Alexjap

Testing , bad results even on training sample after convergence

15

Right now by following the instructions on the Readme: Training procedure seems to converge (perplexity around 1 on toy example), but when we test on the **same** data ( the...