Skip to content
AMD AMD Radeon Cloud Docs

Quickstart

This page covers two separate tasks: calling a model in Token Factory and launching a GPU instance. Neither depends on the other, so start with whichever you need.

Token Factory is on the China site only. Instances are available on both sites; the examples use the China site’s address.

Call a model in Token Factory

Section titled “Call a model in Token Factory”

Calling a model in Token Factory needs no instance and spends no credits.

  1. Open Token Factory and sign in. See Sign in.

  2. On the Your API key card at the top, click Copy next to the API key.

  3. Send a request:

    Terminal window
    export RADEON_API_KEY=rc-...
    

    curl https://developer.amd.com.cn/radeon/api/v1/chat/completions \
      -H "Authorization: Bearer $RADEON_API_KEY" \
      -H "Content-Type: application/json" \
      -d '{"model":"DeepSeek-V4-Flash","messages":[{"role":"user","content":"Say hello in one sentence."}]}'

The same key works for every model in Token Factory. To use another one, change model. To list the models:

Terminal window
curl https://developer.amd.com.cn/radeon/api/v1/models \
  -H "Authorization: Bearer $RADEON_API_KEY"

From Python, point the OpenAI SDK at the same base URL:

from openai import OpenAI


client = OpenAI(base_url="https://developer.amd.com.cn/radeon/api/v1", api_key="rc-...")
response = client.chat.completions.create(
    model="DeepSeek-V4-Flash",
    messages=[{"role": "user", "content": "Say hello in one sentence."}],
)
print(response.choices[0].message.content)

Next: Token Factory for limits and the model list, or the API reference.

Launch a GPU instance

Section titled “Launch a GPU instance”

An instance spends credits every hour it runs, so check your balance first. See Credits.

From the console

Section titled “From the console”
  1. Sign in to the console for your site. See Sign in.
  2. On Home, keep Deploy type at Notebook and Image at AMD OneClick Base.
  3. To keep your files after the instance is gone, turn on Customize and set Workspace storage to Persistent /workspace. See Keep your data.
  4. Click Launch instance. When a node is free, the instance is usually ready within a minute.
  5. When the dialog reads Your instance is ready, click Open notebook to open JupyterLab, or connect over SSH.

If the launch fails, see Launch an instance.

From a terminal

Section titled “From a terminal”

Install Radeon Cloud CLI, sign in, check which resources your account can use, and launch:

Terminal window
python3 -m pip install amd-oneclick-sdk
rcc auth login
rcc catalog list
rcc instance add-key
rcc instance create
rcc instance open

In an interactive terminal, rcc instance create asks you to choose an environment and resources.

Check the GPU

Section titled “Check the GPU”

In a terminal inside the instance:

Terminal window
rocm-smi

It lists the Radeon GPUs attached to the instance.