AI and ML models on local machine is like a hell. We know that and we motivated to create a bridge between local and cloud.

Prerequisite You should have magicpoint.yaml to use bridge!

Let’s continue with an example. If you have an magicpoint application like that;

example_app.py
from magicpoint import Magicpoint, Request, Response
from diffusers import StableDiffusionPipeline 
import torch
import os

app = Magicpoint("MyCoolApp 😎")

@app.init
def init():
    model = StableDiffusionPipeline.from_pretrained(
        "runwayml/stable-diffusion-v1-5",
        revision="fp16",
        torch_dtype=torch.float16,
        cache_dir="./models",
        # Add your own auth token from Huggingface
        use_auth_token=os.environ["HUGGINGFACE_API_KEY"],
    ).to("cuda")
    context = {
        "model": model
    }
    return context

@app.background_task("/generate_async")
def generate_image_async(context, request):
    # do smth
    return 


@app.training("/background")
def training(context : dict, request : Request) -> None:
    # do smth
    return


app.serve()

If you are working on training method you can run following commands to serve on cloud:

magicpoint bridge example_app.py:training

Output will shown like that;

[✓] Creating bridge...
[✓] Deploying magicpoint magic into the bridge...
[✓] Setup completed 
    Your bridge URL is ready: https://each.tech/bridge/eliminate-retirement-x1723xn
Deploy your modelAutoscaling and Concurrency