Update app.py

app.py

@@ -10,37 +10,45 @@ caption_model = BlipForConditionalGeneration.from_pretrained(caption_model_name)
 emotion_model_name = "SamLowe/roberta-base-go_emotions"
 emotion_classifier = pipeline("text-classification", model=emotion_model_name)
 
-def generate_caption_and_analyze_emotions(image):
+def generate_caption_and_analyze_emotions(image, text=None):
     try:
-        # Preprocess the image for caption generation
-        caption_inputs = caption_processor(images=image, return_tensors="pt")
+        if image is not None:
+            # Preprocess the image for caption generation
+            caption_inputs = caption_processor(images=image, return_tensors="pt")
 
-        # Generate caption using the caption model
-        caption_ids = caption_model.generate(**caption_inputs)
+            # Generate caption using the caption model
+            caption_ids = caption_model.generate(**caption_inputs)
 
-        # Decode the output caption
-        decoded_caption = caption_processor.decode(caption_ids[0], skip_special_tokens=True)
-
-        # Perform emotion analysis on the generated caption
+            # Decode the output caption
+            decoded_caption = caption_processor.decode(caption_ids[0], skip_special_tokens=True)
+        else:
+            decoded_caption = text
+        
+        # Perform emotion analysis on the generated caption or provided text
         results = emotion_classifier(decoded_caption)
         sentiment_label = results[0]['label']
         if sentiment_label == 'neutral':
-            sentiment_text = "Sentiment of the image is"
+            sentiment_text = "Sentiment of the text is"
         else:
-            sentiment_text = "Sentiment of the image shows"
+            sentiment_text = "Sentiment of the text shows"
 
-        final_output = f"This image shows '{decoded_caption}' and {sentiment_text} {sentiment_label}."
-        return final_output
+        caption_output = f"Caption: '{decoded_caption}'"
+        sentiment_output = f"{sentiment_text} {sentiment_label}."
+        
+        return caption_output, sentiment_output
     except Exception as e:
-        return f"An error occurred: {e}"
+        return f"An error occurred: {e}", ""
 
 # Define the Gradio interface using the new API
-inputs = gr.Image(label="Upload an image")
-outputs = gr.Textbox(label="Sentiment Analysis")
+image_input = gr.inputs.Image(label="Upload an image")
+text_input = gr.inputs.Textbox(label="Or enter text")
+
+outputs = [gr.outputs.Textbox(label="Generated Caption"), gr.outputs.Textbox(label="Sentiment Analysis")]
 
 # Create the Gradio app
-app = gr.Interface(fn=generate_caption_and_analyze_emotions, inputs=inputs, outputs=outputs)
+app = gr.Interface(fn=generate_caption_and_analyze_emotions, inputs=[image_input, text_input], outputs=outputs)
 
 # Launch the Gradio app
 if __name__ == "__main__":
     app.launch()
+