PathwayGrad/pruning_methods_visual.py at main · CAMP-eXplain-AI/PathwayGrad · GitHub

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
import matplotlib as mpl
import numpy as np
import torch
from torchvision import datasets, models, transforms
from tqdm import tqdm
from src import Pruner, Plot_tools
from src.attribution_methods import vanilla_saliency
import matplotlib.pyplot as plt

# ### Setup Imagenet
#
# ImageNet as of Oct2019 can no longer be downloaded using pytorch.
# https://github.com/pytorch/vision/issues/1453
# To download ImageNet, see http://image-net.org/.

imagenet_dir = '/home/ashkan/data/ILSVRC2012/'
transform = transforms.Compose([transforms.Resize(256),
                                transforms.CenterCrop(224),
                                transforms.ToTensor(),
                                transforms.transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])])
imagenet = datasets.ImageNet(imagenet_dir, download=False, split='val', transform=transform)
classes = imagenet.classes
mpl.rcParams['figure.dpi']= 400


# ### Method to get attributions

def get_attribution(attribution_name, data, model, model_sparsity_threshold):
    make_single_channel = True
    class_id = model(data)
    class_id = class_id.data.max(1)[1].item()
    # print(class_id)
    model.eval()
    if attribution_name == "Gradients":
        vanilla_sal = vanilla_saliency.VanillaSaliency(model, device)
        saliency = vanilla_sal.generate_saliency(data, class_id, make_single_channel)

    elif attribution_name == "NeuronMCT":
        pruner = Pruner.Pruner(model, data, device)
        pruner.prune_neuron_mct(model_sparsity_threshold, debug=False)
        saliency = pruner.generate_saliency(make_single_channel=make_single_channel)
        pruner.remove_handles()

    elif attribution_name == "NeuronIntGrad":
        pruner = Pruner.Pruner(model, data, device)
        pruner.prune_integrad(model_sparsity_threshold, debug=False)
        saliency = pruner.generate_saliency(make_single_channel=make_single_channel)
        pruner.remove_handles()

    if make_single_channel:
        saliency = torch.from_numpy(np.asarray(saliency)).view([1, 224, 224])
    else:
        saliency = torch.from_numpy(np.asarray(saliency)).view([3, 224, 224])
    saliency /= np.max(np.asarray(abs(saliency)).flatten())
    return saliency


def visualize(model, dataloader, images, classes, sparsity_levels):
    global id
    model = model.to(device)
    model.eval()
    name_methods = ["NeuronIntGrad", "NeuronMCT"]
    num_methods = len(name_methods)
    dataiter = iter(dataloader)
    i = 0
    fig = plt.figure(figsize=(7, num_samples*2))
    acts = []
    for chosen in tqdm(range(num_samples)):
        data, _ = dataiter.next()
        data = data.to(device)
        output = model(data.clone())
        output = torch.nn.functional.softmax(output.detach(), dim=1)
        predicted_logit = output.data.max(1)[1].item()
        predicted_prob = output.data.max(1)[0].item()
        image = Plot_tools.reverse_preprocess_imagenet_image(data.clone())
        ax = fig.add_subplot(num_samples*num_methods, len(sparsity_levels)+1, 2*chosen*(len(sparsity_levels)+1) + 1)
        if chosen == 0:
            ax.set_title("Original Image", fontsize=6)
        ax.text(-0.13, 0.5, classes[predicted_logit][0]+"\n"+str("%.2f" % round(predicted_prob*100, 2)+"%"), fontsize=6, rotation=90, horizontalalignment='center', verticalalignment='center', transform=ax.transAxes)
        plt.axis('off')
        plt.imshow(image)
        for j in range(num_methods):
            for k in range(len(sparsity_levels)):
              ax = fig.add_subplot(num_samples*num_methods, len(sparsity_levels)+1, num_methods*chosen*(len(sparsity_levels)+1) + j*(len(sparsity_levels)+1) + 1 + k + 1)
              if sparsity_levels[k] == 0:
                  attribution = get_attribution('Gradients', data.clone(), model, sparsity_levels[k])
              else:
                  attribution = get_attribution(name_methods[j], data.clone(), model, sparsity_levels[k])
              attribution = np.asarray(attribution.squeeze(0))
              if chosen == 0 and j == 0:
                if sparsity_levels[k] != 0:
                  ax.set_title("Sparsity={}".format(sparsity_levels[k]), fontsize=6)
                else:
                  ax.set_title("Original Network", fontsize=6)
              plt.imshow(abs(attribution), cmap='jet', vmin=0, vmax=1)
              if k == len(sparsity_levels)-1:
                  ax.text(1.05, 0.5, name_methods[j], fontsize=5, rotation=-90, horizontalalignment='center', verticalalignment='center', transform=ax.transAxes, fontweight='bold')
              plt.axis('off')
        i += 1
    plt.tight_layout()
    plt.subplots_adjust(wspace=0.05, hspace=0.01)
    plt.savefig('./different_sparsity_'+str(id)+'.png', dpi=300)
    id += 1


device = 'cuda:0' if torch.cuda.is_available() else 'cpu'

num_samples = 1
id = 0
indices = [16305] #butterfly
dataset = torch.utils.data.Subset(imagenet, indices)
dataloader = torch.utils.data.DataLoader(dataset, batch_size=1, shuffle=True)


# # Evaluate on VGG16

dataset = torch.utils.data.Subset(imagenet, indices)
dataloader = torch.utils.data.DataLoader(dataset, batch_size=1, shuffle=True)

model = models.vgg16(pretrained=True)
model_sparsity_threshold = 90 # Threshold computed for 15% output change for Pruner
visualize(model, dataloader, indices, classes, [0, 70, 80, 85, 90, 95, 99])