자주 쓰는 모듈/라이브러리

import pandas as pd
#경고 출력 옵션 {"raise": 오류발생, "warn":경고출력(default), "None":무시}
pd.set_option('mode.chained_assignment', None)

import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline
import seaborn as sns
sns.set_style('darkgrid')

import warnings
warnings.filterwarnings('ignore')

import sys
from glob import glob	
import os
import platform
import random
import sklearn 
import cv2
from PIL import Image


import torch
from torchvision import models
import torchvision.transforms as transforms
from torch.utils.data import Dataset, DataLoader
from torchvision.models.efficientnet import efficientnet_b0, efficientnet_b1
import torch.nn.functional as F
from sklearn.metrics import f1_score 
from tqdm import tqdm

# 시드 고정
def seed_everything(seed: int = 2022):
    random.seed(seed)
    np.random.seed(seed)
    os.environ["PYTHONHASHSEED"] = str(seed)
    torch.manual_seed(seed)
    torch.cuda.manual_seed(seed)  
    torch.backends.cudnn.deterministic = True  
    torch.backends.cudnn.benchmark = True 

# 환경 및 버전 확인
print(f"Setup complete. Using torch {torch.__version__} ({torch.cuda.get_device_properties(0).name if torch.cuda.is_available() else 'CPU'})")
print(f"- os: {platform.platform()}")
print(f"- python: {sys.version}")
print(f"- pandas: {pd.__version__}")
print(f"- numpy: {np.__version__}")
print(f"- sklearn: {sklearn.__version__}")
print(f"- pytorch: {torch.__version__}")
print(f"- opencv: {cv2.__version__}")

# hyper parameter
class config:
    seed = 2022
    device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu") 
    lr = 3e-3
    scheduler = ''	# 가변 lr  optim.lr_scheduler.StepLR(optimizer, step_size=5, gamma=0.9)
    epochs = 100
    batch_size = 16
    # num_workers = 4
    patience = 5
    min_delta = 1e-3
    dropout= 0.5
    optimizer = 'Adam'
    train_5_folds = True
    tta = True

seed_everything(config.seed)

acc_check

def acc_check(net, test_set, epoch, save=1):
    correct = 0
    total = 0
    with torch.no_grad():
        for data in test_set:
            images, labels = data
            images = images.to(device)
            labels = labels.to(device)
            outputs = net(images)

            _, predicted = torch.max(outputs.data, 1)

            total += labels.size(0)
            correct += (predicted == labels).sum().item()
    
    acc = (100 * correct / total)
    print('Accuracy of the network on the 10000 test images: %d %%' % acc)
    if save:
        torch.save(net.state_dict(), "./model/model_epoch_{}_acc_{}.pth".format(epoch, int(acc)))
    return acc

모델 저장 / 불러오기

• pytorch tutorial(kr) - 모델 저장하기/불러오기

# save trained model
def save_model(model, saved_dir, file_name='best_model.pt'):
    os.makedirs(saved_dir, exist_ok=True)
    check_point = {
        'net': model.state_dict()
    }
    output_path = os.path.join(saved_dir, file_name)
    torch.save(model.state_dict(), output_path)

# load model
model_path = './saved/CNN/best_model.pt'
model = CNN().to(device)
checkpoint = torch.load(model_path, map_location=device)
model.load_state_dict(checkpoint)

GPU 캐시 비우기

import gc
gc.collect()
torch.cuda.empty_cache()

경로 확인 및 이동

os.getcwd()		# 현재 경로 확인
os.chdir(r"C:\User")		# 경로 이동
f = open("sample folder/sample.txt", "r")	# mode: r(default), w, a