update asr model

app.py

@@ -1,3 +1,4 @@
+import nemo.collections.asr as nemo_asr
 from transformers import pipeline
 import numpy as np
 import gradio as gr
@@ -14,8 +15,10 @@ def transcribe(audio):
   result = asr_model({"sampling_rate": sr, "raw": y})["text"]
   return result
 
-asr_model_id = "openai/whisper-small.en"
-asr_model = pipeline("automatic-speech-recognition", model=asr_model_id)
+# asr_model_id = "openai/whisper-small.en"
+# asr_model = pipeline("automatic-speech-recognition", model=asr_model_id)
+asr_model = nemo_asr.models.EncDecCTCBPEModel.from_pretrained(model_name="nvidia/parakeet-ctc-0.6b")
+text = asr_model.transcribe(["./Samples/Sample_audios/test.wav"])
 
 with gr.Blocks() as demo:
   with gr.Column():
@@ -23,6 +26,8 @@ with gr.Blocks() as demo:
       """
       # HKU Canteen VA
       """)
+    gr.Markdown(
+      f"{text}")
     va = gr.Chatbot(container=False)
 
     with gr.Row(): # text input

requirements.txt

@@ -1,3 +1,4 @@
 torch
 transformers
+nemo_toolkit['all']
 numpy