Chat model logging

blendsql/models/local/_transformers.py

@@ -1,6 +1,8 @@
 import importlib.util
 from typing import Optional
+from colorama import Fore
 
+from ..._logger import logger
 from .._model import LocalModel, ModelObj
 
 DEFAULT_KWARGS = {"do_sample": True, "temperature": 0.0, "top_p": 1.0}
@@ -47,6 +49,7 @@ def __init__(
         transformers.logging.set_verbosity_error()
         if config is None:
             config = {}
+
         super().__init__(
             model_name_or_path=model_name_or_path,
             requires_config=False,
@@ -61,12 +64,20 @@ def _load_model(self) -> ModelObj:
         from guidance.models import Transformers
         import torch
 
-        return Transformers(
+        lm = Transformers(
             self.model_name_or_path,
             echo=False,
             device_map="cuda" if torch.cuda.is_available() else "cpu",
             **self.load_model_kwargs,
         )
+        # Try to infer if we're in chat mode
+        if lm.engine.tokenizer._orig_tokenizer.chat_template is None:
+            logger.debug(
+                Fore.YELLOW
+                + "chat_template not found in tokenizer config.\nBlendSQL currently only works with chat models"
+                + Fore.RESET
+            )
+        return lm
 
 
 class TransformersVisionModel(TransformersLLM):