Update app.py

app.py

@@ -9,15 +9,11 @@ import traceback
 from pydub import AudioSegment
 import librosa
 from utils.vad import get_speech_timestamps, collect_chunks, VadOptions
-from transformers import MllamaForConditionalGeneration, AutoProcessor, TextIteratorStreamer
-import torch
-from huggingface_hub import login
+from huggingface_hub import InferenceClient
 import os
 from PIL import Image
 from threading import Thread
-ckpt = "meta-llama/Llama-3.2-11B-Vision-Instruct"
-model = MllamaForConditionalGeneration.from_pretrained(ckpt,torch_dtype=torch.bfloat16).to("cpu")
-processor = AutoProcessor.from_pretrained(ckpt)
+client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 r = sr.Recognizer()
 
 @dataclass