adding text box

app.py

@@ -11,7 +11,7 @@ from skimage.measure import label, regionprops
 
 processor = CLIPSegProcessor.from_pretrained("CIDAS/clipseg-rd64-refined")
 model = CLIPSegForImageSegmentation.from_pretrained("CIDAS/clipseg-rd64-refined")
-global classes
+classes = list()
 
 def rescale_bbox(bbox,orig_image_shape=(1024,1024),model_shape=352):
     bbox = np.asarray(bbox)/model_shape
@@ -68,34 +68,6 @@ def add_text(text):
     labels = text.split(',')
     return labels 
 
-# inputt = gr.Image(type="numpy", label="Input Image for Classification")
-
-# with gr.Blocks(title="Zero Shot Object ddetection using Text Prompts") as demo :
-#     gr.Markdown(
-#     """ 
-#     <center>
-#     <h1>
-#     The CLIP Model  
-#     </h1>
-#     A neural network called CLIP which efficiently learns visual concepts from natural language supervision. CLIP can be applied to any visual classification benchmark by simply providing the names of the visual categories to be recognized, similar to the “zero-shot” capabilities of GPT-2 and GPT-3.   
-#     </center>
-#     """
-#     )
-
-#     with gr.Row():
-#         with gr.Column():
-#             inputt = gr.Image(type="numpy", label="Input Image for Classification")
-#             labels = gr.Textbox(label="Enter Label/ labels",placeholder="ex. car,person",scale=4)
-#             button = gr.Button(value="Locate objects")
-#         with gr.Column():
-#             outputs = gr.Image(type="numpy", label="Detected Objects with Selected Category")
-#             # dropdown = gr.Dropdown(labels,label="Select the category",info='Label selection panel')
-        
-#         # labels.submit(add_text, inputs=labels)
-#         button.click(fn=shot,inputs=[inputt,labels],api_name='Get labels')
-
-
-# demo.launch()
 iface = gr.Interface(fn=shot,
                     inputs = ["image","text",gr.Dropdown(classes, label="Category Label",info='Select Categories')],
                     outputs="label",