File size: 5,045 Bytes
df7f40f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3b11ca6
 
 
 
 
 
 
 
 
 
df7f40f
 
 
 
 
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
import gradio as gr
import os
import numpy as np
import pandas as pd
import torch
import timm
from PIL import Image
from torchvision import transforms
import gradio
import warnings
warnings.filterwarnings("ignore")

model = timm.create_model('swinv2_cr_tiny_ns_224.sw_in1k', pretrained=True)
output_shape = 60
model.classifier = torch.nn.Sequential(
    torch.nn.Dropout(p=0.2, inplace=True),
    torch.nn.Linear(in_features=1000,
                   out_features=output_shape,
                   bias=True).to('cpu'))
model.load_state_dict(torch.load('./swin_70_65.pth', map_location=torch.device('cpu')))

preprocess = transforms.Compose([
    transforms.Resize((224, 224)),
    transforms.ToTensor(),
    transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
])

class_names = [
    'Ahaetulla_prasina', 'Bitis_arietans', 'Boa_constrictor', 'Boa_imperator',
    'Bothriechis_schlegelii', 'Bothrops_asper', 'Bothrops_atrox', 'Bungarus_fasciatus',
    'Chrysopelea_ornata', 'Coelognathus_radiatus', 'Corallus_hortulana', 'Coronella_austriaca',
    'Crotaphopeltis_hotamboeia', 'Dendrelaphis_pictus', 'Dolichophis_caspius', 'Drymarchon_melanurus',
    'Drymobius_margaritiferus', 'Elaphe_dione', 'Epicrates_cenchria', 'Erythrolamprus_poecilogyrus',
    'Eunectes_murinus', 'Fowlea_flavipunctata', 'Gonyosoma_oxycephalum', 'Helicops_angulatus',
    'Hierophis_viridiflavus', 'Imantodes_cenchoa', 'Indotyphlops_braminus', 'Laticauda_colubrina',
    'Leptodeira_annulata', 'Leptodeira_ornata', 'Leptodeira_septentrionalis', 'Leptophis_ahaetulla',
    'Leptophis_mexicanus', 'Lycodon_capucinus', 'Malayopython_reticulatus', 'Malpolon_insignitus',
    'Mastigodryas_boddaerti', 'Natrix_helvetica', 'Natrix_maura', 'Natrix_natrix', 'Natrix_tessellata',
    'Ninia_sebae', 'Ophiophagus_hannah', 'Oxybelis_aeneus', 'Oxybelis_fulgidus', 'Oxyrhopus_petolarius',
    'Phrynonax_poecilonotus', 'Psammodynastes_pulverulentus', 'Ptyas_korros', 'Ptyas_mucosa',
    'Python_bivittatus', 'Rhabdophis_tigrinus', 'Sibon_nebulatus', 'Spilotes_pullatus',
    'Tantilla_melanocephala', 'Trimeresurus_albolabris', 'Vipera_ammodytes', 'Vipera_aspis',
    'Vipera_berus', 'Zamenis_longissimus'
]


def predict(image):
    if image is None:
        return "No image provided."

    try:
        input_tensor = preprocess(image)
    except Exception as e:
        return f"Error in preprocessing: {str(e)}"

    input_batch = input_tensor.unsqueeze(0).to('cpu')

    try:
        with torch.no_grad():
            output = model(input_batch)
    except Exception as e:
        return f"Error in model inference: {str(e)}"

    probabilities = torch.nn.functional.softmax(output, dim=1)
    percentages = probabilities[0].cpu().numpy() * 100

    top_n = 5
    combined = list(zip(class_names, percentages))
    sorted_combined = sorted(combined, key=lambda x: x[1], reverse=True)
    top_predictions = sorted_combined[:top_n]
    
    # Generate HTML for progress bars with numbers above
    html_content = "<div style='font-family: Arial, sans-serif;'>"
    for class_label, confidence in top_predictions:
        html_content += f"""
        <div style='margin-bottom: 10px; position: relative;'>
            <div style='display: flex; align-items: center;'>
                <strong style='flex: 1;'>{class_label}</strong>
                <span style='flex-shrink: 0; color: black; margin-left: 10px;'>
                    {confidence:.2f}%
                </span>
            </div>
            <div style='background-color: #f3f3f3; border-radius: 5px; width: 100%; height: 20px; margin-top: 5px;'>
                <div style='background-color: #4CAF50; height: 100%; width: {confidence:.2f}%; border-radius: 5px;'></div>
            </div>
        </div>
        """
    html_content += "</div>"
    
    return html_content


interface = gr.Interface(
    fn=predict,
    inputs=gr.Image(type='pil'),
    outputs=gr.HTML(),
    title="Snake Species Classification (SnakeCLEF2024)",
    description =  """
    <div style='font-family: Arial, sans-serif; line-height: 1.6;'>
        <p>Datasets and classes are referenced from: <a href="https://www.imageclef.org/node/319" target="_blank">ImageCLEF</a> and for more details: <a href="https://github.com/Tanaanan/SnakeCLEF2024_MLCS" target="_blank">GitHub repository</a></p>
        <p style='font-size: smaller;'>This project is part of the course 'Machine Learning Systems (01418262).'</p>
        <p style='font-size: smaller;'>Developed by Tanaanan, Narakorn, Department of Computer Science, Kasetsart University.</p>
    </div>
    """,
    examples=['./sample_imgs/Bitis_arietans.png',
 './sample_imgs/Boa_imperator.png',
 './sample_imgs/Coelognathus_radiatus.png',
 './sample_imgs/Leptodeira_septentrionalis.png',
 './sample_imgs/Natrix_tessellata.png',
 './sample_imgs/Psammodynastes_pulverulentus.png',
 './sample_imgs/Ptyas_mucosa.png',
 './sample_imgs/Vipera_ammodytes.png',
 './sample_imgs/Zamenis_longissimus.png']
,
    live=True
)

if __name__ == "__main__":
    interface.launch()