diff --git a/model-lab/app/main.py b/model-lab/app/main.py
index d72f1a3..aa1c7be 100644
--- a/model-lab/app/main.py
+++ b/model-lab/app/main.py
@@ -1,12 +1,25 @@
 from fastapi import FastAPI, File, UploadFile
+from contextlib import asynccontextmanager
 from PIL import Image
 import io
-from app.model import predict_emotion
+from app.model import predict_emotion, load_model, is_model_ready
 
-app = FastAPI(title="Face Sentiment API")
+@asynccontextmanager
+async def lifespan(app: FastAPI):
+    load_model()
+    yield
+
+app = FastAPI(title="Face Sentiment API", lifespan=lifespan)
 
 @app.post("/predict")
 async def predict(file: UploadFile = File(...)):
     image = Image.open(io.BytesIO(await file.read()))
     emotion = predict_emotion(image)
     return {"emotion": emotion}
+
+@app.get("/health")
+def get_health():
+    if is_model_ready():
+        return {"status" : "ok"}
+    else:
+        return {"status" : "unavailable"}
diff --git a/model-lab/app/model.py b/model-lab/app/model.py
index b6f92c6..16243c6 100644
--- a/model-lab/app/model.py
+++ b/model-lab/app/model.py
@@ -1,11 +1,20 @@
+from typing import Optional
 from transformers import AutoFeatureExtractor, AutoModelForImageClassification
 from PIL import Image
 import torch
 
 MODEL_NAME = "trpakov/vit-face-expression"
 
-extractor = AutoFeatureExtractor.from_pretrained(MODEL_NAME)
-model = AutoModelForImageClassification.from_pretrained(MODEL_NAME)
+extractor = None
+model = None
+
+def load_model():
+    global extractor, model
+    extractor = AutoFeatureExtractor.from_pretrained(MODEL_NAME)
+    model = AutoModelForImageClassification.from_pretrained(MODEL_NAME)
+
+def is_model_ready() -> bool:
+    return extractor is not None and model is not None
 
 def predict_emotion(image: Image.Image):
     inputs = extractor(images=image, return_tensors="pt")