File size: 6,151 Bytes
d45258a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195


import pandas as pd
import numpy as np
import torch
from transformers import RobertaTokenizer, RobertaForSequenceClassification
from torch import nn
from torch.nn import init, MarginRankingLoss
from transformers import BertModel, RobertaModel
from transformers import BertTokenizer, RobertaTokenizer
from torch.optim import Adam
from distutils.version import LooseVersion
from torch.utils.data import Dataset, DataLoader
from torch.utils.tensorboard import SummaryWriter
from datetime import datetime
from torch.autograd import Variable
from transformers import AutoConfig, AutoModel, AutoTokenizer
import nltk
import re
import Levenshtein
import spacy
import en_core_web_sm
import torch.optim as optim
from torch.distributions import Categorical
from numpy import linalg as LA
from transformers import AutoModelForMaskedLM
from nltk.corpus import wordnet
import torch.nn.functional as F
import random
from transformers import get_linear_schedule_with_warmup
from sklearn.metrics import precision_recall_fscore_support
from nltk.corpus import words as wal
from sklearn.utils import resample


# In[56]:


class MyDataset(Dataset):
    def __init__(self,file_name):
        df1 = pd.read_csv(file_name)
        df1 = df1.fillna("")
        res = df1['X']
        self.X_list = res.to_numpy()
        self.y_list = df1['y'].to_numpy()
    def __len__(self):
        return len(self.X_list)
    def __getitem__(self,idx):
        mapi = []
        mapi.append(self.X_list[idx])
        mapi.append(self.y_list[idx])
        return mapi


# In[59]:


class Step1_model(nn.Module):
    def __init__(self, hidden_size=512):
        super(Step1_model, self).__init__()
        self.hidden_size = hidden_size
        self.model = RobertaForSequenceClassification.from_pretrained("microsoft/graphcodebert-base", num_labels=6)
        self.tokenizer = AutoTokenizer.from_pretrained("microsoft/graphcodebert-base")
        self.config = AutoConfig.from_pretrained("microsoft/graphcodebert-base")
        for name, param in self.model.named_parameters():
            param.requires_grad = True
          

    def forward(self, mapi):
        X_init = mapi[0]
        X_init = X_init.replace("[MASK]", " ".join([tokenizer.mask_token] * 1))
        y = mapi[1]
        print(y)
        nl = re.findall(r'[A-Z](?:[a-z]+|[A-Z]*(?=[A-Z]|$))|[a-z]+|\d+', y)
        lb = ' '.join(nl).lower()
        x = tokenizer.tokenize(lb)
        nlab = len(x)
        print(nlab)
        tokens = self.tokenizer.encode_plus(X_init, add_special_tokens=False,return_tensors='pt')
        input_id_chunki = tokens['input_ids'][0].split(510)
        input_id_chunks = []
        mask_chunks  = []
        mask_chunki = tokens['attention_mask'][0].split(510)
        for tensor in input_id_chunki:
            input_id_chunks.append(tensor)
        for tensor in mask_chunki:
            mask_chunks.append(tensor)
        xi = torch.full((1,), fill_value=101)
        yi = torch.full((1,), fill_value=1)
        zi = torch.full((1,), fill_value=102)
        for r in range(len(input_id_chunks)):
            input_id_chunks[r] = torch.cat([xi, input_id_chunks[r]],dim = -1)
            input_id_chunks[r] = torch.cat([input_id_chunks[r],zi],dim=-1)
            mask_chunks[r] = torch.cat([yi, mask_chunks[r]],dim=-1)
            mask_chunks[r] = torch.cat([mask_chunks[r],yi],dim=-1)
        di = torch.full((1,), fill_value=0)
        for i in range(len(input_id_chunks)):
            # get required padding length
            pad_len = 512 - input_id_chunks[i].shape[0]
            # check if tensor length satisfies required chunk size
            if pad_len > 0:
                # if padding length is more than 0, we must add padding
                for p in range(pad_len):
                    input_id_chunks[i] = torch.cat([input_id_chunks[i],di],dim=-1)
                    mask_chunks[i] = torch.cat([mask_chunks[i],di],dim=-1)
        input_ids = torch.stack(input_id_chunks)
        attention_mask = torch.stack(mask_chunks)
        input_dict = {
            'input_ids': input_ids.long(),
            'attention_mask': attention_mask.int()
        }
        with torch.no_grad():
            outputs = self.model(**input_dict)
        last_hidden_state = outputs.logits.squeeze()
        lhs_agg = []
        if len(last_hidden_state) == 1:
            lhs_agg.append(last_hidden_state)
        else:
            for p in range(len(last_hidden_state)):
                lhs_agg.append(last_hidden_state[p])
        lhs = lhs_agg[0]
        for i in range(len(lhs_agg)):
            if i == 0:
                continue
            lhs+=lhs_agg[i]
        lhs/=len(lhs_agg)
    #    print(lhs)
        predicted_prob = torch.softmax(lhs, dim=0)
        if nlab > 6:
            nlab = 6
        pll = -1*torch.log(predicted_prob[nlab-1])
        return {'loss':pll}


# In[60]:


epoch_number = 0
EPOCHS = 5
run_int = 8
tokenizer = AutoTokenizer.from_pretrained("microsoft/graphcodebert-base")
model = Step1_model()
optimizer = optim.AdamW(model.parameters(), lr=2e-5)
myDs=MyDataset('dat1.csv')
train_loader=DataLoader(myDs,batch_size=1,shuffle=False)
best_loss = torch.full((1,), fill_value=100000)


# In[61]:


flag = 0
def train_one_epoch(transformer_model, dataset):
    global flag
    for batch in dataset:
        p = 0
        inputs = batch
        optimizer.zero_grad()
        for i in range(len(inputs[0])):
            l = []
            l.append(inputs[0][i])
            l.append(inputs[1][i])
            opi = transformer_model(l)
            loss = opi['loss']
        loss.backward()
        optimizer.step()
        if p % 1 == 0:
            print('  batch loss: {}'.format(loss))
    return loss


# In[62]:


for epoch in range(EPOCHS):
    print('EPOCH {}:'.format(epoch_number + 1))

    model.train(True)
    avg_loss = train_one_epoch(model,train_loader)
    model.train(False)
    print('LOSS train {}'.format(avg_loss))
    if avg_loss < best_loss:
        best_loss = avg_loss
    model_path = 'var_runs_class/model_{}_{}'.format(run_int, epoch_number)
    torch.save(model.state_dict(), model_path)

    epoch_number += 1


# In[ ]: