Generate text before starting past api

main.py

@@ -22,6 +22,7 @@ from transformers import AutoTokenizer, AutoModelForCausalLM
 
 import transformers
 import torch
+import logging
 
 model_id = "deepcogito/cogito-v1-preview-llama-3B"
 
@@ -33,6 +34,23 @@ pipeline = transformers.pipeline(
 )
 
 
+print("Pipeline loaded")
+logging.info("Pipeline loaded")
+
+messages = [
+    {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
+    {"role": "user", "content": "Give me a short introduction to LLMs."},
+]
+
+outputs = pipeline(
+    messages,
+    max_new_tokens=512,
+)
+
+logging.info("Generated text")
+print(outputs[0]["generated_text"][-1])
+
+
 app = FastAPI()
 
 class EchoMessage(BaseModel):
@@ -54,11 +72,12 @@ async def generate_text(item: Item):
         max_new_tokens=512,
     )
 
-    print("request got")
+    logging.info("request got")
 
     resp = outputs[0]["generated_text"][-1]
 
-    print("Response generated", resp)
+    logging.info("Response generated")
+
     return {"response": resp}
 
 @app.get("/")