Update app.py

app.py

@@ -5,20 +5,25 @@ import torch
 from threading import Thread
 import bitsandbytes as bnb
 
-phi4_model_path = "Compumacy/OpenBioLLm-70B"
+phi4_model_path = "Daemontatox/Qwen3-14B-Griffon"
 
 device = "cuda:0" if torch.cuda.is_available() else "cpu"
 
 # Load model with 4-bit quantization
+from transformers import BitsAndBytesConfig
+
+# Configure 4-bit quantization
+quantization_config = BitsAndBytesConfig(
+    load_in_4bit=True,
+    bnb_4bit_compute_dtype=torch.float16,
+    bnb_4bit_use_double_quant=True,
+    bnb_4bit_quant_type="nf4"
+)
+
 phi4_model = AutoModelForCausalLM.from_pretrained(
     phi4_model_path,
     device_map="auto",
-    load_in_4bit=True,              # Enable 4-bit quantization
-    quantization_config={
-        "bnb_4bit_compute_dtype": torch.float16,
-        "bnb_4bit_use_double_quant": True,
-        "bnb_4bit_quant_type": "nf4"
-    }
+    quantization_config=quantization_config
 )
 phi4_tokenizer = AutoTokenizer.from_pretrained(phi4_model_path)