Launching the DeepSeek-R1 language model in a Yandex Compute Cloud GPU cluster

In this tutorial, you will create a GPU cluster with two VMs to run the DeepSeek-R1 language model.

To run a language model in a GPU cluster:

1. Get your cloud ready.
2. Create a GPU cluster with two VMs.
3. Test cluster state.
4. Run the language model.
5. Test the language model performance.

If you no longer need the resources you created, delete them.

Get your cloud readyGet your cloud ready

Sign up in Yandex Cloud and create a billing account:

1. Navigate to the management console and log in to Yandex Cloud or register a new account.
2. On the Yandex Cloud Billing page, make sure you have a billing account linked and it has the ACTIVE or TRIAL_ACTIVE status. If you do not have a billing account, create one and link a cloud to it.

If you have an active billing account, you can navigate to the cloud page to create or select a folder for your infrastructure to operate in.

Learn more about clouds and folders.

Required paid resourcesRequired paid resources

The infrastructure support costs include:

• Fee for continuously running VMs and disks (see Yandex Compute Cloud pricing).

Create a GPU cluster with two VMs .Create a GPU cluster with two VMs .

Create a GPU clusterCreate a GPU cluster

Add two VMs to the clusterAdd two VMs to the cluster

1. Create the first VM:

   Management console

   1. In the left-hand panel, select Virtual machines.

   2. Click Create virtual machine.

   3. Under Boot disk image, select the Ubuntu 20.04 LTS Secure Boot CUDA 12.2 public image.

   4. In the Availability zone field, select the ru-central1-d availability zone.

   5. Under Disks and file storages, select the SSD disk type and specify its size: 800 GB.

   6. Under Computing resources, navigate to the Custom tab and specify the platform, number of GPUs, and cluster:

      • Platform: AMD Epyc 9474F with Gen2.
      • GPU: 8.
      • GPU cluster: Select the previously created test-gpu-cluster cluster.

   7. Under Access, select SSH key and specify the VM access credentials:

      • In the Login field, enter a username, e.g., ubuntu. Do not use root or other names reserved for the OS purposes. To perform operations requiring root privileges, use the sudo command.

   8. Click Create VM.

2. Similarly, create the second VM.

Test cluster stateTest cluster state

Optionally, you can:

• Test cluster physical state.
• Run parallel jobs in the cluster.
• Test InfiniBand throughput.

Run the language modelRun the language model

1. Connect to both VMs over SSH.

2. Add the current user to the docker group:

   sudo groupadd docker
   sudo usermod -aG docker $USER
   newgrp docker
   

3. Pull the SGLang image to both VMs:

   docker pull lmsysorg/sglang:latest
   

4. Run this command on the first VM:

   docker run --gpus all --device=/dev/infiniband --ulimit memlock=-1 --ulimit stack=67108864 --shm-size 32g --network=host -v ~/.cache/huggingface:/root/.cache/huggingface --name sglang_multinode1 -e GLOO_SOCKET_IFNAME=eth0 -it --rm --ipc=host lmsysorg/sglang:latest python3 -m sglang.launch_server --model-path deepseek-ai/DeepSeek-R1 --tp 16 --nccl-init-addr <IP_address_1>:30000 --nnodes 2 --node-rank 0 --trust-remote-code --host 0.0.0.0 --port 30001 --disable-radix --max-prefill-tokens 126000
   

   Where IP_address_1 is the first VM internal IP address.

5. Run this command on the second VM:

   docker run --gpus all --device=/dev/infiniband --ulimit memlock=-1 --ulimit stack=67108864 --shm-size 32g --network=host -v ~/.cache/huggingface:/root/.cache/huggingface --name sglang_multinode2 -e GLOO_SOCKET_IFNAME=eth0 -it --rm --ipc=host lmsysorg/sglang:latest python3 -m sglang.launch_server --model-path deepseek-ai/DeepSeek-R1 --tp 16 --nccl-init-addr <IP_address_1>:30000 --nnodes 2 --node-rank 1 --trust-remote-code --host 0.0.0.0 --port 30001 --disable-radix --max-prefill-tokens 126000
   

   Where IP_address_1 is the first VM internal IP address.

6. Wait for the server to start:

   The server is fired up and ready to roll!
   

Test the language model performanceTest the language model performance

1. In a new session, connect to the first VM over SSH.

2. Install the openai package:

   sudo apt update
   sudo apt install python3-pip
   pip install openai
   

3. Create a test_model.py script with the following contents:

   import openai
   client = openai.Client(
      base_url="http://127.0.0.1:30001/v1", api_key="EMPTY")
   
   response = client.chat.completions.create(
      model="default",
      messages=[
         {"role": "system", "content": "You are a helpful AI assistant"},
         {"role": "user", "content": "List 3 countries and their capitals."},
      ],
      temperature=0.3,
      max_tokens=1024,
   )
   print(response.choices[0].message.content)
   

4. Run the script:

   python3 test_model.py
   

   Model response example:

   Here are three countries and their capitals:
   
   1. **France** - Paris
   2. **Japan** - Tokyo
   3. **Brazil** - Brasília
   
   Let me know if you'd like more examples! 😊
   

How to delete the resources you createdHow to delete the resources you created

To stop paying for the resources you created:

1. Delete the VM instances in Compute Cloud.
2. Delete the GPU cluster in Compute Cloud.