ValueError when running seq2seq_train.py

[djoker911]

Hi all,
I met ValueError: Sample larger than population when I ran seq2seq_train.py,
and I ran it on python3.5 virtualenv , tensorflow 1.0.0, numpy1.13.1

Is there any parameter or code should I adjust to run it correctly

btw, line 17 in data_util.py could be np.arange(...) not np.arrange(...)

Thanks for help!

[MayurNagar]

change config.batch_size less than 13. The code return is not proper. There are too many places where the code had to be corrected. for validation batch_size should be 1 else same error will pop up.

[nauman-akram]

@MayurNagar did you correct the code and generate some output? please help @anooshac I get this error even if I reduce the batch size or increase data set : IndexError: index 10 is out of bounds for axis 0 with size 1. could understand how to fix this.

[nauman-akram]

[AnjirHossain]

Could this be related to sampling without replacement? I've found link. I'm new to ml so I'm not entirely sure how to conduct generating batches.

[nauman-akram]

@MayurNagar do you have any idea how to test/predict this model?

[suryacaprice]

I have the same doubt , how to test / predict the model ?