Update app.py

app.py

@@ -6,7 +6,7 @@ import torch
 
 
 pipe1 = pipeline("automatic-speech-recognition", model="openai/whisper-base")
-pipe2 = pipeline("summarization", model="facebook/bart-large-cnn")
+#pipe2 = pipeline("summarization", model="facebook/bart-large-cnn")
 #pipe3 = DiffusionPipeline.from_pretrained("stable-diffusion-v1-5/stable-diffusion-v1-5")
 #pipe3.to("cuda" if torch.cuda.is_available() else "cpu")
 
@@ -16,18 +16,20 @@ def audio_to_image(audio):
 
         transcription_text = transcription['text']
         
-        summary = pipe2(transcription_text, max_length=50, min_length=10, do_sample=False)
-        summary_text = summary[0]['summary_text']
+        #summary = pipe2(transcription_text, max_length=50, min_length=10, do_sample=False)
+        #summary_text = summary[0]['summary_text']
        
         #prompt = summary_text
         #image = pipe3(prompt).images[0]
         
         #return image
-        print("Transcription:", transcription_text)
-        print("Summary:", summary_text)
-        return transcription_text, summary_text
+        #print("Transcription:", transcription_text)
+        #print("Summary:", summary_text)
+        #return transcription_text, summary_text
+        return transcription_text
 
     
 #demo = gr.Interface(fn=audio_to_image, inputs=gr.Audio(), outputs="image")
-demo = gr.Interface(fn=audio_to_image, inputs=gr.Audio(), outputs=[gr.Textbox(label="Transcription"), gr.Textbox(label="Summary")])
+#demo = gr.Interface(fn=audio_to_image, inputs=gr.Audio(), outputs=[gr.Textbox(label="Transcription"), gr.Textbox(label="Summary")])
+demo = gr.Interface(fn=audio_to_image, inputs=gr.Audio(), outputs="text")
 demo.launch(share=True)