リモート関数 - CoreWeave Docs

The examples on this page omit auth. To authenticate with your W&B API key, install cwsandbox[wandb] and pass auth=AuthStrategy.WANDB to Sandbox.run(), Sandbox(), or Sandbox.session(). See Get started.

This tutorial shows you how to run Python functions inside a sandbox using remote functions. You can call sandbox-executed code as if it were a local function instead of constructing shell command strings. It’s for developers who want to invoke compute in a sandbox directly from their Python code and collect typed results.

Define and call a remote function

Decorate a function with @session.function() to call it inside the sandbox with .remote(). The call returns a reference you resolve with .result():

from cwsandbox import Sandbox, SandboxDefaults

with Sandbox.session(SandboxDefaults()) as session:
    @session.function()
    def add(x: int, y: int) -> int:
        return x + y

    result = add.remote(2, 3).result()  # 5

Run functions in parallel with `.map()`

Use .map() to dispatch many invocations at once and collect their results. Independent calls run concurrently instead of one at a time:

@session.function()
def square(x: int) -> int:
    return x * x

refs = square.map([(1,), (2,), (3,)])
results = [r.result() for r in refs]  # [1, 4, 9]

Next steps

You can now invoke Python functions inside a sandbox as if they were local calls, and fan out work in parallel with .map(). For serialization modes, .local() testing, and more, see the Remote functions guide.

最終更新日 2026年9月30日

​Define and call a remote function

​Run functions in parallel with .map()

​Next steps

Define and call a remote function

Run functions in parallel with `.map()`

Next steps