import numpy as np
import torch
from torch import nn
from torch.autograd import Variable
from torch.utils.data import TensorDataset, DataLoader, Dataset
import torch.nn.functional as F
from utils import *
from tqdm import tqdm
from trainDataloader import SimDataset, EvalSimDataset, EvalSimWithLabelDataset
from transformers import BertModel, BertConfig, BertTokenizer, BertForSequenceClassification
# %%
tokenizer = BertTokenizer.from_pretrained('./dataset/vocab')
eval_list = load_sim_dev('./dataset/101/c_dev_with_label')
myData_eval = EvalSimWithLabelDataset(tokenizer, './dataset/std_data', 100)
# %%
class SelfAttention(nn.Module):
def __init__(self, hidden_dim):
super().__init__()
self.hidden_dim = hidden_dim
self.projection = nn.Sequential(nn.Linear(hidden_dim, hidden_dim),
nn.ReLU(True),
nn.Linear(hidden_dim, hidden_dim))
def forward(self, encoder_outputs):
batch_size = encoder_outputs.size(0)
# (B, L, H) -> (B , L, 1)
energy = self.projection(encoder_outputs)
from utils import *
from tqdm import tqdm
from trainDataloader import SupremeSimDataset, EvalSimWithLabelDataset
from transformers import BertModel, BertConfig, BertTokenizer, BertForSequenceClassification
# %%
LABEL_ID = '1'
tokenizer = BertTokenizer.from_pretrained('./dataset/vocab')
myDataset = SupremeSimDataset(tokenizer, './dataset/supreme/l{}/s_train'.format(LABEL_ID), './dataset/std_data', 100)
myDataset.make_data()
dataiter = DataLoader(myDataset, batch_size=1024)
eval_list = load_sim_dev('./dataset/supreme/l{}/s_dev'.format(LABEL_ID))
myData_eval = EvalSimWithLabelDataset(tokenizer, './dataset/std_data', 100)
# %%
class SelfAttention(nn.Module):
def __init__(self, hidden_dim):
super().__init__()
self.hidden_dim = hidden_dim
self.projection = nn.Sequential(
nn.Linear(hidden_dim, hidden_dim),
nn.ReLU(True),
nn.Linear(hidden_dim, hidden_dim)
)
def forward(self, encoder_outputs):
batch_size = encoder_outputs.size(0)
# (B, L, H) -> (B , L, 1)