STN-YOLO/demo.py at main · Advanced-Vision-and-Learning-Lab/STN-YOLO · 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
# -*- coding: utf-8 -*-
"""
Main script for LACE experiments
"""
import argparse

import numpy as np
import torch
import torch.nn as nn
import torch.optim as optim
import matplotlib.pyplot as plt
import pdb
from Utils.Initailize_model import initialize_model
from yaml import load , CLoader as Loader
#Turn off plotting
plt.ioff()


def main(args):

    with open(args.yaml , 'r') as stream:
        out = load(stream, Loader=Loader)

    # Name of dataset
    Dataset =out['roboflow']['project']

    # Model(s) to be used
    model_name = args.model

    # Number of classes in dataset
    num_classes = out['nc']

    # Number of runs and/or splits for dataset
    numRuns = out['num_runs']

    # Detect if we have a GPU available
    device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

    print('Starting Experiments...')

    for split in range(0, numRuns):
        #Set same random seed based on split and fairly compare
        torch.manual_seed(split)
        np.random.seed(split)
        np.random.seed(split)
        torch.cuda.manual_seed(split)
        torch.cuda.manual_seed_all(split)
        torch.manual_seed(split)


        # Initialize the histogram model for this run
        model_ft, input_size = initialize_model(model_name, num_classes,
                                                feature_extract=args.feature_extraction,
                                                use_pretrained=args.use_pretrained)


        # Send the model to GPU if available, use multiple if available
        if torch.cuda.device_count() > 1:
            print("Using", torch.cuda.device_count(), "GPUs!")
            # dim = 0 [30, xxx] -> [10, ...], [10, ...], [10, ...] on 3 GPUs
            list_device= np.arange(0, torch.cuda.device_count(), 1)
            list_device.tolist()
        else:
            list_device = device.type

        model_ft.info()
        value=model_ft.info()
        num_params = value[1]

        if args.feature_extraction:
            mode = 'feature_extract'
        else:
            mode = 'fine_tuning'

        # TRAIN VAL AND TEST USING ULTRALETYCS and store in runs
        project = '{}/{}/{}/{}'.format(args.folder,
                                            mode,
                                            Dataset,
                                            model_name)
        name='Run_{}'.format(split+1)
        #TRAIN
        model_ft.train(data=args.yaml, epochs= args.num_epochs, batch=args.train_batch_size ,imgsz=args.resize_size,
                        device= list_device,pretrained=args.use_pretrained, project = project, name= name,save=args.save_results,
                       seed = split)

        #pretrained and freeze already there ( implement later) & dropout

        #  VALIDATE
        # model_ft.val(save=args.save_results,batch = args.val_batch_size)

        # TEST
        # model_ft.val(split='test',save=args.save_results)
        # If explicitly want to see the model info such as the number of the parameters and much more.
        #model_ft.info()
        print('**********Run ' + str(split + 1) + model_name + ' Finished**********')


def parse_args():
    parser = argparse.ArgumentParser(description='Run Angular Losses and Baseline experiments for dataset')
    parser.add_argument('--save_results', default=True, action=argparse.BooleanOptionalAction,
                        help='Save results of experiments(default: True)')
    parser.add_argument('--folder', type=str, default='path/to/your/Saved_Models/',
                        help='Location to save models')
    parser.add_argument('--yaml', type=str, default='/path/to/your/data.yaml',
                        help='Location to save models')
    parser.add_argument('--feature_extraction', default=True, action=argparse.BooleanOptionalAction,
                        help='Flag for feature extraction. False, train whole model. True, only update fully connected/encoder parameters (default: True)')
    parser.add_argument('--use_pretrained', default=True, action=argparse.BooleanOptionalAction,
                        help='Flag to use pretrained model from ImageNet or train from scratch (default: True)')
    parser.add_argument('--train_batch_size', type=int, default=4,
                        help='input batch size for training (default: 128)')
    parser.add_argument('--val_batch_size', type=int, default=32,
                        help='input batch size for validation (default: 512)')
    parser.add_argument('--test_batch_size', type=int, default=32,
                        help='input batch size for testing (default: 256)')
    parser.add_argument('--num_epochs', type=int, default=15,
                        help='Number of epochs to train each model for (default: 50)')
    parser.add_argument('--resize_size', type=int, default=640,
                        help='Resize the image before center crop. (default: 256)')
    parser.add_argument('--lr', type=float, default=0.001,
                        help='learning rate (default: 0.01)')
    parser.add_argument('--model', type=str, default='YOLOV8L',
                        help='backbone architecture to use (default: 0.01)')
    parser.add_argument('--use-cuda', action='store_true', default=True,
                        help='enables CUDA training')
    args = parser.parse_args()
    return args

if __name__ == "__main__":
    args = parse_args()
    use_cuda = args.use_cuda and torch.cuda.is_available()
    device = torch.device("cuda" if use_cuda else "cpu")
    main(args)