testing french bulldog imagenet

app.py

@@ -24,12 +24,12 @@ BLENHEIM_SPANIEL_CLASS = 156
 def get_image_data(image_name):
     """Gets the image data and model."""
     image, model_and_data = [None, None]
-    if (image_name == "imagenet_diego.png"): 
+    if (image_name == "imagenet_diego"): 
         image = get_dataset_by_name("imagenet_diego", get_label=False)
         model_and_data = process_imagenet_get_model(image)
-    elif (image_name == "mnist_9.png"): 
-        image = get_dataset_by_name(image_name, get_label=False)
-        model_and_data = process_mnist_get_model(image)
+    elif (image_name == "imagenet_french_bulldog"): 
+        image = get_dataset_by_name("imagenet_french_bulldog", get_label=False)
+        model_and_data = process_imagenet_get_model(image)
     
 
     return image, model_and_data
@@ -49,7 +49,12 @@ def segmentation_generation(image_name, c_width, n_top, n_gif_imgs):
     segs = model_and_data["xtest_segs"]
     get_model = model_and_data["model"]
     label = model_and_data["label"]
-    print("LABEL:", label)
+    
+
+    if (image_name == 'imagenet_diego'):
+        label = 156
+    elif (image_name == 'imagenet_french_bulldog'):
+        label = 245
 
     # Unpack instance and segments
     instance = xtest[0]
@@ -62,7 +67,7 @@ def segmentation_generation(image_name, c_width, n_top, n_gif_imgs):
     xtrain = get_xtrain(segments)
 
     prediction = np.argmax(cur_model(xtrain[:1]), axis=1)
-    assert prediction == BLENHEIM_SPANIEL_CLASS, f"Prediction is {prediction} not {BLENHEIM_SPANIEL_CLASS}"
+    assert prediction == label, f"Prediction is {prediction} not {label}"
 
     # Compute explanation
     exp_init = BayesLocalExplanations(training_data=xtrain,
@@ -72,7 +77,7 @@ def segmentation_generation(image_name, c_width, n_top, n_gif_imgs):
                                               verbose=True)
     rout = exp_init.explain(classifier_f=cur_model,
                             data=np.ones_like(xtrain[0]),
-                            label=BLENHEIM_SPANIEL_CLASS,
+                            label=label,
                             cred_width=cred_width,
                             focus_sample=False,
                             l2=False)
@@ -93,11 +98,12 @@ if __name__ == "__main__":
         segmentation_generation, 
         [
             inp,
-            gr.inputs.Slider(minimum=0.01, maximum=0.8, step=0.001, default=0.1, label="cred_width", optional=False),
+            gr.inputs.Slider(minimum=0.01, maximum=0.8, step=0.01, default=0.1, label="cred_width", optional=False),
             gr.inputs.Slider(minimum=1, maximum=10, step=1, default=5, label="n_top_segs", optional=False),
             gr.inputs.Slider(minimum=10, maximum=50, step=1, default=20, label="n_gif_images", optional=False), 
         ], 
         outputs=out, 
-        examples=[["imagenet_diego.png", 0.01, 7, 50]]
+        examples=[["imagenet_diego", 0.01, 7, 50], 
+                  ["imagenet_french_bulldog", 0.05, 5, 50]]
     )
-    iface.launch(show_error=True, enable_queue=True)
+    iface.launch(enable_queue=True)