ConvTimeNetClassifierSKTimeWrapper
Python template · sinapsis_sktime.templates.sktime_classifiers
Overview
Template for SkTime classification model training and inference for the deep_learning module.
ConvTimeNetClassifier — inherited documentation
ConvTimeNet for time series classification.
ConvTimeNet is a hierarchical pure convolutional model designed. Unlike prevalent methods centered around self-attention mechanisms, ConvTimeNet introduces two key innovations:
1. A deformable patch layer that adaptively perceives local patterns of temporally dependent basic units in a data-driven manner. 2. Hierarchical pure convolutional blocks that capture dependency relationships among the representations of basic units at different scales.
The model employs a large kernel mechanism allowing convolutional blocks to be deeply stacked, achieving a larger receptive field. This architecture effectively models both local patterns and their multi-scale dependencies within a single model, addressing common challenges in time series analysis such as adaptive perception of local patterns and multi-scale dependency capture.
This classifier has been wrapped around implementations from [1]_, [2]_ and [3]_.
Parameters ---------- d_model : int Hidden dimension size for model processing. patch_size : int Size of patches for sequence splitting. patch_stride : int Stride length for patch creation. dropout : float, optional (default=0) Dropout rate to apply to layers. d_ff : int, optional (default=128) Dimension of feedforward network. dw_ks : int or list, optional (default=3) Depthwise convolution kernel size(s). Can be a single int or list of ints. device : str, optional (default="cpu") Device to use for computation ("cpu" or "cuda"). num_epochs : int, optional (default=16) The number of epochs to train the model. batch_size : int, optional (default=8) The size of each mini-batch during training. criterion : callable, optional (default=None) The loss function to use. If None, CrossEntropyLoss will be used. criterion_kwargs : dict, optional (default=None) Additional keyword arguments to pass to the loss function. optimizer : str, optional (default=None) The optimizer to use. If None, Adam will be used. optimizer_kwargs : dict, optional (default=None) Additional keyword arguments to pass to the optimizer. lr : float, optional (default=0.001) The learning rate to use for the optimizer. verbose : bool, optional (default=False) Whether to print progress information during training. random_state : int, optional (default=None) Seed to ensure reproducibility.
Examples -------- >>> from sktime.classification.deep_learning import ConvTimeNetClassifier >>> import numpy as np >>> # Create a sample multivariate time series dataset >>> # 48 samples, 3 variables, length 128 >>> X = np.random.randn(16 * 3, 3, 128).astype("float32") >>> y = np.array([0, 1, 2] * 16) # 3 classes >>> # Create and fit the classifier >>> clf = ConvTimeNetClassifier( ... patch_size=4, ... patch_stride=2, ... d_model=64, ... d_ff=128, ... dw_ks=[5, 7, 9], ... batch_size=8, ... device="cpu", ... random_state=10 ... ) # doctest: +SKIP >>> clf.fit(X, y) # doctest: +SKIP ConvTimeNetClassifier(...) >>> # Make predictions >>> y_pred = clf.predict(X) # doctest: +SKIP >>> y_proba = clf.predict_proba(X) # doctest: +SKIP
References ---------- .. [1] Cheng, M., Yang, J., Pan, T., Liu, Q., & Li, Z. (2024). ConvTimeNet: A deep hierarchical fully convolutional model for multivariate time series analysis. arXiv preprint arXiv:2403.01493. https://arxiv.org/abs/2403.01493 .. [2] https://github.com/Mingyue-Cheng/ConvTimeNet
Capabilities
Capabilities: not setInstallation
Install the published package using pip or uv.
pip install sinapsis-sktime==1.0.0uv pip install sinapsis-sktime==1.0.0uv add sinapsis-sktime==1.0.0Configuration
Effective fields from the inspector schema, including nested Pydantic model fields.
Optional
Default: null
Optional
Default: "/sinapsis-hub/private/platform/sinapsis-hub/playground/publish"
Path where the trained model will be saved. convtimenetclassifier_init d_model typing.Optional[int]: default: <class 'inspect._empty'> patch_size typing.Optional[int]: default: <class 'inspect._empty'> patch_stride… Read full field description
Path where the trained model will be saved. convtimenetclassifier_init d_model typing.Optional[int]: default: <class 'inspect._empty'> patch_size typing.Optional[int]: default: <class 'inspect._empty'> patch_stride typing.Optional[int]: default: <class 'inspect._empty'> dropout typing.Optional[float]: default: 0 d_ff <class 'int'>: default: 128 dw_ks <class 'int'>: default: 3 device <class 'str'>: default: cpu num_epochs <class 'int'>: default: 16 batch_size <class 'int'>: default: 8 criterion <class 'NoneType'>: default: None criterion_kwargs typing.Optional[dict]: default: None optimizer typing.Optional[str]: default: None optimizer_kwargs typing.Optional[dict]: default: None lr <class 'float'>: default: 0.001 verbose <class 'bool'>: default: False random_state typing.Optional[int]: default: None
Required
Default: null
Required
Default: null
Optional
Default: 0.001
Optional
Default: 128
Optional
Default: 3
Optional
Default: "cpu"
Optional
Default: null
Optional
Default: 0
Optional
Default: false
Optional
Default: null
Optional
Default: null
Optional
Default: 8
Optional
Default: 16
Optional
Default: null
Optional
Default: null
Optional
Default: null
Optional
Default: null
Optional
Default: null
Configuration example
Generated from documented defaults. Replace any placeholders and complete required nested settings before running.
templates:
- template_name: "ConvTimeNetClassifierSKTimeWrapper"
class_name: "ConvTimeNetClassifierSKTimeWrapper"
attributes:
root_dir: "/sinapsis-hub/private/platform/sinapsis-hub/playground/publish"
model_save_path: "<required>"
convtimenetclassifier_init:
lr: 0.001
d_ff: 128
dw_ks: 3
device: "cpu"
d_model: null
dropout: 0
verbose: false
criterion: null
optimizer: null
batch_size: 8
num_epochs: 16
patch_size: null
patch_stride: null
random_state: null
criterion_kwargs: null
optimizer_kwargs: null
Inheritance
sinapsis_sktime.templates.sktime_classifiers.SKTimeDeepLearningClassifiersEnvironment variables
Environment definitions belong to their canonical Python modules.