Update app.py

app.py

@@ -1,10 +1,23 @@
 import gradio as gr
 from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
 
-MODEL_ID = "janajankovic/autotrain-juhh6-uwiv9"  # your fine-tuned model repo id
+import torch
+from transformers import AutoTokenizer, pipeline
+from peft import AutoPeftModelForCausalLM  # comes from `peft`
 
-tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
-model = AutoModelForCausalLM.from_pretrained(MODEL_ID)
+
+
+BASE_MODEL = "cjvt/GaMS-1B-Chat"
+ADAPTER_ID = "janajankovic/autotrain-juhh6-uwiv9"
+
+# Tokenizer from the base model
+tokenizer = AutoTokenizer.from_pretrained(BASE_MODEL)
+
+# Load base+adapter via AutoPeftModel
+model = AutoPeftModelForCausalLM.from_pretrained(
+    ADAPTER_ID,
+    torch_dtype=torch.float32,  # CPU in this Space
+)
 
 pipe = pipeline(
     "text-generation",