Serve with FastAPI
Wrap inference in a REST endpoint.
Serve with FastAPI is a free Deep Learning Academy lesson on CoddyKit — lesson 4 of 4. You can read the complete lesson below for free — then practise it hands-on in the browser with a built-in code editor and a 24/7 AI tutor. It is part of the Deep Learning Academy learning path, one of 4 lessons in the course, and your progress syncs across the web and the CoddyKit app.
From Saved Model to Live API
A trained model only helps users when it is reachable. FastAPI wraps your model in a web endpoint they can call over HTTP. 🌍
Why FastAPI Fits Inference
FastAPI is fast, async-friendly, and auto-generates docs. That makes it a clean way to serve model predictions as a REST service.
Load the Model Once at Startup
Load your model a single time when the server boots, not on every request. Loading per call would make each prediction painfully slow.
model = torch.jit.load('model.pt')
model.eval()Create the App
You start by creating a FastAPI instance. This object holds your routes and becomes the server you run.
from fastapi import FastAPI
app = FastAPI()Validate Input with Pydantic
Define a Pydantic model for the request body so FastAPI checks the shape and types of incoming data automatically.
from pydantic import BaseModel
class Item(BaseModel):
features: list[float]Define a Predict Route
A POST route receives the validated data, runs the model, and returns the result as JSON the client can read.
@app.post('/predict')
def predict(item: Item):
...Run Inference Without Gradients
Wrap the forward pass in torch.no_grad() so the server skips gradient tracking and saves time and memory on every call.
with torch.no_grad():
output = model(x)Return a Clean JSON Response
Convert the tensor output to plain Python numbers before returning, since raw tensors are not directly JSON serializable.
return {'prediction': output.argmax().item()}Serve It with Uvicorn
Uvicorn is the server that runs your FastAPI app. One command brings your prediction endpoint online. 🚀
uvicorn main:app --host 0.0.0.0 --port 8000Explore the Auto Docs
FastAPI builds interactive docs at the /docs path, so you and your clients can test the endpoint right in the browser.
Add a Health Check Route
A tiny health endpoint lets load balancers confirm the server is alive, a small touch that makes deployment far more reliable.
@app.get('/health')
def health():
return {'status': 'ok'}Quick Check
You want each request validated for correct fields and types automatically. What does that?
Recap: Your Model Is Live
You served a model with FastAPI: load once at startup, validate with Pydantic, predict under no_grad, and run it on Uvicorn. 🎉
Frequently asked questions
Is the “Serve with FastAPI” lesson free?
Yes — the full text of “Serve with FastAPI” is free to read here on the web, and the Deep Learning Academy course includes 4 lessons in total. To practise it interactively (a built-in code editor and a 24/7 AI tutor) and unlock the rest of the Deep Learning Academy course, upgrade to CoddyKit PRO.
What will I learn in “Serve with FastAPI”?
Wrap inference in a REST endpoint. You practise Deep Learning Academy with hands-on code you run directly in the browser, and a 24/7 AI tutor answers your questions as you work through the lesson.
Do I need any experience to start Deep Learning Academy?
No prior experience is required. Deep Learning Academy on CoddyKit is structured for beginners through advanced learners; this is — lesson 4 of 4, so you can start here or from the beginning and move at your own pace.
How long does the “Serve with FastAPI” lesson take?
Most CoddyKit lessons take about 5–10 minutes. Each one is bite-sized and interactive, so you make steady progress and pick up exactly where you left off across the web and the app.
Can I write and run code in this Deep Learning Academy lesson?
Yes. Every Deep Learning Academy lesson includes a built-in code editor, so you write and run real code right in your browser and get instant AI feedback — no local setup required.