TimeVQVAE

This is an official Github repository for the PyTorch implementation of TimeVQVAE from our paper "Vector Quantized Time Series Generation with a Bidirectional Prior Model", AISTATS 2023.

TimeVQVAE is a robust time series generation model that utilizes vector quantization for data compression into the discrete latent space (stage1) and a bidirectional transformer for the prior learning (stage2).

Install / Environment setup

You should first create a virtual environment, and activate the environment. Then you can install the necessary libraries by running the following command.

$ pip install -r requirements.txt

Dataset Download

The UCR archive datasets are automatically downloaded if you run any of the training command below such as $ python stage1.py. If you just want to download the datasets only without running the training, run

$ python preprocessing/preprocess_ucr.py

Usage

Configuration

• configs/config.yaml: configuration for dataset, data loading, optimizer, and models (i.e., encoder, decoder, vector-quantizer, and MaskGIT)
• config/sconfig_cas.yaml: configuration for running CAS, Classification Accuracy Score (= TSTR, Training on Synthetic and Test on Real).

Run

:rocket: The stage 1 and stage 2 training can be performed with the following command:

$ python stage1.py --dataset_names CBF --gpu_device_idx 0

$ python stage2.py --dataset_names CBF --gpu_device_idx 0

Note that you need to specify a dataset of your interest in configs/config.yaml.

:bulb: The training pipeline is as follows:

• Run stage1.py and it saves trained encoders, decoders, and vector-quantizers for LF and HF.
• Run stage2.py and it saves the prior model (i.e., bidirectional transformer).
  • stage2.py includes an evaluation step which is performed right after the stage 2 training. The evaluation includes a visualization plot of test samples (from a test set) versus generated samples, FID score, and IS (Inception Score).

:rocket: If you want to run stage 1 and stage 2 at the same time, use the following command. You can specify dataset(s) and a GPU device in the command line for stages12_all_ucr.py.

$ python stage12_all_ucr.py --dataset_names CBF BME --gpu_device_idx 0

:rocket: CAS can be performed with the following command:

$ python run_CAS.py  --dataset_names CBF BME --gpu_device_idx 0

Google Colab

(NB! make sure to change your notebook setting to GPU.)

A Google Colab notebook is available for time series generation with the pretrained VQVAE. The usage is simple:

1. User Settings: specify dataset_name and n_samples_to_generate.
2. Sampling: Run the unconditional sampling and class-conditional sampling.

Note that the pretrained models are automatically downloaded within the notebook. In case you're interested, the pretrained models are stored in here.

Remarks

• The full result tables for FID, IS, and CAS are available in results/.

Updates

Enhanced Sampling Scheme [2]

We have published a follow-up paper [2] that enhances the sampling process by resolving its existing limitations, which in turn results in considerably higher fidelity. To be more precise, we first sample a token set with a naive iterative decoding (existing sampling process) and remove the less-likely tokens, and resample the tokens with a better realism approach for tokens. The figure below illustrates the overview of [2].

The visual results are presented in the following figure:

:star: We have included the Enhanced Sampling Scheme from [2] in this repository. We can use it by setting MaskGIT.ESS.use = True in configs/config.yaml.

Citation

[1]
@inproceedings{lee2023vector,
  title={Vector Quantized Time Series Generation with a Bidirectional Prior Model},
  author={Lee, Daesoo and Malacarne, Sara and Aune, Erlend},
  booktitle={International Conference on Artificial Intelligence and Statistics},
  pages={7665--7693},
  year={2023},
  organization={PMLR}
}

[2]
@article{lee2023masked,
  title={Masked Generative Modeling with Enhanced Sampling Scheme},
  author={Lee, Daesoo and Aune, Erlend and Malacarne, Sara},
  journal={arXiv preprint arXiv:2309.07945},
  year={2023}
}