Update app.py

app.py

@@ -2,6 +2,7 @@ import gradio as gr
 import pickle
 import pandas as pd
 
+
 data = pickle.load(open("OOV_Train_2.pkl", "rb"))
 data = pd.DataFrame(
     data, 
@@ -40,27 +41,26 @@ Use the next page to check if your test-items (i.e. adjective, noun and hypernym
 
 examples = [["A red car is a vehicle"], ["A fake smile is a smile"], ["A small cat is a small animal"]]
 
-
-interface_words = gr.Interface(
-            fn=test_input,
-            inputs=gr.Textbox(label="Input:word_1,word2,...,word_n"),
-            outputs=gr.outputs.Label(label="In training-distribution?"),
-            examples=examples_w,
+interface_model = gr.Interface.load(
+            "huggingface/lorenzoscottb/bert-base-cased-PLANE-ood-2",
+            description=description,
+            examples=examples,
+            title=title,
 )
 
 
-
-description_e = """
+description_w = """
 You can use this page to test if a set of words was included in the training data used to tune the model. As in the samples below, use as input a series of words separated solely by a comma (e.g. *red,car,vehicle*).
 """
 
 examples_w = [["red,car,vehicle"], ["fake,smile"], ["small,cat,animal"]]
 
-interface_model = gr.Interface.load(
-            "huggingface/lorenzoscottb/bert-base-cased-PLANE-ood-2",
-            description=description,
-            examples=examples,
-            title=title,
+interface_words = gr.Interface(
+            fn=test_input,
+            inputs=gr.Textbox(label="Input:word_1,word2,...,word_n"),
+            outputs=gr.outputs.Label(label="In training-distribution?"),
+            description=description_w,
+            examples=examples_w,
 )
 
 gr.TabbedInterface(