made few changes

app.py

@@ -3,6 +3,7 @@ from fastapi.responses import HTMLResponse, RedirectResponse, JSONResponse
 from pydantic import BaseModel
 from typing import List
 import os                                    # ← add this
+
 from clearml import Model, Task
 import torch
 from configs import add_args
@@ -11,6 +12,9 @@ import argparse
 from argparse import Namespace
 from peft import PeftModel, PeftConfig, get_peft_model, LoraConfig
 
+# maximum token length for inputs
+MAX_SOURCE_LENGTH = 512
+
 # Load endpoints & creds
 CLEARML_API_HOST     = os.environ["CLEARML_API_HOST"]
 CLEARML_WEB_HOST     = os.environ["CLEARML_WEB_HOST"]
@@ -66,8 +70,7 @@ print("Base model architecture and tokenizer loaded.")
 # Download the fine-tuned weights via ClearML using your injected creds
 task = Task.get_task(task_id="9cc33fb4d1d54378b691188c5e230253")
 finetuned_weights_path = task.artifacts["lora-pytorch-bin"].get_local_copy()
-adapter_dir = os.path.dirname(finetuned_weights_path)
-print(f"Fine-tuned adapter weights downloaded to directory: {adapter_dir}")
+print(f"Fine-tuned adapter weights downloaded to directory: {os.path.dirname(finetuned_weights_path)}")
 
 # Create LoRA configuration matching the fine-tuned checkpoint
 lora_cfg = LoraConfig(