Compute — run Python on cloud GPUs

Compute

1 min read Original article ↗

Compute starts a fresh GPU machine for your Python function, streams logs to your terminal, and shuts the machine down when the run finishes.

Fine-tuning modelsReinforcement learningBatch inference

A run, end to end

$ compute run simple_mlp.py::train --gpu H100

  1. estimaterate locked · confirmation required
  2. provisionfresh H100-SXM requested
  3. ready80 GB HBM3 available
  4. stdoutstep 1/3 · loss 0.8421
  5. result{"status":"ok","artifact":"result.json"}
  6. terminatedmachine gone · receipt final

Provider usage$0.80

Platform fee$0.06

Total debit$0.86

Providers

Choose a GPU provider.

RunPod, Hot Aisle, and Vast.ai are available. See all GPUs.

  • RunPodSecure H100 capacity

  • Hot AisleMI300X capacity

  • Vast.aiSpot instances

Quickstart

Your first run, in four steps.

  1. 01

    Sign up

    Create a Compute account.

  2. 02

    Buy credits

    Add $10 or more in prepaid credit.

  3. 03

    Install the CLI

    Run the install command.

    curl -fsSL https://compute.cx/install.sh | sh
  4. 04

    Run a function

    Choose a Python entry point and GPU.

    compute run simple_mlp.py::train --gpu H100

Use cases

Choose a workload.

Each guide includes code and a command.

Pricing

Provider rate + 7.5%.

The provider rate is locked before launch. Compute adds a 7.5% fee. There are no subscriptions. If a machine never becomes ready, you pay $0.

See pricing
Get started

Run your first function.

Create an account and add credit, then install the CLI.