Discover Hailo models

Start with precompiled Hailo models that run out of the box—learn how to pick the right variant for your device.

Estimated read time: 3 minutes

Where the model comes from: Hailo Model Zoo

In the First inference guide, we specified the model by:

• model_name="yolov8n_coco--640x640_quant_hailort_multidevice_1"

• zoo_url="degirum/hailo"

• inference_host_address="@local"

• model_properties={"device_type": ["HAILORT/HAILO8", "HAILORT/HAILO8L"]}

That zoo_url points to the Hailo Model Zoo, a curated collection of models precompiled for HailoRT and tested to run on Hailo devices. We highlight this zoo because it lets you run examples immediately, without any conversion steps.

Any model compiled for Hailo can be used with the same pattern. The advanced guide shows how to bring your own model (BYOM).

Why we start with the Hailo Model Zoo

• Instant results: models are precompiled and follow predictable I/O, so examples “just work”

• Consistent setup: you focus on model_name, zoo_url, and device_type—everything else stays the same

• Easy to swap: change only the model_name to try a different model

Example models

# ImageNet classification model
model_spec = ModelSpec(
    model_name="yolov8s_silu_imagenet--224x224_quant_hailort_hailo8l_1",
    zoo_url="degirum/hailo",
    inference_host_address="@local",
    model_properties={"device_type": ["HAILORT/HAILO8", "HAILORT/HAILO8L"]},
)

# COCO object detection model
model_spec = ModelSpec(
    model_name="yolov8n_coco--640x640_quant_hailort_hailo8_1",
    zoo_url="degirum/hailo",
    inference_host_address="@local",
    model_properties={"device_type": ["HAILORT/HAILO8", "HAILORT/HAILO8L"]},
)

# Face detection model
model_spec = ModelSpec(
    model_name="yolov8n_relu6_face--640x640_quant_hailort_multidevice_1",
    zoo_url="degirum/hailo",
    inference_host_address="@local",
    model_properties={"device_type": ["HAILORT/HAILO8", "HAILORT/HAILO8L"]},
)

Hailo-8 vs. Hailo-8L: picking the right device type

Compatibility rule of thumb

A model compiled for Hailo-8L will run on Hailo-8, but not the other way around.

That's why our starter examples use 8L-compatible or multidevice builds. We also specify device_type as a list—so the same code runs on both boards.

• For examples: We prioritize plug-and-play. Your snippet should "just work," whether you use Hailo-8 or Hailo-8L.

• For deployments: A model compiled for Hailo-8L will run on Hailo-8, but won't be optimized. For best performance (e.g., FPS, latency, power), use a model compiled specifically for your hardware.

Example: picking a YOLOv8n variant

• Runs on both (portable): yolov8n_coco--640x640_quant_hailort_multidevice_1

  • Great for docs and demos—this model works on Hailo-8 and Hailo-8L with the same code.

• Optimized for Hailo-8: yolov8n_coco--640x640_quant_hailort_hailo8_1

  • Best for production on Hailo-8; this model does not support Hailo-8L.

# Portable demo (both Hailo-8 and Hailo-8L)
model_spec = ModelSpec(
    model_name="yolov8n_coco--640x640_quant_hailort_multidevice_1",
    zoo_url="degirum/hailo",
    inference_host_address="@local",
    model_properties={"device_type": ["HAILORT/HAILO8", "HAILORT/HAILO8L"]},
)

# Optimized for Hailo-8 (recommended for real deployments on Hailo-8)
model_spec = ModelSpec(
    model_name="yolov8n_coco--640x640_quant_hailort_hailo8_1",
    zoo_url="degirum/hailo",
    inference_host_address="@local",
    model_properties={"device_type": ["HAILORT/HAILO8"]},
)

Bring your own Hailo model (advanced)

When you’re ready, the advanced guide shows how to:

• Package your Hailo-compiled artifacts with a PySDK model.json that defines inputs, outputs, preprocessing, and postprocessing.

• Point PySDK to your private or local model zoo.

• Rescan the zoo as you add models.

• Use the same ModelSpec pattern—only the zoo_url and model_name change.

Bottom line: Start fast with the Hailo Model Zoo, then graduate to bring in your own Hailo models, all using the same stable PySDK workflow.