search

cesium-ml / cesium

Machine Learning Time-Series Platform
Other
670 stars 101 forks source link

Error: too many indices for array #253

Closed friedlysol closed 7 years ago

friedlysol commented 7 years ago

I'm loading up a larger dataset and I get this erro when loading into the model.

The header file looks like this:

filename,target 22090057.txt,a 22090282.txt,a 22109627.txt,a 22109630.txt,a 22110423.txt,a 22113257.txt,a 22126905.txt,a 22127034.txt,a 22128488.txt,a

Individual files look similair to: 12,1.53 13,1.64 14,1.64 15,1.66 16,1.74 17,1.76 18,1.82 19,1.90 20,1.96 21,1.95 22,1.96 23,2.03 24,2.07 25,2.09 26,2.17 27,2.20 28,2.26 29,2.40 30,2.44

Loading the demo data, or loading small subset of the data I have problem with works fine. Any suggestions?

acrellin commented 7 years ago

Can you provide the full traceback? And how large is the dataset?

friedlysol commented 7 years ago

Where do I get the traceback from? Nothing shows up in the console.

Here is the data I load: https://github.com/friedlysol/cesium253

friedlysol commented 7 years ago

Anything that I can do to further identify the problem?

acrellin commented 7 years ago

@friedlysol the issue is with the time series with only a single epoch. The minimum number of epochs per time series is two. We will update the error message to be more informative. Thanks for bringing this to our attention.