|
|
|
|
|
|
|
|
|
|
|
|
|
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 |
|
|
|
|
|
|
|
|
|
|
|
class MyDataset(Dataset): |
|
def __init__(self,file_name): |
|
df1 = pd.read_csv(file_name) |
|
df1 = df1[230000:] |
|
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 |
|
|
|
|
|
|
|
|
|
|
|
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/codebert-base", num_labels=6) |
|
self.tokenizer = AutoTokenizer.from_pretrained("microsoft/codebert-base") |
|
self.config = AutoConfig.from_pretrained("microsoft/codebert-base") |
|
|
|
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)): |
|
pad_len = 512 - input_id_chunks[i].shape[0] |
|
if pad_len > 0: |
|
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]) |
|
|
|
pred = torch.argmax(predicted_prob).item() |
|
pred+=1 |
|
print(pred) |
|
predicted = torch.tensor([pred], dtype = float) |
|
if pred == nlab: |
|
l2 = 0 |
|
else: |
|
l2 = 1 |
|
actual = torch.tensor([nlab], dtype = float) |
|
l1 = Variable(torch.tensor([(actual-predicted)**2],dtype=float),requires_grad = True) |
|
return {'loss1':l1, 'loss2':l2} |
|
|
|
|
|
|
|
|
|
|
|
epoch_number = 0 |
|
EPOCHS = 5 |
|
run_int = 0 |
|
tokenizer = AutoTokenizer.from_pretrained("microsoft/codebert-base") |
|
model = Step1_model() |
|
myDs=MyDataset('dat_test.csv') |
|
train_loader=DataLoader(myDs,batch_size=2,shuffle=True) |
|
best_loss = torch.full((1,), fill_value=100000) |
|
|
|
|
|
|
|
|
|
|
|
flag = 0 |
|
def train_one_epoch(transformer_model, dataset): |
|
global flag |
|
tot_loss1 = 0.0 |
|
tot_loss2 = 0.0 |
|
cnt = 0 |
|
for batch in dataset: |
|
p = 0 |
|
inputs = batch |
|
for i in range(len(inputs[0])): |
|
cnt += 1 |
|
l = [] |
|
l.append(inputs[0][i]) |
|
l.append(inputs[1][i]) |
|
opi = transformer_model(l) |
|
loss1 = opi['loss1'] |
|
loss2 = opi['loss2'] |
|
tot_loss1 += loss1 |
|
tot_loss2 += loss2 |
|
|
|
tot_loss1/=cnt |
|
tot_loss2/=cnt |
|
print('MSE: ') |
|
print(tot_loss1) |
|
print('Acc: ',tot_loss2) |
|
return {'tot loss1': tot_loss1,'tot_loss2':tot_loss2} |
|
|
|
|
|
|
|
|
|
model.eval() |
|
avg_loss = train_one_epoch(model,train_loader) |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|