Update app.py

app.py

@@ -9,6 +9,8 @@ import logging
 import time
 import uuid
 import soundfile as sf
+from PIL import Image
+
 # model.py apache license 2.0 Copyright      2022-2023  Xiaomi Corp.        (authors: Fangjun Kuang)
 from model import get_pretrained_model, language_to_models
 # demo for a input given image transform into text interpretation, and those text put a speech text to be played 
@@ -63,16 +65,14 @@ def text_to_speech(language: str, repo_id: str, text: str, sid: str, speed: floa
 
     return filename, build_html_output(info)
 
-demo = gr.Blocks()
 
-with demo:
+with gr.Blocks() as demo:
     language_choices = list(language_to_models.keys())
-    inputsImg=gr.Image(type='pil')
+    inputsImg=gr.Image(type='PIL')
     idx=0
-    text_output = image_to_text_model(inputsImg)
-
+    text_output = image_to_text_model(inputsImg)[0]
     print(text_output)
-    for txt in text_output:        
+    for txt in t ext_output:        
         output_txt[idx] = gr.Textbox(label=text_output,lines=1,max_lines=1,value=text_output,placeholder="Interpretation")
         input_sid = gr.Textbox(
                 label="Speaker ID",