Forecaster

TinyTimeMixerForecaster

TinyTimeMixer Forecaster for Zero-Shot Forecasting of Multivariate Time Series.

Quickstart

python

from sktime.forecasting.ttm import TinyTimeMixerForecaster

estimator = TinyTimeMixerForecaster(model_path='ibm/TTM', revision='main', validation_split=0.2, config=None, training_args=None, compute_metrics=None, callbacks=None, broadcasting=False, use_source_package=False, fit_strategy='minimal')

Parameters(10)

model_pathstr, default=”ibm/TTM”

Path to the Huggingface model to use for forecasting. This can be either:

The name of a Huggingface repository (e.g., “ibm/TTM”)
A local path to a folder containing model files in a format supported by transformers. In this case, ensure that the directory contains all necessary files (e.g., configuration, tokenizer, and model weights).
If this parameter is None, fit_strategy should be full to allow full fine tuning of the model loaded from pretrained/provided config, else ValueError is raised.

revision: str, default=”main”

Revision of the model to use:

“main”: For loading model with context_length of 512 and prediction_length of 96.
“1024_96_v1”: For loading model with context_length of 1024 and prediction_length of 96.

This param becomes irrelevant when model_path is None

validation_splitfloat, default=0.2

Fraction of the data to use for validation

configdict, default={}

Configuration to use for the model. See the transformers documentation for details.

training_argsdict, default={}

Training arguments to use for the model. See transformers.TrainingArguments for details. Note that the output_dir argument is required.

compute_metricslist, default=None

List of metrics to compute during training. See transformers.Trainer for details.

callbackslist, default=[]

List of callbacks to use during training. See transformers.Trainer

broadcastingbool, default=False

if True, multiindex data input will be broadcasted to single series. For each single series, one copy of this forecaster will try to fit and predict on it. The broadcasting is happening inside automatically, from the outerside api perspective, the input and output are the same, only one multiindex output from predict.

use_source_packagebool, default=False

If True, the model and configuration will be loaded directly from the source package tsfm_public.models.tinytimemixer. This is useful if you want to bypass the local version of the package or when working in an environment where the latest updates from the source package are needed. If False, the model and configuration will be loaded from the local version of package maintained in sktime because of model’s unavailability on pypi. To install the source package, follow the instructions here [4].

fit_strategystr, default=”minimal”

Strategy to use for fitting (fine-tuning) the model. This can be one of the following: - “zero-shot”: Uses pre-trained model as it is. If model path is None

with this strategy, ValueError is raised.

“minimal”: Fine-tunes only a small subset of the model parameters, allowing for quick adaptation with limited computational resources. If model path is None with this strategy, ValueError is raised.
“full”: Fine-tunes all model parameters, which may result in better performance but requires more computational power and time. Allows model path to be None.

Examples

>>> from sktime.forecasting.ttm import TinyTimeMixerForecaster
>>> from sktime.datasets import load_airline
>>> y = load_airline ()
>>> forecaster = TinyTimeMixerForecaster ()
>>> # performs zero-shot forecasting, as default config (unchanged) is used
>>> forecaster. fit (y, fh = [1, 2, 3 ]) TinyTimeMixerForecaster(
... )
>>> y_pred = forecaster. predict ()
>>> from sktime.forecasting.ttm import TinyTimeMixerForecaster
>>> from sktime.datasets import load_tecator
>>> 
>>> # load multi-index dataset
>>> y = load_tecator (
... return_type = "pd-multiindex",
... return_X_y = False
... )
>>> y. drop (['class_val' ], axis = 1, inplace = True)
>>> 
>>> # global forecasting on multi-index dataset
>>> forecaster = TinyTimeMixerForecaster (
... model_path = None,
... fit_strategy = "full",
... config = {
... "context_length": 8,
... "prediction_length": 2
... },
... training_args = {
... "num_train_epochs": 1,
... "output_dir": "test_output",
... "per_device_train_batch_size": 32,
... },
... )
>>> 
>>> # model initialized with random weights due to None model_path
>>> # and trained with the full strategy.
>>> forecaster. fit (y, fh = [1, 2, 3 ]) TinyTimeMixerForecaster(
... )
>>> y_pred = forecaster. predict () Example with exogenous variables:
>>> from sktime.forecasting.ttm import TinyTimeMixerForecaster
>>> from sktime.datasets import load_longley
>>> from sktime.split import temporal_train_test_split
>>> y, X = load_longley ()
>>> y_train, _, X_train, X_future = temporal_train_test_split (y, X, test_size = 2)
>>> 
>>> # Initialize forecaster
>>> forecaster = TinyTimeMixerForecaster (
... model_path = None,
... fit_strategy = "full",
... config = {
... "context_length": 8,
... "prediction_length": 2
... },
... training_args = {
... "max_steps": 10,
... "output_dir": "test_output",
... "per_device_train_batch_size": 4,
... "report_to": "none",
... },
... )
>>> 
>>> # Fit with exogenous variables
>>> forecaster. fit (y_train, X = X_train, fh = [1, 2 ]) TinyTimeMixerForecaster(
... )
>>> 
>>> # Predict with exogenous variables
>>> y_pred = forecaster. predict (X = X_future)

References

[1 ] https://github.com/ibm-granite/granite-tsfm/tree/main/tsfm_public/models/tinytimemixer [2 ] Ekambaram, V., Jati, A., Dayama, P., Mukherjee, S., Nguyen, N.H., Gifford, W.M., Reddy, C. and Kalagnanam, J., 2024. Tiny Time Mixers (TTMs): Fast Pre-trained Models for Enhanced Zero/Few-Shot Forecasting of Multivariate Time Series. CoRR. [3 ] https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/tutorial/ttm_tutorial.ipynb [4 ] https://github.com/ibm-granite/granite-tsfm/tree/ttm