File size: 5,265 Bytes
3f2b233 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 |
import shap
import torch
import torch.nn as nn
import torch.nn.functional as F
from torchvision import models
import numpy as np
import matplotlib.pyplot as plt
# Custom BasicBlock to avoid in-place operations
class CustomBasicBlock(nn.Module):
expansion = 1
def __init__(self, in_planes, planes, stride=1, downsample=None, groups=1, base_width=64, dilation=1, norm_layer=None):
super(CustomBasicBlock, self).__init__()
if norm_layer is None:
norm_layer = nn.BatchNorm2d
self.conv1 = nn.Conv2d(in_planes, planes, kernel_size=3, stride=stride, padding=1, bias=False)
self.bn1 = norm_layer(planes)
self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, padding=1, bias=False)
self.bn2 = norm_layer(planes)
self.downsample = downsample
self.stride = stride
def forward(self, x):
identity = x.clone()
out = self.conv1(x)
out = self.bn1(out)
out = F.relu(out.clone(), inplace=False)
out = self.conv2(out)
out = self.bn2(out)
if self.downsample is not None:
identity = self.downsample(x.clone())
out = out.clone() + identity # Clone before addition to avoid in-place modification
out = F.relu(out.clone(), inplace=False)
return out
# Custom ResNet using CustomBasicBlock
class CustomResNet(nn.Module):
def __init__(self, block, layers, num_classes=1000):
super(CustomResNet, self).__init__()
self.inplanes = 64
self.conv1 = nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3, bias=False)
self.bn1 = nn.BatchNorm2d(64)
self.relu = nn.ReLU(inplace=False)
self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
self.layer1 = self._make_layer(block, 64, layers[0])
self.layer2 = self._make_layer(block, 128, layers[1], stride=2)
self.layer3 = self._make_layer(block, 256, layers[2], stride=2)
self.layer4 = self._make_layer(block, 512, layers[3], stride=2)
self.avgpool = nn.AdaptiveAvgPool2d((1, 1))
self.fc = nn.Linear(512 * block.expansion, num_classes)
for m in self.modules():
if isinstance(m, nn.Conv2d):
nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')
elif isinstance(m, nn.BatchNorm2d):
nn.init.constant_(m.weight, 1)
nn.init.constant_(m.bias, 0)
def _make_layer(self, block, planes, blocks, stride=1):
norm_layer = nn.BatchNorm2d
downsample = None
if stride != 1 or self.inplanes != planes * block.expansion:
downsample = nn.Sequential(
nn.Conv2d(self.inplanes, planes * block.expansion, kernel_size=1, stride=stride, bias=False),
norm_layer(planes * block.expansion),
)
layers = []
layers.append(block(self.inplanes, planes, stride, downsample, groups=1, base_width=64, dilation=1, norm_layer=norm_layer))
self.inplanes = planes * block.expansion
for _ in range(1, blocks):
layers.append(block(self.inplanes, planes, groups=1, base_width=64, dilation=1, norm_layer=norm_layer))
return nn.Sequential(*layers)
def forward(self, x):
x = self.conv1(x)
x = self.bn1(x)
x = self.relu(x.clone()) # Clone to avoid in-place operation
x = self.maxpool(x)
x = self.layer1(x.clone()) # Clone to avoid in-place operation
x = self.layer2(x.clone()) # Clone to avoid in-place operation
x = self.layer3(x.clone()) # Clone to avoid in-place operation
x = self.layer4(x.clone()) # Clone to avoid in-place operation
x = self.avgpool(x)
x = torch.flatten(x, 1)
x = self.fc(x.clone()) # Clone to avoid in-place operation
return x
# Initialize the custom model with pre-trained weights
model = CustomResNet(CustomBasicBlock, [2, 2, 2, 2])
state_dict = models.resnet18(weights=models.ResNet18_Weights.IMAGENET1K_V1).state_dict()
model.load_state_dict(state_dict)
model.eval()
# Initialize SHAP explainer with the custom model
explainer = shap.DeepExplainer(model, torch.randn(1, 3, 224, 224))
# Generate SHAP values for an input image
sample_image = torch.randn(1, 3, 224, 224)
shap_values = explainer.shap_values(sample_image, check_additivity=False)
# Convert SHAP values and sample image to numpy for SHAP visualization
shap_values_class_0 = shap_values[0][0] # Extract SHAP values for the first class
sample_image_np = sample_image.squeeze().permute(1, 2, 0).detach().numpy()
# Normalize sample image and SHAP values to range [0, 1] for visualization
sample_image_np = np.clip(sample_image_np, 0, 1)
shap_min, shap_max = shap_values_class_0.min(), shap_values_class_0.max()
shap_values_class_0 = (shap_values_class_0 - shap_min) / (shap_max - shap_min)
# Ensure both `sample_image_np` and `shap_values_class_0` are NumPy arrays with correct shapes for image_plot
sample_image_np = np.array([sample_image_np]) # Add batch dimension for SHAP
shap_values_class_0 = np.array([shap_values_class_0]) # Add batch dimension for SHAP
# Visualize SHAP values for the first class
shap.image_plot(shap_values_class_0, sample_image_np) |