megha444

class BERT_Arch(nn.Module):

    def __init__(self, bert_model):
      
      super(BERT_Arch, self).__init__()

      self.bert = bert_model
     
      # relu activation function
      self.relu =  nn.ReLU()

megha444

# freeze all the parameters
for params in bert_model.parameters():
    params.requires_grad = False

megha444

from torch.utils.data import DataLoader, TensorDataset, SequentialSampler, RandomSampler

# wrap tensors
valdata = TensorDataset(valseq, valmask, valy)
b_size = 32
# sampler for sampling the data during training
valsampler = SequentialSampler(valdata)

# dataLoader for validation set
valDataLoader = DataLoader(valdata, sampler = valsampler, batch_size = b_size)

megha444

# tokenizing and encoding the sequences in the validation set
tokensval = tokenizer.batch_encode_plus(
    valtext.tolist(),
    max_length = 25,
    pad_to_max_length=True,
    truncation=True
)

# tokenizeing and encoding the sequences in the training set
tokenstrain = tokenizer.batch_encode_plus (

megha444

# get length of all the messages in the train set
sequence_len = [len(i.split()) for i in traintext]

pd.Series(sequence_len).hist(bins = 30)

megha444

# importing BERT-base pretrained model
bert_model= AutoModel.from_pretrained('bert-base-uncased')
# Loading the BERT tokenizer
bert_tokenizer = BertTokenizerFast.from_pretrained('bert-base-uncased')

#Tokenizer

# sample data
sample_text = ["this is a bert tutorial", "we will fine tune the bert model"]

megha444

# split train dataset into train, validation and test sets
traintext, temptext, trainlabels, templabels = train_test_split(df['text'], df['label'], 
                                                                    random_state=2018, 
                                                                    test_size=0.3, 
                                                                    stratify=df['label'])


valtext, testtext, vallabels, testlabels = train_test_split(temptext, templabels, 
                                                                random_state=2018, 
                                                                test_size=0.5, 

megha444

import torch
import pandas as pd
import torch.nn as nn
import transformers
import numpy as np
from sklearn.model_selection import train_test_split
from transformers import AutoModel, BertTokenizerFast
from sklearn.metrics import classification_report

# specify your GPU

megha444

train_data=new_dataset[:987]
valid_data=new_dataset[987:]
valid_data['Predictions']=predicted_closing_price
plt.plot(train_data["Close"])
plt.plot(valid_data[['Close',"Predictions"]])

megha444

lstm_model.save("saved_model.h5")