Update app.py

app.py

@@ -1,11 +1,13 @@
 import spaces
 import gradio as gr
-from huggingface_hub import InferenceClient
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
+from threading import Thread
 
 """
 For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
 """
-client = InferenceClient("cognitivecomputations/dolphin-2.8-mistral-7b-v02")
+# client = InferenceClient("cognitivecomputations/dolphin-2.8-mistral-7b-v02")
 
 @spaces.GPU
 def respond(