predcoding/script.py at master · nikkiguo/predcoding · 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
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
# pylint: disable=not-callable
# pylint: disable=no-member

import numpy as np
import torch
import os
import mnist_utils
import functions as F
from network import PredictiveCodingNetwork
import argparse


class AttrDict(dict):
    __setattr__ = dict.__setitem__
    __getattr__ = dict.__getitem__


def main(cf):
    print(f"device [{cf.device}]")

    if cf.dataset == "mnist":
        print("loading MNIST data...")
        train_set = mnist_utils.get_mnist_train_set()
        test_set = mnist_utils.get_mnist_test_set()
    elif cf.dataset == "fashion_mnist":
        print("loading Fashion MNIST data...")
        train_set = mnist_utils.get_fashion_mnist_train_set()
        test_set = mnist_utils.get_fashion_mnist_test_set()


    img_train = mnist_utils.get_imgs(train_set)
    img_test = mnist_utils.get_imgs(test_set)
    label_train = mnist_utils.get_labels(train_set)
    label_test = mnist_utils.get_labels(test_set)


    if cf.data_size is not None:
        test_size = cf.data_size // 5
        img_train = img_train[:, 0 : cf.data_size]
        label_train = label_train[:, 0 : cf.data_size]
        img_test = img_test[:, 0:test_size]
        label_test = label_test[:, 0:test_size]

    msg = "img_train {} img_test {} label_train {} label_test {}"
    print(msg.format(img_train.shape, img_test.shape, label_train.shape, label_test.shape))

    print("performing preprocessing...")
    if cf.apply_scaling:
        img_train = mnist_utils.scale_imgs(img_train, cf.img_scale)
        img_test = mnist_utils.scale_imgs(img_test, cf.img_scale)
        label_train = mnist_utils.scale_labels(label_train, cf.label_scale)
        label_test = mnist_utils.scale_labels(label_test, cf.label_scale)

    if cf.apply_inv and cf.act_fn != F.RELU:
        img_train = F.f_inv(img_train, cf.act_fn)
        img_test = F.f_inv(img_test, cf.act_fn)

    model = PredictiveCodingNetwork(cf)

    with torch.no_grad():
        for epoch in range(cf.n_epochs):
            print(f"\nepoch {epoch}")

            img_batches, label_batches = mnist_utils.get_batches(img_train, label_train, cf.batch_size, cf.percent_data_used, cf.subsample_idx)
            print(f"training on {len(img_batches)} batches of size {cf.batch_size}")
            model.train_epoch(img_batches, label_batches, epoch_num=epoch)

            img_batches, label_batches = mnist_utils.get_batches(img_test, label_test, cf.batch_size, cf.percent_data_used, cf.subsample_idx)
            print(f"testing on {len(img_batches)} batches of size {cf.batch_size}")
            accs = model.test_epoch(img_batches, label_batches)
            print(f"average accuracy {np.mean(np.array(accs))}")

            np.random.seed(cf.seed)
            perm = np.random.permutation(img_train.shape[1])
            img_train = img_train[:, perm]
            label_train = label_train[:, perm]

    # Save model state_dict
    filepath = "models/"
    filename = f'net_{cf.dataset}_{cf.act_fn}_{cf.optim}_seed{cf.seed}_samplsize{cf.percent_data_used}_samplidx{cf.subsample_idx}.pth'
    full_path = os.path.join(filepath, filename)
    model.save(full_path)


if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    # Add arguments
    parser.add_argument("--data", type=str)
    parser.add_argument("--o", type=str, help="Optimizer name")
    parser.add_argument("--af", type=str, help="Activation function name")

    # Parse the arguments
    args = parser.parse_args()

    cf = AttrDict()

    cf.n_epochs = 16
    cf.data_size = None
    cf.batch_size = 128
    cf.seed = 20
    cf.percent_data_used = 0.2
    cf.subsample_idx = 0 # e.g. subsample_idx = 0, 1, 2, 3, 4 when percent_data_used = 0.2

    cf.apply_inv = True
    cf.apply_scaling = True
    cf.label_scale = 0.94
    cf.img_scale = 1.0

    if (not args.data) or (args.data == "mnist"): # default is mnist
        cf.dataset = "mnist"
    elif args.data == "fashion_mnist":
        cf.dataset = args.data

    # elif args.data == "cifar10":
    #     print("Using cifar10 dataset")
    cf.neurons = [784, 128, 128, 128, 10]
    cf.n_layers = len(cf.neurons)
    cf.act_fn = F.RELU
    cf.var_out = 1
    cf.vars = torch.ones(cf.n_layers)

    cf.itr_max = 50
    cf.beta = 0.1
    cf.div = 2
    cf.condition = 1e-6
    cf.d_rate = 0

    # optim parameters
    cf.l_rate = 1e-3
    cf.optim = "ADAM"
    cf.eps = 1e-8
    cf.decay_r = 0.9
    cf.beta_1 = 0.9
    cf.beta_2 = 0.999

    cf.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
    main(cf)