I am following the quick start in:
python preprocess.py -train_src data/src-train.txt -train_tgt data/tgt-train.txt -valid_src data/src-val.txt -valid_tgt data/tgt-val.txt -save_data data/demo
I trained 2million fr-en sentece pairs, everything went smooth, but the translation result looks very bad, even the sentence in the training source file are not translated correclty and report unknown words. Any suggestion? I follow the quick start step by step.