Models

sleepKIT provides a number of model architectures that can be used for training sleep-monitoring tasks. While a number of off-the-shelf models exist, they are often not efficient nor optimized for real-time, edge applications. To address this, sleepKIT provides a model factory that allows you to easily create and train customized models via helia-edge. helia-edge includes a growing number of state-of-the-art models that can be easily configured and trained using high-level parameters. The models are designed to be efficient and well-suited for real-time, edge applications. Most of the models are based on state-of-the-art architectures that have been modified to allow for more fine-grain customization. In addition, the models support 1D variants to allow for training on time-series data. Please check helia-edge for list of available models and their configurations.

---

Available Models

• TCN: A CNN leveraging dilated convolutions (key=tcn)
• U-Net: A CNN with encoder-decoder architecture for segmentation tasks (key=unet)
• U-NeXt: Registered as an alias for the U-Net builder (key=unext)
• EfficientNetV2: A CNN leveraging MBConv blocks (key=efficientnetv2)
• MobileOne: A CNN aimed at sub-1ms inference (key=mobileone)
• ResNet: A popular CNN often used for vision tasks (key=resnet)
• Composer: Compose an architecture through the registered composer builder.

• Bring-Your-Own-Model: Add a custom model architecture to sleepKIT.

---

Model Factory

sleepKIT includes a model factory, sk.ModelFactory, that eases the processes of creating models for training. The factory allows you to create models by specifying the model key and the model parameters. The factory will then create the model using the specified parameters. The factory also allows you to register custom models that can be used for training. By leveraring a factory, a task only needs to provide the architecture key and the parameters, and the factory will take care of the rest.

The model factory provides the following methods:

• sk.ModelFactory.register: Register a custom model
• sk.ModelFactory.unregister: Unregister a custom model
• sk.ModelFactory.has: Check if a model is registered
• sk.ModelFactory.get: Get a model from the factory
• sk.ModelFactory.list: List all available models

---

Usage

Defining a model in configuration file

A model can be created when invoking a command via the CLI by setting architecture in the configuration file. The task will use the supplied name to get the registered model and instantiate with the provided parameters.

Merge this architecture fragment into your task configuration. It does not include dataset, feature or training settings.

Configuration fragment

{
    "architecture": {
        "name": "tcn",
        "params": {
            "input_kernel": [1, 3],
            "input_norm": "batch",
            "blocks": [
                {"depth": 1, "branch": 1, "filters": 12, "kernel": [1, 3], "dilation": [1, 1], "dropout": 0.10, "ex_ratio": 1, "se_ratio": 0, "norm": "batch"},
                {"depth": 1, "branch": 1, "filters": 20, "kernel": [1, 3], "dilation": [1, 1], "dropout": 0.10, "ex_ratio": 1, "se_ratio": 2, "norm": "batch"},
                {"depth": 1, "branch": 1, "filters": 28, "kernel": [1, 3], "dilation": [1, 2], "dropout": 0.10, "ex_ratio": 1, "se_ratio": 2, "norm": "batch"},
                {"depth": 1, "branch": 1, "filters": 36, "kernel": [1, 3], "dilation": [1, 4], "dropout": 0.10, "ex_ratio": 1, "se_ratio": 2, "norm": "batch"},
                {"depth": 1, "branch": 1, "filters": 40, "kernel": [1, 3], "dilation": [1, 8], "dropout": 0.10, "ex_ratio": 1, "se_ratio": 2, "norm": "batch"}
            ],
            "output_kernel": [1, 3],
            "include_top": true,
            "use_logits": true,
            "model_name": "tcn"
        }
    }
}

After completing the task configuration, train with:

Terminal

sleepkit --mode train --task stage --config configuration.json

Defining a model in code

Alternatively, the model can be created directly in code using the following snippet:

Python example

import keras
import sleepkit as sk

architecture = {
    "name": "tcn",
    "params": {
        "input_kernel": [1, 3],
        "input_norm": "batch",
        "blocks": [
            {"depth": 1, "branch": 1, "filters": 12, "kernel": [1, 3], "dilation": [1, 1], "dropout": 0.10, "ex_ratio": 1, "se_ratio": 0, "norm": "batch"},
            {"depth": 1, "branch": 1, "filters": 20, "kernel": [1, 3], "dilation": [1, 1], "dropout": 0.10, "ex_ratio": 1, "se_ratio": 2, "norm": "batch"},
            {"depth": 1, "branch": 1, "filters": 28, "kernel": [1, 3], "dilation": [1, 2], "dropout": 0.10, "ex_ratio": 1, "se_ratio": 2, "norm": "batch"},
            {"depth": 1, "branch": 1, "filters": 36, "kernel": [1, 3], "dilation": [1, 4], "dropout": 0.10, "ex_ratio": 1, "se_ratio": 2, "norm": "batch"},
            {"depth": 1, "branch": 1, "filters": 40, "kernel": [1, 3], "dilation": [1, 8], "dropout": 0.10, "ex_ratio": 1, "se_ratio": 2, "norm": "batch"}
        ],
        "output_kernel": [1, 3],
        "include_top": True,
        "use_logits": True,
        "model_name": "tcn"
    }
}

inputs = keras.Input(shape=(256,1), dtype="float32")
num_classes = 5

model = sk.ModelFactory.get(architecture["name"])(
    inputs=inputs,
    params=architecture["params"],
    num_classes=num_classes,
)

model.summary()

---