Update app.py

app.py

@@ -9,18 +9,14 @@ def alexa(audio):
     return speech
 
 def speech_to_text(audio):
-    audio_to_text = pipeline("automatic-speech-recognition", model="openai/whisper-tiny")
-    if audio == None:
-        raise gr.Error("Submit an audio file")
-    else:  
-        text = audio_to_text(audio)["text"]
+    audio_to_text = pipeline("automatic-speech-recognition", model="openai/whisper-tiny")    
+    text = audio_to_text(audio,generate_kwargs={"task": "transcribe", "language": "english"})["text"]
     return text
 
 def text_generation(text):
     model = AutoModelForCausalLM.from_pretrained("microsoft/Phi-3-mini-128k-instruct", 
-    torch_dtype="auto", 
-    trust_remote_code=True, )
-    tokenizer = AutoTokenizer.from_pretrained("microsoft/Phi-3-mini-128k-instruct")
+    trust_remote_code=True)
+    tokenizer = AutoTokenizer.from_pretrained("microsoft/Phi-3-mini-128k-instruct"), trust_remote_code=True)
 
     messages = [
         {"role": "user", "content": text}
@@ -33,8 +29,8 @@ def text_generation(text):
         "do_sample": True,
     }
     
-    text_gen= pipeline("text-generation", model="microsoft/Phi-3-mini-128k-instruct", tokenizer=tokenizer, trust_remote_code = True)
-    response = text_gen(messages, **generation_args)
+    text_gen= pipeline("text-generation", model=model, tokenizer=tokenizer, trust_remote_code = True)
+    response = text_gen(messages, generation_args)
     return response[0]["generated_text"]
 
 def text_to_speech(text):
@@ -45,4 +41,4 @@ def text_to_speech(text):
 gr.Interface(
   fn=alexa,
   inputs=gr.Audio(type="filepath"),
-  outputs=[gr.Audio(label="Audio", type="numpy",  autoplay=True)]).launch()
+  outputs=[gr.Audio(label="Audio", type="numpy",  autoplay=False)]).launch()