Quickstart

Running your AI workloads on Intel® Gaudi® accelerators can be accomplished in just a few simple steps. In this quick guide, we show how to run inference with GPT-2 model on Intel Gaudi 2 accelerators using the 🤗 Optimum for Intel Gaudi library.

Optimum for Intel Gaudi library is optimized for running various AI workloads on Intel Gaudi accelerators and it contains fully documented inference, training and fine-tuning examples. Please refer to the Optimum for Intel Gaudi GitHub page for more information.

Accessing Intel Gaudi AI Accelerator

To access an Intel Gaudi AI accelerator node in the Intel® Tiber™ AI Cloud, you will go to Intel Tiber AI Cloud and access the hardware instances to select the Intel Gaudi AI accelerator platform for deep learning and follow the steps to start and connect to the node.

Docker Setup

Now that you have access to the node, you will use the latest Intel Gaudi AI Accelerator docker image by executing the docker run command which will automatically download and run the docker. At the time of writing this guide, latest Gaudi docker version was 1.21.0:

release=1.21.0
os=ubuntu22.04
torch=2.6.0
docker_image=vault.habana.ai/gaudi-docker/$release/$os/habanalabs/pytorch-installer-$torch:latest

Execute docker run command:

docker run -itd \
    --name Gaudi_Docker \
    --runtime=habana \
    -e HABANA_VISIBLE_DEVICES=all \
    -e OMPI_MCA_btl_vader_single_copy_mechanism=none \
    --cap-add=sys_nice \
    --net=host \
    --ipc=host \
    ${docker_image}

Optimum for Intel Gaudi Setup

Check latest release of Optimum for Intel Gaudi here. At the time of writing this guide, latest Optimum for Intel Gaudi release version was v1.18.0, which is paired with Intel Gaudi Software release version 1.21.0. Install Optimum for Intel Gaudi as follows:

git clone -b v1.18.0 https://github.com/huggingface/optimum-habana
pip install ./optimum-habana

All available examples are under optimum-habana/examples.

Here is text-generation example, to run Llama-2 7B text generation example on Gaudi, complete the prerequisite setup:

cd ~/optimum-habana/examples/text-generation
pip install -r requirements.txt

To be able to run gated models like Llama-2 7B, you should:

• Have a 🤗 account
• Agree to the terms of use of the model in its model card
• Set your token as explained here
• Login to your account using the HF CLI: run huggingface-cli login before launching your script

Single Device Inference

Run single Gaudi device (HPU) inference with Llama-2 7B model:

PT_HPU_LAZY_MODE=1 python run_generation.py \
    --model_name_or_path meta-llama/Llama-2-7b-hf \
    --use_hpu_graphs \
    --use_kv_cache \
    --max_new_tokens 100 \
    --do_sample \
    --prompt "Here is my prompt"

Multi-Device Inference

With a multi-device Gaudi system, such as one with 8 HPUs, you can perform distributed inference using libraries like Microsoft® DeepSpeed. Gaudi-specific fork of the library is maintained by Intel at https://github.com/HabanaAI/DeepSpeed.

To install the library compatible with the same Gaudi software release stack, use:

pip install git+https://github.com/HabanaAI/DeepSpeed.git@1.21.0

With DeepSpeed successfully installed we can now run a distributed GPT-2 inference on an 8 HPU system as follows:

number_of_devices=8 \
PT_HPU_LAZY_MODE=1 python ../gaudi_spawn.py --use_deepspeed --world_size ${number_of_devices} \
run_generation.py \
    --model_name_or_path meta-llama/Llama-2-7b-hf \
    --use_hpu_graphs \
    --use_kv_cache \
    --max_new_tokens=100 \
    --do_sample \
    --prompt="Here is my prompt"

Training on Gaudi

🤗 Optimum for Intel Gaudi contains a number of examples demonstrating single and multi Gaudi device training/fine-tuning.

For example, a number of language models can be trained with the scripts provided language modeling examples section.

As an illustration, let us run GPT-2 single and multi card training examples on Gaudi.

Install prerequisites with:

cd ~/optimum-habana/examples/language-modeling
pip install -r requirements.txt

To train GPT-2 model on a single card, use:

PT_HPU_LAZY_MODE=1 python run_clm.py \
    --model_name_or_path gpt2 \
    --dataset_name wikitext \
    --dataset_config_name wikitext-2-raw-v1 \
    --per_device_train_batch_size 4 \
    --per_device_eval_batch_size 4 \
    --do_train \
    --do_eval \
    --output_dir /tmp/test-clm \
    --gaudi_config_name Habana/gpt2 \
    --use_habana \
    --use_lazy_mode \
    --use_hpu_graphs_for_inference \
    --throughput_warmup_steps 3

To train GPT-2 model using multi-card Gaudi system:

number_of_devices=8 \
PT_HPU_LAZY_MODE=1 python ../gaudi_spawn.py --use_deepspeed --world_size ${number_of_devices} \
run_clm.py \
    --model_name_or_path gpt2 \
    --dataset_name wikitext \
    --dataset_config_name wikitext-2-raw-v1 \
    --per_device_train_batch_size 4 \
    --per_device_eval_batch_size 4 \
    --do_train \
    --do_eval \
    --output_dir /tmp/test-clm \
    --gaudi_config_name Habana/gpt2 \
    --use_habana \
    --use_lazy_mode \
    --use_hpu_graphs_for_inference \
    --gradient_checkpointing \
    --use_cache False \
    --throughput_warmup_steps 3

Diffusion Workloads

🤗 Optimum for Intel Gaudi also features HPU-optimized support for the 🤗 Diffusers library. Thus, you can deploy Stable Diffusion and similar diffusion models on Gaudi and enable text-to-image generation and other diffusion-based workloads.

Before running Stable Diffusion inference example on Gaudi, complete the prerequisite setup:

cd ~/optimum-habana/examples/stable-diffusion
pip install -r requirements.txt

Here is an example of running Stable Diffusion text to image inference on Gaudi:

PT_HPU_LAZY_MODE=1 python text_to_image_generation.py \
    --model_name_or_path CompVis/stable-diffusion-v1-4 \
    --prompts "An image of a squirrel in Picasso style" \
    --num_images_per_prompt 10 \
    --batch_size 1 \
    --image_save_dir /tmp/stable_diffusion_images \
    --use_habana \
    --use_hpu_graphs \
    --gaudi_config Habana/stable-diffusion \
    --bf16

Also, here is an example of modifying a basic 🤗 Diffusers Stable Diffusion pipeline call to work with Gaudi using the Optimum for Intel Gaudi library:

- from diffusers import DDIMScheduler, StableDiffusionPipeline
+ from optimum.habana.diffusers import GaudiDDIMScheduler, GaudiStableDiffusionPipeline

model_name = "CompVis/stable-diffusion-v1-4"

- scheduler = DDIMScheduler.from_pretrained(model_name, subfolder="scheduler")
+ scheduler = GaudiDDIMScheduler.from_pretrained(model_name, subfolder="scheduler")

- pipeline = StableDiffusionPipeline.from_pretrained(
+ pipeline = GaudiStableDiffusionPipeline.from_pretrained(
    model_name,
    scheduler=scheduler,
+   use_habana=True,
+   use_hpu_graphs=True,
+   gaudi_config="Habana/stable-diffusion",
)

outputs = pipeline(
    ["An image of a squirrel in Picasso style"],
    num_images_per_prompt=16,
+   batch_size=4,
)

In addition, sample scripts for fine-tuning diffusion models are given in Stable Diffusion training section.

A more comprehensive list of examples in Optimum for Intel Gaudi is given next.

Ready-to-Use Examples

Now that you have run a full inference case, you can go back to the Optimum for Intel Gaudi validated models to see more options for running inference.

Here are examples for various modalities and tasks that can be used out of the box:

• Text

  • language modeling
  • multi node training
  • protein folding
  • question answering
  • sentence transformers training
  • summarization
  • table detection
  • text classification
  • text feature extraction
  • text generation
  • translation
  • trl

• Audio

  • audio classification
  • speech recognition
  • text to speech

• Images

  • object detection
  • object segementation
  • image classification
  • image to text
  • contrastive image text
  • stable diffusion
  • visual question answering
  • zero-shot object detection

• Video

  • stable-video-diffusion
  • video-classification

To learn more about how to adapt 🤗 Transformers or Diffusers scripts for Intel Gaudi, check out Script Adaptation guide.