app.py

import torch
import torch.nn as nn
import numpy as np
from torchvision import models, transforms
import time
import os
import copy
import pickle
from PIL import Image
import datetime
import gdown
import urllib.request
import gradio as gr
import markdown

# load model state and class names from gdrive
# issue accessing the link. updated permissions
#https://drive.google.com/file/d/1m9C-WMfKRDCmScxTh8JmcoFtymxAqjS3/view?usp=sharing
# url = 'https://drive.google.com/uc?id=1m9C-WMfKRDCmScxTh8JmcoFtymxAqjS3'
path_class_names = "./class_names_restnet_leeds_butterfly.pkl"

# MAX_TRIES = 10
# i = 0
# while(i<MAX_TRIES):
#     try:
#         gdown.download(url, path_class_names, quiet=False)
#         break
#     except:
#         pass
#     i += 1
    
# gdown.download(url, path_class_names, quiet=False)
#https://drive.google.com/file/d/1qxaWnYwLIwWGrGg9uehG7h2W227SXGKq/view?usp=sharing
# url = 'https://drive.google.com/uc?id=1qxaWnYwLIwWGrGg9uehG7h2W227SXGKq'
path_model = "./model_state_restnet_leeds_butterfly.pth"
# gdown.download(url, path_model, quiet=False)
# i = 0
# while(i<MAX_TRIES):
#     try:
#         gdown.download(url, path_model, quiet=False)
#         break
#     except:
#         pass
#     i += 1

# example images
url = "https://upload.wikimedia.org/wikipedia/commons/thumb/f/f8/Red_postman_butterfly_%28Heliconius_erato%29.jpg/1599px-Red_postman_butterfly_%28Heliconius_erato%29.jpg"
path_input = "./h_erato.jpg"
urllib.request.urlretrieve(url, filename=path_input)

url = "https://upload.wikimedia.org/wikipedia/commons/thumb/6/63/Monarch_In_May.jpg/1024px-Monarch_In_May.jpg"
path_input = "./d_plexippus.jpg"
urllib.request.urlretrieve(url, filename=path_input)

url = "https://drive.google.com/uc?id=1A7WgDrQ_RLO6JOQiYhkH_hj_EKcbpmOl"
path_input = "./v_cardui.jpg"
urllib.request.urlretrieve(url, filename=path_input)

url = "https://drive.google.com/uc?id=1CiWShQYIm2N0fkVaWJpftlXZFqwjsXhA"
path_input = "./p_cresphontes.jpg"
urllib.request.urlretrieve(url, filename=path_input)

url = "https://drive.google.com/uc?id=1r8rbkUwTSIZL0MQVgU-WjDGwvLXuwYPG"
path_input = "./p_rapae.jpg"
urllib.request.urlretrieve(url, filename=path_input)


# normalisation
data_transforms_test = transforms.Compose([
        transforms.Resize(256),
        transforms.CenterCrop(224),
        transforms.ToTensor(),
        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
    ])

class_names = pickle.load(open(path_class_names, "rb"))

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

model_ft = models.resnet18(pretrained=True)
num_ftrs = model_ft.fc.in_features
model_ft.fc = nn.Linear(num_ftrs, len(class_names))
model_ft = model_ft.to(device)
model_ft.load_state_dict(copy.deepcopy(torch.load(path_model,device)))

# Proper labeling
id_to_name = {
    '001_Danaus Plexippus': 'Danaus plexippus - Monarch',
    '002_Heliconius Charitonius': 'Heliconius charitonius - Zebra Longwing',
    '003_Heliconius Erato': 'Heliconius erato - Red Postman',
    '004_Junonia Coenia': 'Junonia coenia - Common Buckeye',
    '005_Lycaena Phlaeas': 'Lycaena phlaeas - Small Copper',
    '006_Nymphalis Antiopa': 'Nymphalis antiopa - Mourning Cloak',
    '007_Papilio Cresphontes': 'Papilio cresphontes - Giant Swallowtail',
    '008_Pieris Rapae': 'Pieris rapae - Cabbage White',
    '009_Vanessa Atalanta': 'Vanessa atalanta - Red Admiral',
    '010_Vanessa Cardui': 'Vanessa cardui - Painted Lady',
}

def do_inference(img):
  img_t = data_transforms_test(img)
  batch_t = torch.unsqueeze(img_t, 0)
  model_ft.eval()
  # We don't need gradients for test, so wrap in 
  # no_grad to save memory
  with torch.no_grad():
      batch_t = batch_t.to(device)
      # forward propagation
      output = model_ft( batch_t)
      # get prediction
      probs = torch.nn.functional.softmax(output, dim=1)
      output = torch.argsort(probs, dim=1, descending=True).cpu().numpy()[0].astype(int)
      probs = probs.cpu().numpy()[0]
      probs = probs[output]
      labels = np.array(class_names)[output]
      
      if(probs[0] < 0.5):
          return {"No butterfly":1.0}
      else:  
        return {id_to_name[labels[i]]: round(float(probs[i]),2) for i in range(len(labels))}

im = gr.inputs.Image(shape=(512, 512), image_mode='RGB', 
                         invert_colors=False, source="upload",
                         type="pil")

title = "Butterfly Classification Demo"
description = "A pretrained ResNet18 CNN trained on the Leeds Butterfly Dataset. Libraries: PyTorch, Gradio."
examples = [['./h_erato.jpg'],['d_plexippus.jpg'],['v_cardui.jpg'],['p_cresphontes.jpg'],['p_rapae.jpg']]
article_text = markdown.markdown('''

<h1 style="color:white">PyTorch image classification - A pretrained ResNet18 CNN trained on the <a href="http://www.josiahwang.com/dataset/leedsbutterfly/" target="_blank">Leeds Butterfly Dataset</a></h1>
<br>
<p>The Leeds Butterfly Dataset consists of 832 images in 10 classes:</p>
<ul>
  <li>Danaus plexippus - Monarch</li>
  <li>Heliconius charitonius - Zebra Longwing</li>
  <li>Heliconius erato - Red Postman</li>
  <li>Lycaena phlaeas - Small Copper</li>
  <li>Junonia coenia - Common Buckeye</li>
  <li>Nymphalis antiopa - Mourning Cloak</li>
  <li>Papilio cresphontes - Giant Swallowtail</li>
  <li>Pieris rapae - Cabbage White</li>
  <li>Vanessa atalanta - Red Admiral</li>
  <li>Vanessa cardui - Painted Lady</li>
</ul>
<br>
<p>Part of a dissertation project. Author: <a href="https://github.com/ttheland" target="_blank">ttheland</a></p>
''')

iface = gr.Interface(
    do_inference,
    im,
    gr.outputs.Label(num_top_classes=3),
    live=False,
    interpretation=None,
    title=title,
    description=description,
    article= article_text,
    examples=examples,
    theme="dark-peach"
)

iface.test_launch()

iface.launch(share=True, enable_queue=True)