Updated.

app.py

@@ -23,10 +23,18 @@ embedding_model = SentenceTransformer("all-MiniLM-L6-v2")
 model_id = "BioMistral/BioMistral-7B"
 
 tokenizer = AutoTokenizer.from_pretrained(model_id)
+bnb_config = BitsAndBytesConfig(
+    load_in_4bit=True,
+    bnb_4bit_quant_type="nf4",
+    bnb_4bit_use_double_quant=True,
+    bnb_4bit_compute_dtype=torch.float32,
+)
+
 generation_model = AutoModelForCausalLM.from_pretrained(
     model_id,
+    device_map={"": "cpu"},
+    quantization_config=bnb_config,
     torch_dtype=torch.float32,
-    device_map={"": "cpu"}
 )
 
 # ----------------------