Update app.py

app.py

@@ -1,6 +1,8 @@
 import os
+from PIL import Image, ImageDraw, ImageFont
 import gradio as gr
-from helper import load_image_from_url, render_results_in_image
+from helper import load_image_from_url, render_results_in_image 
+from helper import summarize_predictions_natural_language
 from transformers import pipeline
 from transformers.utils import logging
 logging.set_verbosity_error()
@@ -9,7 +11,12 @@ from helper import ignore_warnings
 ignore_warnings()
 
 
-od_pipe = pipeline("object-detection", "./models/facebook/detr-resnet-50")
+
+
+od_pipe = pipeline("object-detection", "facebook/detr-resnet-50")
+tts_pipe = pipeline("text-to-speech",
+                    model="kakao-enterprise/vits-ljs")
+
 
 def get_pipeline_prediction(pil_image):
     
@@ -17,14 +24,26 @@ def get_pipeline_prediction(pil_image):
     
     processed_image = render_results_in_image(pil_image,
                                             pipeline_output)
-    return processed_image
+    
+    text = summarize_predictions_natural_language(pipeline_output)
+    print(text)
+    narrated_text = tts_pipe(text)
+
+    #print (narrated_text)
+    print(narrated_text["audio"][0])
+    print (narrated_text["sampling_rate"])
+    return processed_image, (narrated_text["sampling_rate"], narrated_text["audio"][0] )
+    #return processed_image
+
 
 demo = gr.Interface(
   fn=get_pipeline_prediction,
   inputs=gr.Image(label="Input image", 
                   type="pil"),
-  outputs=gr.Image(label="Output image with predicted instances",
-                   type="pil")
+  outputs=[gr.Image(label="Output image with predicted instances",
+                   type="pil"), gr.Audio(label="Narration", type="numpy",  autoplay=True)]
+  #outputs=gr.Image(label="Output image with predicted instances",
+  #                 type="pil")
 )
 
-demo_blocks.queue().launch(server_name="0.0.0.0", server_port=7860)
+demo.launch(server_name="0.0.0.0", server_port=7860)