DeepfakeGenome_Codebase / training /detectors /effort_patch_shuffle.py
shunliwang
update
8bc3305
import os
import math
import datetime
import logging
import numpy as np
from sklearn import metrics
from typing import Union
from collections import defaultdict
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torch.nn import DataParallel
from torch.utils.tensorboard import SummaryWriter
from metrics.base_metrics_class import calculate_metrics_for_train, calculate_acc_for_train
from .base_detector import AbstractDetector
from detectors import DETECTOR
from networks import BACKBONE
from loss import LOSSFUNC
import albumentations as A
import loralib as lora
from transformers import AutoProcessor, CLIPModel, ViTModel, ViTConfig
logger = logging.getLogger(__name__)
def get_clip_visual(model_name = "openai/clip-vit-base-patch16"):
processor = AutoProcessor.from_pretrained(model_name)
model = CLIPModel.from_pretrained(model_name)
return processor, model.vision_model
def shuffle_patches(images: torch.Tensor, patch_size: int = 14) -> torch.Tensor:
"""
Apply patch-level shuffling to the input images.
images: [B, C, H, W]
patch_size: patch size used by ViT (for example, 16)
Returns: an image tensor with the same shape [B, C, H, W]
"""
B, C, H, W = images.shape
assert H % patch_size == 0 and W % patch_size == 0, \
f"H ({H}) and W ({W}) must be divisible by patch_size ({patch_size})"
num_patches_h = H // patch_size
num_patches_w = W // patch_size
num_patches = num_patches_h * num_patches_w
# [B, C, H, W] -> [B, C, num_patches_h, patch_size, num_patches_w, patch_size]
images = images.view(B, C, num_patches_h, patch_size, num_patches_w, patch_size)
# -> [B, num_patches_h, num_patches_w, C, patch_size, patch_size]
images = images.permute(0, 2, 4, 1, 3, 5).contiguous()
# -> [B, num_patches, C, patch_size, patch_size]
images = images.view(B, num_patches, C, patch_size, patch_size)
# Shuffle patch order independently for each image.
# permutation shape: [B, num_patches]
perms = torch.stack(
[torch.randperm(num_patches, device=images.device) for _ in range(B)],
dim=0
)
# Use advanced indexing to perform the shuffle.
batch_idx = torch.arange(B, device=images.device).unsqueeze(1).expand(B, num_patches)
images = images[batch_idx, perms] # [B, num_patches, C, patch_size, patch_size]
# Restore the original image shape.
images = images.view(B, num_patches_h, num_patches_w, C, patch_size, patch_size)
# -> [B, C, num_patches_h, patch_size, num_patches_w, patch_size]
images = images.permute(0, 3, 1, 4, 2, 5).contiguous()
# -> [B, C, H, W]
images = images.view(B, C, H, W)
return images
def get_aug_transform():
return A.Compose([
A.HorizontalFlip(p=0.5),
A.RandomBrightnessContrast(brightness_limit=0.2, contrast_limit=0.2, p=0.5),
A.HueSaturationValue(p=0.3),
A.ImageCompression(quality_lower=40, quality_upper=100, p=0.1),
A.GaussNoise(p=0.1),
A.MotionBlur(p=0.1),
A.CLAHE(p=0.1),
A.ChannelShuffle(p=0.1),
A.Cutout(p=0.1),
A.RandomGamma(p=0.3),
A.GlassBlur(p=0.3),
])
def data_aug(images: torch.Tensor) -> torch.Tensor:
is_gpu = images.is_cuda
aug = get_aug_transform()
# Step 1: convert the batch tensor to batch numpy arrays (BHWC uint8 0-255).
imgs_np = images.cpu().detach().numpy()
imgs_np = np.transpose(imgs_np, (0, 2, 3, 1)) # BCHW -> BHWC
imgs_np = (imgs_np * 255).astype(np.uint8)
# Step 2: augment images one by one to avoid KeyError from batch-style arguments.
imgs_aug_np = []
for img in imgs_np:
# Pass a single image with `image=img`, which is natively supported by Albumentations.
aug_img = aug(image=img)["image"]
imgs_aug_np.append(aug_img)
imgs_aug_np = np.array(imgs_aug_np) # convert back to batch numpy arrays
# Step 3: convert back to a tensor while preserving the original logic.
aug_tensor = torch.from_numpy(imgs_aug_np).permute(0, 3, 1, 2)
aug_tensor = aug_tensor.float() / 255.0
# Restore the original device.
if is_gpu:
aug_tensor = aug_tensor.cuda()
return aug_tensor
@DETECTOR.register_module(module_name='effort_shuffle_ensemble')
class Effort_Shuffle_Ensenble_Detector(nn.Module):
def __init__(self, config=None):
super().__init__()
self.config = config
self.backbone = self.build_backbone(config)
self.head = nn.Linear(1024, config['backbone_config']['num_classes'])
#self.head1 = nn.Linear(1024, config['backbone_config']['num_classes'])
self.loss_func = nn.CrossEntropyLoss()
self.prob, self.label = [], []
self.correct, self.total = 0, 0
#self.backbone2=self.build_clip_backbone(config)
def build_clip_backbone(self,config):
_, backbone = get_clip_visual(model_name=config['pretrained'])
return backbone
def build_backbone(self, config):
# Download model
# https://huggingface.co/openai/clip-vit-large-patch14
# mean: [0.48145466, 0.4578275, 0.40821073]
# std: [0.26862954, 0.26130258, 0.27577711]
# ViT-L/14 224*224
clip_model = CLIPModel.from_pretrained(self.config["pretrained"])
# Apply SVD to self_attn layers only
# ViT-L/14 224*224: 1024-1
clip_model.vision_model = apply_svd_residual_to_self_attn(clip_model.vision_model, r=1024-1)
for name, param in clip_model.vision_model.named_parameters():
print('{}: {}'.format(name, param.requires_grad))
num_param = sum(p.numel() for p in clip_model.vision_model.parameters() if p.requires_grad)
num_total_param = sum(p.numel() for p in clip_model.vision_model.parameters())
print('Number of total parameters: {}, tunable parameters: {}'.format(num_total_param, num_param))
return clip_model.vision_model
def features(self, data_dict: dict) -> torch.tensor:
# data_dict['image']: torch.Size([32, 3, 224, 224])
if self.training:
#aug_image=data_aug(data_dict['image'])
shuffle_images=shuffle_patches(data_dict['image'],14)
feat = self.backbone(shuffle_images)['pooler_output']
#feat1=self.backbone2(shuffle_images)['pooler_output']
else:
feat = self.backbone(data_dict['image'])['pooler_output']
#feat1=self.backbone2(data_dict['image'])['pooler_output']
# feat torch.Size([32, 1024])
return feat#,feat1
def classifier(self, features: torch.tensor) -> torch.tensor:
return self.head(features)
# def get_losses(self, data_dict: dict, pred_dict: dict) -> dict:
# label = data_dict['label']
# pred = pred_dict['cls']
# loss = self.loss_func(pred, label)
# # Regularization term
# lambda_reg = 0.1
# orthogonal_losses = []
# for module in self.backbone.modules():
# if isinstance(module, SVDResidualLinear):
# # Apply orthogonal constraints to the U_residual and V_residual matrix
# orthogonal_losses.append(module.compute_orthogonal_loss())
# if orthogonal_losses:
# reg_term = sum(orthogonal_losses)
# loss += lambda_reg * reg_term
# loss_dict = {'overall': loss}
# return loss_dict
def compute_weight_loss(self):
weight_sum_dict = {}
num_weight_dict = {}
for name, module in self.backbone.named_modules():
if isinstance(module, SVDResidualLinear):
weight_curr = module.compute_current_weight()
if str(weight_curr.size()) not in weight_sum_dict.keys():
weight_sum_dict[str(weight_curr.size())] = weight_curr
num_weight_dict[str(weight_curr.size())] = 1
else:
weight_sum_dict[str(weight_curr.size())] += weight_curr
num_weight_dict[str(weight_curr.size())] += 1
loss2 = 0.0
for k in weight_sum_dict.keys():
_, S_sum, _ = torch.linalg.svd(weight_sum_dict[k], full_matrices=False)
loss2 += -torch.mean(S_sum)
loss2 /= len(weight_sum_dict.keys())
return loss2
def get_losses(self, data_dict: dict, pred_dict: dict) -> dict:
label = data_dict['label'] # Tensor of shape [batch_size]
pred = pred_dict['cls'] # Tensor of shape [batch_size, num_classes]
# Compute overall loss using all samples
loss = self.loss_func(pred, label)
# Create masks for real and fake classes
mask_real = label == 0 # Boolean tensor
mask_fake = label == 1 # Boolean tensor
# Compute loss for real class
if mask_real.sum() > 0:
pred_real = pred[mask_real]
label_real = label[mask_real]
loss_real = self.loss_func(pred_real, label_real)
else:
# No real samples in batch
loss_real = torch.tensor(0.0, device=pred.device)
# Compute loss for fake class
if mask_fake.sum() > 0:
pred_fake = pred[mask_fake]
label_fake = label[mask_fake]
loss_fake = self.loss_func(pred_fake, label_fake)
else:
# No fake samples in batch
loss_fake = torch.tensor(0.0, device=pred.device)
# loss2 = self.compute_weight_loss()
# overall_loss = loss + loss2
# Return a dictionary with all losses
loss_dict = {
'overall': loss,
'real_loss': loss_real,
'fake_loss': loss_fake,
# 'erank_loss': loss2
}
return loss_dict
def get_train_metrics(self, data_dict: dict, pred_dict: dict) -> dict:
label = data_dict['label']
pred = pred_dict['cls']
# compute metrics for batch data
# auc, eer, acc, ap = calculate_metrics_for_train(label.detach(), pred.detach())
# metric_batch_dict = {'acc': acc, 'auc': auc, 'eer': eer, 'ap': ap}
acc, mAP = calculate_acc_for_train(label.detach(), pred.detach(), self.config['backbone_config']['num_classes'])
metric_batch_dict = {'acc': acc, 'mAP': mAP}
return metric_batch_dict
def forward(self, data_dict: dict, inference=False) -> dict:
# get the features by backbone
features= self.features(data_dict)
# get the prediction by classifier
pred = self.classifier(features)
#features=features+f1
#pred=pred+pred1
# get the probability of the pred
# prob = torch.softmax(pred, dim=1)[:, 1]
prob = torch.softmax(pred, dim=1)
# build the prediction dict for each output
pred_dict = {'cls': pred, 'prob': prob, 'feat': features}
return pred_dict
# Custom module to represent the residual using SVD components
class SVDResidualLinear(nn.Module):
def __init__(self, in_features, out_features, r, bias=True, init_weight=None):
super(SVDResidualLinear, self).__init__()
self.in_features = in_features
self.out_features = out_features
self.r = r # Number of top singular values to exclude
# Original weights (fixed)
self.weight_main = nn.Parameter(torch.Tensor(out_features, in_features), requires_grad=False)
if init_weight is not None:
self.weight_main.data.copy_(init_weight)
else:
nn.init.kaiming_uniform_(self.weight_main, a=math.sqrt(5))
# Bias
if bias:
self.bias = nn.Parameter(torch.Tensor(out_features))
nn.init.zeros_(self.bias)
else:
self.register_parameter('bias', None)
def compute_current_weight(self):
if self.S_residual is not None:
return self.weight_main + self.U_residual @ torch.diag(self.S_residual) @ self.V_residual
else:
return self.weight_main
def forward(self, x):
if hasattr(self, 'U_residual') and hasattr(self, 'V_residual') and self.S_residual is not None:
# Reconstruct the residual weight
residual_weight = self.U_residual @ torch.diag(self.S_residual) @ self.V_residual
# Total weight is the fixed main weight plus the residual
weight = self.weight_main + residual_weight
else:
# If residual components are not set, use only the main weight
weight = self.weight_main
return F.linear(x, weight, self.bias)
def compute_orthogonal_loss(self):
if self.S_residual is not None:
# According to the properties of orthogonal matrices: A^TA = I
UUT = torch.cat((self.U_r, self.U_residual), dim=1) @ torch.cat((self.U_r, self.U_residual), dim=1).t()
VVT = torch.cat((self.V_r, self.V_residual), dim=0) @ torch.cat((self.V_r, self.V_residual), dim=0).t()
# print(self.U_r.size(), self.U_residual.size()) # torch.Size([1024, 1023]) torch.Size([1024, 1])
# print(self.V_r.size(), self.V_residual.size()) # torch.Size([1023, 1024]) torch.Size([1, 1024])
# UUT = self.U_residual @ self.U_residual.t()
# VVT = self.V_residual @ self.V_residual.t()
# Construct an identity matrix
UUT_identity = torch.eye(UUT.size(0), device=UUT.device)
VVT_identity = torch.eye(VVT.size(0), device=VVT.device)
# Using frobenius norm to compute loss
loss = 0.5 * torch.norm(UUT - UUT_identity, p='fro') + 0.5 * torch.norm(VVT - VVT_identity, p='fro')
else:
loss = 0.0
return loss
def compute_keepsv_loss(self):
if (self.S_residual is not None) and (self.weight_original_fnorm is not None):
# Total current weight is the fixed main weight plus the residual
weight_current = self.weight_main + self.U_residual @ torch.diag(self.S_residual) @ self.V_residual
# Frobenius norm of current weight
weight_current_fnorm = torch.norm(weight_current, p='fro')
loss = torch.abs(weight_current_fnorm ** 2 - self.weight_original_fnorm ** 2)
# loss = torch.abs(weight_current_fnorm ** 2 + 0.01 * self.weight_main_fnorm ** 2 - 1.01 * self.weight_original_fnorm ** 2)
else:
loss = 0.0
return loss
def compute_fn_loss(self):
if (self.S_residual is not None):
weight_current = self.weight_main + self.U_residual @ torch.diag(self.S_residual) @ self.V_residual
weight_current_fnorm = torch.norm(weight_current, p='fro')
loss = weight_current_fnorm ** 2
else:
loss = 0.0
return loss
# Function to replace nn.Linear modules within self_attn modules with SVDResidualLinear
def apply_svd_residual_to_self_attn(model, r):
for name, module in model.named_children():
if 'self_attn' in name:
# Replace nn.Linear layers in this module
for sub_name, sub_module in module.named_modules():
if isinstance(sub_module, nn.Linear):
# Get parent module within self_attn
parent_module = module
sub_module_names = sub_name.split('.')
for module_name in sub_module_names[:-1]:
parent_module = getattr(parent_module, module_name)
# Replace the nn.Linear layer with SVDResidualLinear
setattr(parent_module, sub_module_names[-1], replace_with_svd_residual(sub_module, r))
else:
# Recursively apply to child modules
apply_svd_residual_to_self_attn(module, r)
# After replacing, set requires_grad for residual components
for param_name, param in model.named_parameters():
if any(x in param_name for x in ['S_residual', 'U_residual', 'V_residual']):
param.requires_grad = True
else:
param.requires_grad = False
return model
# Function to replace a module with SVDResidualLinear
def replace_with_svd_residual(module, r):
if isinstance(module, nn.Linear):
in_features = module.in_features
out_features = module.out_features
bias = module.bias is not None
# Create SVDResidualLinear module
new_module = SVDResidualLinear(in_features, out_features, r, bias=bias, init_weight=module.weight.data.clone())
if bias and module.bias is not None:
new_module.bias.data.copy_(module.bias.data)
new_module.weight_original_fnorm = torch.norm(module.weight.data, p='fro')
# Perform SVD on the original weight
U, S, Vh = torch.linalg.svd(module.weight.data, full_matrices=False)
# Determine r based on the rank of the weight matrix
r = min(r, len(S)) # Ensure r does not exceed the number of singular values
# Keep top r singular components (main weight)
U_r = U[:, :r] # Shape: (out_features, r)
S_r = S[:r] # Shape: (r,)
Vh_r = Vh[:r, :] # Shape: (r, in_features)
# Reconstruct the main weight (fixed)
weight_main = U_r @ torch.diag(S_r) @ Vh_r
# Calculate the frobenius norm of main weight
new_module.weight_main_fnorm = torch.norm(weight_main.data, p='fro')
# Set the main weight
new_module.weight_main.data.copy_(weight_main)
# Residual components (trainable)
U_residual = U[:, r:] # Shape: (out_features, n - r)
S_residual = S[r:] # Shape: (n - r,)
Vh_residual = Vh[r:, :] # Shape: (n - r, in_features)
if len(S_residual) > 0:
new_module.S_residual = nn.Parameter(S_residual.clone())
new_module.U_residual = nn.Parameter(U_residual.clone())
new_module.V_residual = nn.Parameter(Vh_residual.clone())
new_module.S_r = nn.Parameter(S_r.clone(), requires_grad=False)
new_module.U_r = nn.Parameter(U_r.clone(), requires_grad=False)
new_module.V_r = nn.Parameter(Vh_r.clone(), requires_grad=False)
else:
new_module.S_residual = None
new_module.U_residual = None
new_module.V_residual = None
new_module.S_r = None
new_module.U_r = None
new_module.V_r = None
return new_module
else:
return module