Update app.py

app.py

@@ -10,15 +10,21 @@ from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStream
 # ---------------------
 # Model + Tokenizer
 # ---------------------
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
 MODEL_ID = "Ephraimmm/PIDGIN_gemma-3"
 
-bnb_config = BitsAndBytesConfig(
-    load_in_4bit=True,
-    bnb_4bit_quant_type="nf4",
-    bnb_4bit_use_double_quant=True,
-    bnb_4bit_compute_dtype=torch.bfloat16 if torch.cuda.is_bf16_supported() else torch.float16,
+print("Loading quantized model...")
+model = AutoModelForCausalLM.from_pretrained(
+    MODEL_ID,
+    device_map="auto",              # Let HF handle GPU placement
+    torch_dtype="auto",             # Match the quantization dtype
+    trust_remote_code=True          # Required for Unsloth models
 )
 
+tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, trust_remote_code=True)
+
+
 print("Loading model...")
 model = AutoModelForCausalLM.from_pretrained(
     MODEL_ID,