update endpoint helper files

handler.py

@@ -1,5 +1,5 @@
-from typing import Any, Dict
 import time
+from typing import Any, Dict
 
 import torch
 from peft import PeftConfig, PeftModel
@@ -13,6 +13,25 @@ from infer_local import (
 )
 
 
+DEFAULT_BASE_MODEL = "Qwen/Qwen2.5-Coder-0.5B-Instruct"
+
+
+def as_bool(value: Any) -> bool:
+    if isinstance(value, bool):
+        return value
+    if isinstance(value, str):
+        return value.strip().lower() in {"1", "true", "yes", "y", "on"}
+    return bool(value)
+
+
+def clamp_int(value: Any, default: int, minimum: int, maximum: int) -> int:
+    try:
+        parsed = int(value)
+    except (TypeError, ValueError):
+        parsed = default
+    return max(minimum, min(maximum, parsed))
+
+
 class EndpointHandler:
     def __init__(self, path: str = ""):
         self.path = path or "."
@@ -22,7 +41,7 @@ class EndpointHandler:
 
         if adapter_weights_present:
             peft_config = PeftConfig.from_pretrained(self.path)
-            base_model_name = peft_config.base_model_name_or_path
+            base_model_name = peft_config.base_model_name_or_path or DEFAULT_BASE_MODEL
             self.tokenizer = AutoTokenizer.from_pretrained(base_model_name)
             base_model = AutoModelForCausalLM.from_pretrained(
                 base_model_name,
@@ -62,8 +81,8 @@ class EndpointHandler:
             }
 
         parameters = data.get("parameters", {}) or {}
-        max_new_tokens = int(parameters.get("max_new_tokens", 320))
-        do_sample = bool(parameters.get("do_sample", False))
+        max_new_tokens = clamp_int(parameters.get("max_new_tokens"), 320, 1, 1024)
+        do_sample = as_bool(parameters.get("do_sample", False))
 
         prompt_text = build_instruction_prompt(user_prompt)
         inputs = self.tokenizer(prompt_text, return_tensors="pt").to(self.device)