Hassan ISMAIL FAWAZ

Thanks for sharing the issue. I am not sure if such problem is due to using another version of tensorflow. I will be updating the repository soon to make sure...

I am not sure what's the problem here but based on the link you have sent, it says that the time series length is 0.

Okay so in order to reproduce results on the unequal length datasets, you should use the normalized version (to have equal length). it can be found here: https://www.cs.ucr.edu/~eamonn/time_series_data_2018/

There is a folder inside the archive that contains a normalized version of the unormalized datasets.

Everything should be on this [page](https://www.cs.ucr.edu/~eamonn/time_series_data_2018/).

Yes, you can use these datasets: https://arxiv.org/abs/1811.00075 Shape of input should be: NxLxM N: number of training samples L: Length of the multivariate time series M: The number of dimensions

try to load the model with JIT and use ``` with torch.no_grad(): model.encode_image(...) ``` Or maybe you already use that ... sorry if this does not help I was wondering...

I am not quite sure, but can you try adding more classifiers since 4 is not enough maybe ? Since the test is also based on [friedmann](https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.friedmanchisquare.html) ``` Due to...

This is due to not running the cython script in [utils/build-cython.sh](https://github.com/hfawaz/aaltd18/blob/master/utils/build-cython.sh)

Hi, Indeed we are working on multivariate data and will keep you up to date.