dropout

[zakarianamikaz]

python train_wc.py --train_file ./data/train_data.txt --dev_file ./data/dev_data.txt --test_file ./data/test_data.txt --caseless --fine_tune --high_way --co_train --least_iters 100 --dropout 0.5
setting:
Namespace(batch_size=10, caseless=True, char_dim=30, char_hidden=100, char_layers=1, checkpoint='./checkpoint/', clip_grad=5.0, co_train=True, dev_file='./data/dev_data.txt', dropout=0.5, emb_file='./embedding/glove.6B.100d.txt', epoch=200, eva_matrix='fa', fine_tune=False, gpu=0, high_way=True, highway_layers=2, lambda0=1, least_iters=100, load_check_point='', load_opt=False, lr=0.01, lr_decay=0.05, mini_count=5, momentum=0.9, patience=15, rand_embedding=False, small_crf=True, start_epoch=0, test_file='./data/test_data.txt', train_file='./data/train_data.txt', unk='unk', update='sgd', word_dim=100, word_hidden=100, word_layers=1)
loading corpus
constructing coding table
feature size: '3240'
loading embedding
embedding size: '3712'
constructing dataset
building model
Traceback (most recent call last):
File "train_wc.py", line 138, in
if_highway=args.high_way, in_doc_words=in_doc_words, highway_layers=args.highway_layers)
File "/home/usr/Downloads/arabic-ner-master/lmbilstmcrf/lm_lstm_crf.py", line 27, in init
dropout=dropout_ratio)
File "/home/usr/anaconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/rnn.py", line 425, in init
super(LSTM, self).init('LSTM', *args, **kwargs)
File "/home/usr/anaconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/rnn.py", line 31, in init
raise ValueError("dropout should be a number in range [0, 1] "
ValueError: dropout should be a number in range [0, 1] representing the probability of an element being zeroed