import gradio as gr from torchvision import transforms import torch from utils import CustomResnet, main_inference, get_misclassified_images, get_gradcam inv_normalize = transforms.Normalize( mean=[-0.50/0.23, -0.50/0.23, -0.50/0.23], std=[1/0.23, 1/0.23, 1/0.23] ) model = CustomResnet() classes = ('plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck') targets = None device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') model.load_state_dict(torch.load("best_model.pth",map_location=torch.device("cpu")),strict=False) model.to(device) # Define the input and output components of the Gradio app input_component = gr.inputs.Image(shape=(32, 32)) num_of_output_classes = gr.inputs.Slider(minimum=0, maximum=10, default=5, step=1,label="Top class count") # Adding a checkbox to the interface to show/hide misclassified images show_misclassified_checkbox = gr.inputs.Checkbox(default=False, label="Show Misclassified Images") # Input field to specify the number of misclassified images to display num_images_input = gr.inputs.Slider(minimum=0, maximum=20, default=15, step=5,label="Missclassified Images Count") # Adding a checkbox to the interface to show/hide GradCAM output show_gradcam_checkbox = gr.inputs.Checkbox(default=False, label="Show GradCAM Output") # Slider for adjusting the opacity of the GradCAM overlay opacity_slider = gr.inputs.Slider(minimum=0, maximum=1, default=0.7,step=0.1, label="GradCAM Opacity") #layer_options = ['layer1', 'layer2', 'layer3'] layer_options = ['convblock1', 'resblock1','convblock2',"convblock3","resblock2"] layer_input = gr.inputs.Dropdown(layer_options,label="Select a Layer",default="convblock3") gr.Interface( fn=lambda image, num_of_output_classes,show_misclassified, num_images, show_gradcam, opacity,layer: [main_inference(num_of_output_classes,classes,model,image), get_misclassified_images(show_misclassified, num_images) if show_misclassified else None, get_gradcam(model,image, opacity,layer) if show_gradcam else None], inputs=[input_component, num_of_output_classes,show_misclassified_checkbox, num_images_input, show_gradcam_checkbox, opacity_slider,layer_input], outputs=[gr.outputs.Label(), gr.Image(shape=(500, 500)), gr.Image(shape=(250, 250))], title="CIFAR10 Trained on Custom Residual CNN Architecture", examples=[ ["example_images/example_1.png",5,True,5,True,0.2,'convblock3'], # You can provide your own example input values here ["example_images/example_2.png",5,False,5,True,0.3,'convblock3'], ["example_images/example_3.png",5,True,15,False,0.2,'convblock3'] , ["example_images/example_4.png",5,True,20,True,0.5,'convblock3'] , ["example_images/example_5.png",5,False,5,False,0.2,'convblock3'] , ["example_images/example_6.png",5,True,10,True,0.3,'convblock3'] , ["example_images/example_7.png",5,True,5,True,0.4,'convblock3'] , ["example_images/example_8.png",5,False,5,False,0.6,'convblock3'] , ["example_images/example_9.png",5,True,20,False,0.2,'convblock3'] , ["example_images/example_10.png",5,False,5,True,0.7,'convblock3'] ], layout="horizontal" ).launch()