Skip to content

Running Hugging Face Transformers and Diffusers on an NVIDIA GH200 instance#

Hugging Face provides several powerful Python libraries that provide easy access to a wide range of pre-trained models. Among the most popular are Diffusers, which focuses on diffusion-based generative AI, and Transformers, which supports common AI/ML tasks across several different modalities. This tutorial demonstrates how to use these libraries to generate images and chatbot-style responses on an On-Demand Cloud (ODC) instance backed with the NVIDIA GH200 Grace Hopper Superchip.

Setting up your environment#

Launch your GH200 instance#

Begin by launching a GH200 instance:

  1. In the Lambda Cloud console, navigate to the SSH keys page, click Add SSH Key, and then add or generate a SSH key.
  2. Navigate to the Instances page and click Launch Instance.
  3. Follow the steps in the instance launch wizard.
    • Instance type: Select 1x GH200 (96 GB).
    • Region: Select an available region.
    • Filesystem: Don't attach a filesystem.
    • SSH key: Use the key you created in step 1.
  4. Click Launch instance.
  5. Review the EULAs. If you agree to them, click I agree to the above to start launching your new instance. Instances can take up to five minutes to fully launch.

Set up your Python virtual environment#

Next, create a new Python virtual environment and install the required libraries:

  1. In the Lambda Cloud console, navigate to the Instances page, find the row for your instance, and then click Launch in the Cloud IDE column. JupyterHub opens in a new window.
  2. In JupyterHub's Launcher tab, under Other, click Terminal to open a new terminal.

  3. In your terminal, create a Python virtual environment:

    python -m venv --system-site-packages hf-tests

  4. Activate the virtual environment:

    source hf-tests/bin/activate

  5. Install the Hugging Face Transformers library, Diffusers library, and other dependencies:

    pip install transformers diffusers["torch"] tf-keras==2.17.0 accelerate

Using Hugging Face Transformers and Diffusers#

Now that you've set up your environment, you can create and run Python programs based on Hugging Face Transformers and Diffusers. This section provides a few example programs to get you started.

Generate a chatbot response with the Transformers library#

To generate a chatbot-style response with the Hugging Face Transformers library:

  1. Open a new Python file named test_transformers.py for editing:

    nano test_transformers.py

  2. Paste the following Hugging Face Transformers test script into the file:

    import transformers
import torch

model_id = "facebook/opt-1.3b"

pipeline = transformers.pipeline(
    "text-generation", model=model_id, model_kwargs={"torch_dtype": torch.bfloat16}, device_map="auto"
)

output = pipeline("How is ice cream made?")
print(output)

  3. Save and exit.

  4. Run the script:

    python test_transformers.py

    You should get a result similar to the following:

    [{'generated_text': 'How is ice cream made?\n\nIce cream is made by mixing milk, sugar, and'}]

To learn more about how to use the Transformers library, see the Transformers section in the Hugging Face docs.

Generate an image with the Diffusers library#

To generate a prompt-based image with the Hugging Face Diffusers library:

  1. Open a new Python file named test_diffusers.py for editing:

    nano test_diffusers.py

  2. Paste the following Hugging Face Diffusers test script into the file. Feel free to change the prompt if desired:

    from diffusers import DiffusionPipeline
import torch

pipeline = DiffusionPipeline.from_pretrained("stable-diffusion-v1-5/stable-diffusion-v1-5", torch_dtype=torch.float16)
pipeline.to("cuda")
image = pipeline("An image of an elephant in the style of Matisse").images[0]
image.save("elephant_matisse.png")

  3. Save and exit.

  4. Run the script:

    python test_diffusers.py

    The resulting image file appears in JupyterHub's left nav. Double-click it to view the image:

    Generated image of an elephant in the style of Matisse

To learn more about how to use the Diffusers library, see the Diffusers section in the Hugging Face docs.

Cleaning up#

When you're done with your instance, terminate it to avoid incurring unnecessary costs:

  1. In the Lambda Cloud console, navigate to the Instances page.
  2. Select the checkboxes of the instances you want to delete.
  3. Click Terminate. A dialog appears.
  4. Follow the instructions and then click Terminate instances to terminate your instances.

Next steps#