About Code release for "SimMTM: A Simple Pre-Training Framework for Masked Time-Series Modeling" (NeurIPS 2023 Spotlight), https://arxiv.org/abs/2302.00861
107
stars
13
forks
source link
Works for Multivariate Time Series Classification too? #6
It seems that in the paper, you have used a multivariate time series classification dataset (GESTURE). However, the code seems to work on univariate time series, and there is no configuration for the GESTURE. Directly using your code for MTS classification results in huge reconstruction loss. Could you please kindly share if your SimMTM work on MTS classification self supervised training? If so, could you share the implementation?
Hello, thank you for sharing your valuable work.
It seems that in the paper, you have used a multivariate time series classification dataset (GESTURE). However, the code seems to work on univariate time series, and there is no configuration for the GESTURE. Directly using your code for MTS classification results in huge reconstruction loss. Could you please kindly share if your SimMTM work on MTS classification self supervised training? If so, could you share the implementation?
Thank you in advance!