Update app.py

app.py

@@ -1,7 +1,7 @@
 import gradio as gr
 import gradio as gr
 import torch
-from transformers import AutoTokenizer, Mistral3ForConditionalGeneration, TextIteratorStreamer,BitsAndBytesConfig
+from transformers import AutoTokenizer, Mistral3ForConditionalGeneration, TextIteratorStreamer
 from threading import Thread
 import re
 import time
@@ -13,7 +13,7 @@ import spaces
 
 # Model configuration
 MODEL_NAME = "yasserrmd/SinaReason-Magistral-2509"
-DEVICE = "cuda" 
+DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
 
 # Medical system prompt as recommended by the model card
 MEDICAL_SYSTEM_PROMPT = """
@@ -44,20 +44,14 @@ class SinaReasonMedicalChat:
                 "mistralai/Magistral-Small-2509",
                 tokenizer_type="mistral"
             )
-            bnb_config = BitsAndBytesConfig(
-                load_in_4bit=True,
-                bnb_4bit_use_double_quant=True,
-                bnb_4bit_quant_type="nf4",         # normal float4
-                bnb_4bit_compute_dtype=torch.bfloat16   # computation precision
-            )
-                        
+         
             # Add padding token if not present
             if self.tokenizer.pad_token is None:
                 self.tokenizer.pad_token = self.tokenizer.eos_token
             
             self.model = Mistral3ForConditionalGeneration.from_pretrained(
                 MODEL_NAME,
-                quantization_config=bnb_config
+                dtype="auto"
             )