_size = loader.batch_size for i, batch in enumerate(loader.next_batch(False)): if i >= 1: break seqs, pos_rel, neg_rel = batch seqs_trans = np.transpose(seqs.cpu().data.numpy()) pos_rel_trans = pos_rel.cpu().data.numpy() neg_rel_trans = np.transpose(neg_rel.cpu().data.numpy()) for j in range(5): question = ' '.join(word_vocab.convert_to_word(seqs_trans[j])) print(question) pos_rel_ = rel_vocab.convert_to_word([pos_rel_trans[j]]) print(pos_rel_) neg_rel_ = ' | '.join(rel_vocab.convert_to_word(neg_rel_trans[j])) print(neg_rel_)
torch.cuda.set_device(args.gpu)
torch.cuda.manual_seed(args.seed)
if torch.cuda.is_available() and not args.cuda:
print(
"Warning: You have Cuda but do not use it. You are using CPU for training"
)
# load word vocab for questions, relation vocab for relations
word_vocab = torch.load(args.vocab_file)
word_vocab.add_start_token() # 加了替换sub_text的分隔符
print('load word vocab, size: %s' % len(word_vocab))
rel_vocab = torch.load(args.rel_vocab_file)
print('load relation vocab, size: %s' % len(rel_vocab))
# load data
train_loader = SeqRankingLoader(args.train_file, len(rel_vocab), args.gpu)
print('load train data, batch_num: %d\tbatch_size: %d' %
(train_loader.batch_num, train_loader.batch_size))
valid_loader = SeqRankingLoader(args.valid_file, len(rel_vocab), args.gpu)
print('load valid data, batch_num: %d\tbatch_size: %d' %
(valid_loader.batch_num, valid_loader.batch_size))
os.makedirs(args.save_path, exist_ok=True)
# define models
config = args
config.n_cells = config.n_layers
if config.birnn:
config.n_cells *= 2
print(config)
print("Note: You are using GPU for training")
torch.cuda.set_device(args.gpu)
torch.cuda.manual_seed(args.seed)
if torch.cuda.is_available() and not args.cuda:
print(
"Warning: You have Cuda but do not use it. You are using CPU for training"
)
# load word vocab for questions, relation vocab for relations
word_vocab = torch.load(args.vocab_file)
print('load word vocab, size: %s' % len(word_vocab))
rel_vocab = torch.load(args.rel_vocab_file)
print('load relation vocab, size: %s' % len(rel_vocab))
# load data
train_loader = SeqRankingLoader(args.train_file, args.neg_size, len(rel_vocab),
args.gpu)
print('load train data, batch_num: %d\tbatch_size: %d' %
(train_loader.batch_num, train_loader.batch_size))
valid_loader = SeqRankingLoader(args.valid_file, args.neg_size, len(rel_vocab),
args.gpu)
print('load valid data, batch_num: %d\tbatch_size: %d' %
(valid_loader.batch_num, valid_loader.batch_size))
os.makedirs(args.save_path, exist_ok=True)
# define models
config = args
config.n_cells = config.n_layers
if config.birnn:
config.n_cells *= 2