updated model.py

model.py

@@ -126,7 +126,7 @@
 from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
 from functools import lru_cache
 import logging
-
+import asyncio
 logger = logging.getLogger(__name__)
 _model_loaded = False
 _tokenizer = None
@@ -157,3 +157,37 @@ def load_model_sync():
     except Exception as e:
         logger.error(f"❌ Failed to load model: {e}")
         raise
+
+
+async def load_model_async():
+    global _tokenizer, _model, _model_loaded
+    if _model_loaded:
+        return
+
+    config = get_model_config()
+    model_id = config["model_id"]
+
+    try:
+        _tokenizer = AutoTokenizer.from_pretrained(model_id)
+        _model = AutoModelForSeq2SeqLM.from_pretrained(model_id)
+        _model.eval()
+        _model_loaded = True
+        logger.info(f"✅ Model {model_id} loaded successfully.")
+    except Exception as e:
+        logger.error(f"❌ Failed to load model: {e}")
+        raise
+
+def get_model():
+    if not _model_loaded:
+        raise ValueError("Model not loaded yet")
+    return _tokenizer, _model
+
+def is_model_loaded():
+    return _model_loaded
+
+def get_model_info():
+    return {
+        "model_id": get_model_config()["model_id"],
+        "loaded": _model_loaded,
+        "loading": not _model_loaded
+    }