forked from rizwan09/LanModeledProgramGeneartion-master
-
Notifications
You must be signed in to change notification settings - Fork 0
/
main.py
162 lines (134 loc) · 7.13 KB
/
main.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
###############################################################################
# Author: Md Rizwan Parvez
# Project: LanModeledProgramGeneration
# Date Created: 4/1/2017
# Some codes are from Wasi Ahmad main.py
# File Description: This is the main script from where all experimental
# execution begins.
###############################################################################
import torch.nn as nn
import util, data #, helper, train
import torch, random
from torch import optim
import model_rnd
import time, math, os
# from encoder import EncoderRNN
from embedding_layer import Embedding_Drop_Layer
from torch.autograd import Variable
import train
args = util.get_args()
# args = get_args()
# Set the random seed manually for reproducibility.
print ('='*90, '\nWARNING:::: please fix nepochs, dictionary lower case, batchify, data path, batch size, trim_data for non divisible by batch size, Glove embedding initialization (model_rnd.py 45), pickle_file_name!!!!!\n', '='*89)
#print ('='*90, '\nWARNING:::: if you have insatnce based lstm you need to init model for each batch!!!!!\n', '='*89)
torch.manual_seed(args.seed)
if torch.cuda.is_available():
if not args.cuda:
print('='*90, "\nWARNING: You have a CUDA device, so you should probably run with --cuda\n", '='*89)
else:
#torch.cuda.set_device(2)
torch.cuda.manual_seed(args.seed)
###############################################################################
# Load data
###############################################################################
#### fix this
# corpus = data.Corpus(args.data)
corpus = data.Corpus(args)
print('Train set size = ', len(corpus.train_data), len(corpus.train_label))
# print('Development set size = ', len(corpus.dev))
# print('Test set size = ', len(corpus.test))
print('Vocabulary size = ', len(corpus.dictionary))
###############################################################################
# load_emb
###############################################################################
#### fix this
#file_name = 'train_corpus_3' + 'embeddings_index.p'
#file_name = 'PBT.p'
#embeddings_index = util.get_initial_embeddings(file_name, args.data_path, args.word_vectors_directory, args.Glove_filename, corpus.dictionary)
#print('Number of OOV words = ', len(corpus.dictionary) - len(embeddings_index))
###############################################################################
# batchify
###############################################################################
#### fix this
train_data_trimed, train_label_trimed = util.batchify(corpus.train_data, corpus.train_label, args.batch_size, args.cuda) #[82915, 20] batch size = 20, it's kinda seq len in gen sense
##### fix this
#valid_batches = util.batchify(corpus.valid, args.batch_size, args.cuda)
valid_data_trimed, valid_label_trimed = util.batchify(corpus.valid_data, corpus.valid_label, args.batch_size, args.cuda) #[82915, 20] batch size = 20, it's kinda seq len in gen sense
#### fix this
#test_batches = util.batchify(corpus.test, args.batch_size, args.cuda)
test_data_trimed, test_label_trimed = util.batchify(corpus.test_data, corpus.test_label, args.batch_size, args.cuda) #[82915, 20] batch size = 20, it's kinda seq len in gen sense
assert len(train_data_trimed) == len(train_label_trimed)
assert len(valid_data_trimed) == len(valid_label_trimed)
assert len(test_data_trimed) == len(test_label_trimed)
# print (batchify([2,3,4,3,4,355,4,342,90], 2))
print('train_batches: size: ', len(train_data_trimed) ) #, 'seq len: ', len(train_data_trimed[0]), '1st instance: ', train_data_trimed[0][:50], '1st label: ', train_label_trimed[0][:50] )# , train_batches[0][0].sentence1)
# ###############################################################################
# # Build the model
# ###############################################################################
model_f = model_rnd.LanguageModel(corpus.dictionary, args)
model_b = model_rnd.LanguageModel(corpus.dictionary, args)
if args.cuda:
torch.cuda.set_device(args.gpu)
model_f.cuda()
model_b.cuda()
# ###############################################################################
# # Dummy use the model
####fix this
# ###############################################################################
# list = [[4,14], [14,4]]
# l_t = torch.LongTensor(list)
# # list_var = Variable(l_t)
# # print('========================== before calling model forward', file = sys.stderr)
# print(model(l_t, l_t)[0][1])
# ###############################################################################
# # Train the model
# ###############################################################################
## loss: CrossEntropyLoss :: Combines LogSoftMax and NLLoss in one single class
if args.debug_mode:
train_data_trimed = train_data_trimed[:50]
valid_data_trimed = valid_data_trimed[:50]
test_data_trimed = test_data_trimed[:50]
#print(len(train_data_trimed))
#print(train_data_trimed[::-1][-1][::-1])
#exit()
train = train.Train(model_f, model_b, corpus.dictionary, 'CrossEntropyLoss')
train.train_epochs(train_data_trimed, train_label_trimed , valid_data_trimed, valid_label_trimed)
###############################################################################
# fix this
###############################################################################
#train_data = train_batches
#val_data = valid_batches
#test_data = test_batches
#ntokens = len(corpus.dictionary)
#model = model.RNNModel(args.model, ntokens, args.emsize, args.nhid, args.nlayers, args.dropout, args.tied)
#if args.cuda:
# model.cuda()
criterion = nn.CrossEntropyLoss(size_average=True)
###############################################################################
# testing code
###############################################################################
# Load the best saved model.
#with open(args.save, 'rb') as f:
# model = torch.load(f)
if os.path.isfile(os.path.join(args.log_dir, 'forward_model_best.pth.tar')) and os.path.isfile(os.path.join(args.log_dir, 'backward_model_best.pth.tar')):
print("=> Starting loading best models for testing")
checkpoint_forward= torch.load(os.path.join(args.log_dir, 'forward_model_best.pth.tar'))
checkpoint_backward= torch.load(os.path.join(args.log_dir, 'backward_model_best.pth.tar'))
args.start_epoch = checkpoint_forward['epoch']
best_perplexity_forward = checkpoint_forward['perplexity']
best_perplexity_backward = checkpoint_backward['perplexity']
model_f.load_state_dict(checkpoint_forward['state_dict'])
model_b.load_state_dict(checkpoint_backward['state_dict'])
print("=> Finished loading best models for testing")
# Run on test data.
#print ('using simple forward')
#util.evaluate(valid_data_trimed, valid_label_trimed , model_f, corpus.dictionary, criterion, 0, None, 'forward')
#print('-' * 89)
#print ('using simple backward')
#util.evaluate(valid_data_trimed, valid_label_trimed , model_b, corpus.dictionary, criterion, 0, None, 'backward')
#print('-' * 89)
test_loss = train.validate(test_data_trimed, test_label_trimed, model_f, model_b, args.nepochs, is_test = True)
print('=' * 89)
print('| End of training | test loss {:.2f} | test ppl {:5.2f} |'.format(
test_loss, math.exp(test_loss)))
print('=' * 89)