목록분류 전체보기 (295)

UOMOP

Low Attention Selection Performance (CR : 1/6, 1/24, 1/96) (PS : 2)

import matplotlib.pyplot as plt X = [0, 15, 30] Y_0 = [15.752, 18.992, 19.364] Y_25 = [15.761, 18.909, 19.272] Y_50 = [15.718, 18.706, 19.015] Y_75 = [15.572, 18.021, 18.232] Y_1 = [12.095, 12.097, 12.099] plt.plot(X, Y_0, 'o-', color='blue', label='MR = 0%') plt.plot(X, Y_25, 's--', color='red', label='MR = 25%') plt.plot(X, Y_50, '^-', color='green', label='MR = 50%') plt.plot(X, Y_75, 'x:', c..
Main 2024. 3. 15. 18:36
Random Selection

import matplotlib.pyplot as plt import torchvision.transforms as transforms import torch.nn.functional as F import math import torch import torchvision import torch.nn as nn import torch.optim as optim import torch.nn.functional as f from torch.utils.data import DataLoader, Dataset import time from params import * import os from tqdm import tqdm device = torch.device("cuda:0" if torch.cuda.is_av..
Main 2024. 3. 15. 18:31
High Attention Selection

import matplotlib.pyplot as plt import torchvision.transforms as transforms import torch.nn.functional as F import math import torch import torchvision import torch.nn as nn import torch.optim as optim import torch.nn.functional as f from torch.utils.data import DataLoader, Dataset import time from params import * import os from tqdm import tqdm device = torch.device("cuda:0" if torch.cuda.is_av..
Main 2024. 3. 15. 18:29
Low Attention Selection

import matplotlib.pyplot as plt import torchvision.transforms as transforms import torch.nn.functional as F import math import torch import torchvision import torch.nn as nn import torch.optim as optim import torch.nn.functional as f from torch.utils.data import DataLoader, Dataset import time from params import * import os from tqdm import tqdm device = torch.device("cuda:0" if torch.cuda.is_av..
Main 2024. 3. 15. 18:28
Patch selection with zero padding

import torch import torchvision import torchvision.transforms as transforms import matplotlib.pyplot as plt import seaborn as sns def calculate_correlation(patches): flattened_patches = patches.view(patches.size(0), -1) mean_centered = flattened_patches - flattened_patches.mean(dim=1, keepdim=True) norm = mean_centered.norm(dim=1, keepdim=True) norm[norm == 0] = 1 correlation_matrix = torch.mm(m..
Main 2024. 3. 13. 15:25
cifar10 patch correlation map

import torch import torchvision import torchvision.transforms as transforms import torch import matplotlib.pyplot as plt import seaborn as sns import torch import torchvision import torchvision.transforms as transforms import matplotlib.pyplot as plt import seaborn as sns def calculate_correlation(patches): flattened_patches = patches.view(patches.size(0), -1) mean_centered = flattened_patches -..
Main 2024. 3. 11. 14:47
Prev 1 ··· 12 13 14 15 16 17 18 ··· 50 Next