Update app.py

app.py

@@ -1,8 +1,6 @@
 import gradio as gr
 import torch
 from transformers import pipeline
-from nemo.collections.asr.models import EncDecMultiTaskModel
-from transformers import VitsTokenizer, VitsModel
 
 # Load Canary ASR model
 canary_model = EncDecMultiTaskModel.from_pretrained('nvidia/canary-1b')
@@ -10,8 +8,8 @@ decode_cfg = canary_model.cfg.decoding
 decode_cfg.beam.beam_size = 1
 canary_model.change_decoding_strategy(decode_cfg)
 
-# Load Phi-3 Mini-128K-Instruct LLM model
-phi_3_model_id = "microsoft/Phi-3-mini-128k-instruct"
+# Load Phi-3 Mini-4K-Instruct LLM model
+phi_3_model_id = "microsoft/Phi-3-mini-4k-instruct"
 phi_3_pipeline = pipeline("text-generation", model=phi_3_model_id, trust_remote_code=True)
 
 # Load VITS TTS model
@@ -23,7 +21,7 @@ def transcribe_audio(audio):
     return transcribed_text
 
 def generate_response(prompt):
-    response = phi_3_pipeline(prompt)[0]['generated_text']
+    response = phi_3_pipeline(prompt, max_length=50, num_return_sequences=1)[0]['generated_text']
     return response
 
 def synthesize_speech(text):