인프런 커뮤니티 질문&답변

첨부해주신 코드를 복사 붙여넣기 했는데, 오류가 나네요.. ㅠㅠ

오류명은 :

---> 21 class CustomDataset(Dataset):

22 def init(self, file_list, transform=None):

23 self.file_list = file_list

TypeError: str() argument 'encoding' must be str, not tuple

dataset_train = CustomDataset(train_list, transform=transforms_for_train)
dataset_valid = CustomDataset(val_list, transform=transforms_for_val_test)
dataset_test = CustomDataset(test_list, transform=transforms_for_val_test)

CustomDataset 클래스를 통해, train, valid, test를 만들어주는 과정에서, str() arguments를 받아야 한다는 것 같은데.. 왜 오류가 나는지 잘 모르겠습니다.

from torchvision import transforms

input_size = 224
transforms_for_train =  transforms.Compose([
    transforms.RandomResizedCrop(input_size, scale=(0.5, 1.0)),
    transforms.RandomHorizontalFlip(),
    transforms.RandomVerticalFlip(),
    transforms.RandomRotation(10),
    transforms.ToTensor(),
    transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])

transforms_for_val_test = transforms.Compose([
    transforms.Resize(input_size),
    transforms.CenterCrop(input_size),
    transforms.ToTensor(),
    transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])

#class Dataset
class CustomDataset(Dataset):
    def __init__(self, file_list, transform=None):
        self.file_list = file_list
        self.transform = transform

    def __len__(self):
        return len(self.file_list)

    def __getitem__(self, idx):
        img_path = self.file_list[idx]
        if img_path.split('/')[-1][-3:] == 'jpg':
            img = Image.open(img_path)
            if self.transform is not None:
                img_transform = self.transform(img)
                label = img_path.split('/')[-1].split('.')[0]
                if label == 'dog':
                    label = 1
                elif label == 'cat':
                    label = 0
        return img_transform, label

dataset_train = CustomDataset(train_list, transform=transforms_for_train)
dataset_valid = CustomDataset(val_list, transform=transforms_for_val_test)
dataset_test = CustomDataset(test_list, transform=transforms_for_val_test)

# from torch.utils.data import DataLoader # 데이터 로더 클래스
#
# train_batches = DataLoader(dataset=dataset_train, batch_size=8, shuffle=True)
# val_batches = DataLoader(dataset=dataset_valid, batch_size=8, shuffle=False)
# test_batches = DataLoader(dataset=dataset_test, batch_size=8, shuffle=False)

강의 내용과 같이 len(train_list), len(val_list)가 잘 출력이 됨을 볼 수 있고, Reference 항목에서 특정 img의 사이즈를 출력해볼 떄도, (500, 374) 라고 잘 출력됨을 알 수 있었습니다.

train할 데이터와 test 데이터는 강의 중에 배운 os 명령어 및 with zip 명령어를 사용해서

<오류 화면 입니다.>

전체 작성한 코드 첨부드립니다.

import zipfile

Dataset = 'dogs-vs-cats-redux-kernels-edition'

with zipfile.ZipFile('../BecomeProfessional/' + Dataset + ".zip", "r") as z:
    z.extractall(".")

import pandas as pd
import numpy as np
import warnings
warnings.filterwarnings('ignore')
from PIL import Image
import torch
import torch.nn as nn
from torch.utils.data import DataLoader, Subset
from torchvision import datasets, transforms
from sklearn.model_selection import train_test_split
from copy import deepcopy
import matplotlib.pyplot as plt
import os, shutil

import random
import os

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

seed = 42 # seed값 설정
random.seed(seed) # 파이썬 난수 생성기
os.environ['PYTHONHASHSEED'] = str(seed) # 해시 시크릿값 고정
np.random.seed(seed) # 넘파이 난수 생성기

torch.manual_seed(seed) # 파이토치 CPU 난수 생성기
torch.backends.cudnn.deterministic = True # 확정적 연산 사용 설정
torch.backends.cudnn.benchmark = False # 벤치마크 기능 사용 해제
torch.backends.cudnn.enabled = False # cudnn 기능 사용 해제

if device == 'cuda':
    torch.cuda.manual_seed(seed)
    torch.cuda.manual_seed_all(seed)

original_data_path = './Dog_Cat/'
os.makedirs('dogsvscats', exist_ok=True)
import zipfile
with zipfile.ZipFile(os.path.join(original_data_path, 'train.zip')) as train_zip:
    train_zip.extractall('./Dog_Cat/dogsvscats')
with zipfile.ZipFile(os.path.join(original_data_path, 'test.zip')) as test_zip:
    test_zip.extractall('./Dog_Cat/dogsvscats')

import glob
print(glob.glob('./Dog_Cat/dogsvscats/train/*'))

출력: ./Dog_Cat/dogsvscats/train\\cat.11346.jpg', './Dog_Cat/dogsvscats/train\\cat.11347.jpg', './Dog_Cat/dogsvscats/train\\cat.11348.jpg', './Dog_Cat/dogsvscats/train\\cat.11349.jpg', ' .... (엄청 많이 출력됩니다)

train_dir = './Dog_Cat/dogsvscats/train'
test_dir = './Dog_Cat/dogsvscats/test'
all_train_files = glob.glob(os.path.join(train_dir, '*.jpg'))
test_list = glob.glob(os.path.join(test_dir, '*.jpg'))
train_labels = [path.split('/')[-1].split('.')[0] for path in all_train_files]
train_list, val_list = train_test_split(all_train_files, test_size=0.1, stratify=train_labels, random_state=seed)
print(len(train_list), len(val_list))

출력: 22500 2500

train_list[0]

출력: './Dog_Cat/dogsvscats/train\\cat.4814.jpg'

img = Image.open('./Dog_Cat/dogsvscats/train/cat.4814.jpg')
img.size

출력: (500, 374)

import IPython
image1 = IPython.display.Image(filename='./Dog_Cat/dogsvscats/train/cat.4814.jpg')
display(image1)

from torchvision import transforms

input_size = 224
transforms_for_train =  transforms.Compose([
    transforms.RandomResizedCrop(input_size, scale=(0.5, 1.0)),
    transforms.RandomHorizontalFlip(),
    transforms.RandomVerticalFlip(),
    transforms.RandomRotation(10),
    transforms.ToTensor(),
    transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])

transforms_for_val_test = transforms.Compose([
    transforms.Resize(input_size),
    transforms.CenterCrop(input_size),
    transforms.ToTensor(),
    transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])

#class Dataset
class CustomDataset(Dataset):
    def __init__(self, file_list, transform=None):
        self.file_list = file_list
        self.transform = transform

    def __len__(self):
        return len(self.file_list)

    def __getitem__(self, idx):
        img_path = self.file_list[idx]
        if img_path.split('/')[-1][-3:] == 'jpg':
            img = Image.open(img_path)
            if self.transform is not None:
                img_transform = self.transform(img)
                label = img_path.split('/')[-1].split('.')[0]
                if label == 'dog':
                    label = 1
                elif label == 'cat':
                    label = 0
        return img_transform, label

dataset_train = CustomDataset(train_list, transform=transforms_for_train)
dataset_valid = CustomDataset(val_list, transform=transforms_for_val_test)
dataset_test = CustomDataset(test_list, transform=transforms_for_val_test)

오류 발생