Update app.py

app.py

@@ -62,7 +62,7 @@ def generate_llm_response(text, model_id="ccibeekeoc42/Llama3.1-8b-base-SFT-2024
         else:
             raise e
 
-generate_llm_response("Explain Deep Learning in Igbo")
+# generate_llm_response("Explain Deep Learning in Igbo")
 
 
 # Loading the ST Model (Whisper) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
@@ -186,11 +186,11 @@ def replace_numbers_with_words(text, lang="en"):
     # Replace all numbers in the text
     return re.sub(r'\b\d+\b', replace, text)
 
-llm_response = generate_llm_response("Explain Deep Learning in Igbo")
-llm_response_cleaned = normalize_text(cleanup_text(replace_numbers_with_words(llm_response, "yo"), "yo"))
+# llm_response = generate_llm_response("Explain Deep Learning in Igbo")
+# llm_response_cleaned = normalize_text(cleanup_text(replace_numbers_with_words(llm_response, "yo"), "yo"))
 
-print(f"LLM Response: {llm_response}")
-print(f"LLM Response Cleaned: {llm_response_cleaned}")
+# print(f"LLM Response: {llm_response}")
+# print(f"LLM Response Cleaned: {llm_response_cleaned}")
 
 # returning spech from text (and bringing to CPU)
 def synthesise(text):
@@ -214,14 +214,17 @@ def speech_to_speech_translation(audio):
     print(f"Transcribed: {transcribed_text}")
 
     # Generate LLM Response
+    print("Now making LLM Call ~~~~~~~~~~~~~~~~~~~~~~~~")
     llm_response = generate_llm_response(transcribed_text)
     llm_response_cleaned = normalize_text(cleanup_text(replace_numbers_with_words(llm_response, "yo"), "yo"))
     print(f"LLM Response: {llm_response}")
     print(f"LLM Response Cleaned: {llm_response_cleaned}")
 
     # Text to Speech
+    print("Synthesizing Speech ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~")
     synthesised_speech = synthesise(llm_response_cleaned)
     synthesised_speech = (synthesised_speech.numpy() * max_range).astype(np.int16)
+    print("Speech Synthesis Completed~~~~~~~~~~~~~~~~~~~")
 
     return transcribed_text, (16000, synthesised_speech), llm_response