Update app.py

app.py

@@ -11,6 +11,7 @@ pipe = pipeline(
 translator = pipeline("text2text-generation", model="Baghdad99/saad-hausa-text-to-english-text")
 tts = pipeline("text-to-speech", model="Baghdad99/english_voice_tts")
 
+# Define the function to translate speech
 def translate_speech(audio):
     # Separate the sample rate and the audio data
     sample_rate, audio_data = audio
@@ -19,35 +20,24 @@ def translate_speech(audio):
     output = pipe(audio_data)
     print(f"Output: {output}")  # Print the output to see what it contains
 
-    # Check if the output contains 'transcription'
-    if 'transcription' in output:
-        transcription = output["transcription"]
+    # Check if the output contains 'text'
+    if 'text' in output:
+        transcription = output["text"]
     else:
-        print("The output does not contain 'transcription'")
+        print("The output does not contain 'text'")
         return
 
-    # Rest of your code...
+    # Use the translation pipeline to translate the transcription
+    translated_text = translator(transcription, return_tensors="pt", padding=True)
 
-# # Define the function to translate speech
-# def translate_speech(audio):
-#     # Separate the sample rate and the audio data
-#     sample_rate, audio_data = audio
+    # Use the text-to-speech pipeline to synthesize the translated text
+    synthesised_speech = tts(translated_text, return_tensors='pt')
 
-#     # Use the speech recognition pipeline to transcribe the audio
-#     transcription = pipe(audio_data)["transcription"]
-
-#     # Use the translation pipeline to translate the transcription
-#     translated_text = translator(transcription, return_tensors="pt", padding=True)
-
-#     # Use the text-to-speech pipeline to synthesize the translated text
-#     synthesised_speech = tts(translated_text, return_tensors='pt')
-
-#     # Define the max_range variable
-#     max_range = 32767  # You can adjust this value based on your requirements
-#     synthesised_speech = (synthesised_speech.numpy() * max_range).astype(np.int16)
-
-#     return 16000, synthesised_speech
+    # Define the max_range variable
+    max_range = 32767  # You can adjust this value based on your requirements
+    synthesised_speech = (synthesised_speech.numpy() * max_range).astype(np.int16)
 
+    return 16000, synthesised_speech
 
 # Define the Gradio interface
 iface = gr.Interface(