Upload app.py

app.py

@@ -1,3 +1,4 @@
+
 import gradio as gr
 from transformers import AutoTokenizer, AutoModelForCausalLM
 from peft import PeftModel
@@ -8,7 +9,8 @@ LORA = "GilbertAkham/openlm-llama-lora-codetrans"   # your LoRA on HF
 tokenizer = AutoTokenizer.from_pretrained(BASE)
 model = AutoModelForCausalLM.from_pretrained(
     BASE,
-    load_in_8bit=True,          # fits on T4/low VRAM
+    load_in_8bit=True,
+    use_fast=False,          # fits on T4/low VRAM
     device_map="auto"
 )
 model = PeftModel.from_pretrained(model, LORA)
@@ -34,3 +36,4 @@ demo = gr.Interface(
 )
 
 demo.launch(share=True)   # share=True gives you a free public link
+