Time series Timeseries Deep Learning Machine Learning Python Pytorch fastai | State-of-the-art Deep Learning library for Time Series and Sequences in Pytorch / fastai
I've encountered an issue with the get_splits() function when using balance=True and specifying the test_size. It appears that the function does not behave correctly in this scenario.
The output should still contain the same number of items per class when test_size is specified. However, it seems that something is going wrong when the test_size is used in conjunction with balance=True.
I would appreciate any help in resolving this issue. Thank you!
Description:
Hello everyone,
I've encountered an issue with the
get_splits()
function when usingbalance=True
and specifying thetest_size
. It appears that the function does not behave correctly in this scenario.Reproduction Steps:
os : Linux-5.4.0-100-generic-x86_64-with-glibc2.31 python : 3.10.10 tsai : 0.3.6 fastai : 2.7.12 fastcore : 1.5.29 torch : 2.0.0
Here's a code snippet to reproduce the issue:
It worked correctly:
However, when the
test_size
parameter is also specified, the behavior is incorrect:The output is:
The output should still contain the same number of items per class when
test_size
is specified. However, it seems that something is going wrong when thetest_size
is used in conjunction withbalance=True
.I would appreciate any help in resolving this issue. Thank you!