Update inference

import argparse

import torch

import preprocess

import train

import models

from transformers import BertTokenizer, RobertaTokenizer, BertModel, RobertaModel, RobertaPreTrainedModel, RobertaConfig,  BertConfig, BertPreTrainedModel, PreTrainedModel, AutoConfig, AutoModel, AutoTokenizer

import re 

import models

import train

from torch.utils.data import DataLoader, RandomSampler

# Get user input

print("Enter a sentence: ")

sentence = input()

sentence = sentence.split()

print("Enter the label: 0 for literal, 1 for non-literal")

label = int(input())

data_sample = {"sentence": sentence, "pos": target_pos, "label": label}

print(data_sample)

filepath = "./saved_models/bert_baseline.pt"

model=models.BertForWordClassification.from_pretrained("bert-base-uncased")

#tokenizer=AutoTokenizer.from_pretrained(args.architecture)

# Load model

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

model.load_state_dict(torch.load(filepath))

model.eval()		#loads saved model

model=models.BertForWordClassification.from_pretrained("bert-base-uncased")

model_path = "saved_models/bert_baseline.pth"

model = torch.load(model_path, map_location=device)

train_dataset = [{"sentence": ["Yet", "how", "many", "times", "has", "America", "sided", "with", "Israeli", "aggression", "against", "the", "people", "of", "Palestine?"], "pos": [5, 6], "label": 1}]

train_sampler = RandomSampler(train_dataset)

train_dataloader=DataLoader(train_dataset, sampler=train_sampler, batch_size=1)

tokenizer=AutoTokenizer.from_pretrained("bert-base-uncased")

train_sampler = RandomSampler(data_sample)

train_dataloader=DataLoader(data_sample, sampler=train_sampler, batch_size=1)

for batch in train_dataloader:

	inputs = {'input_ids': batch[0],

	start_positions=batch[3]

	end_positions=batch[4]

	outputs=model(**inputs)

	print("Outputs: ", outputs)