Unable to translate with factored model

[Lichao Zhu]

Hello, 

I am reproducing Netamus factored translation pipeline (FR->EN) following these instructions : https://github.com/rsennrich/wmt16-scripts/tree/master 

I used Stanza Pipeline to produce linguistic annotations in conll format, and trained an rnn model from scratch. But when I tried to use it to translate, this error message showed up :

"ERROR: Expected 5 factors, but input word has 1"

Here's my setting : 

CUDA_VISIBLE_DEVICES=0 python $nematus/nematus/translate.py \

     -m $prefix \
     -i $script_dir/wmt16-scripts/factored_sample/data_fren/dev.factors.fr \
     -o $script_dir/wmt16-scripts/factored_sample/data_fren/dev.output.en -k 5 -n

Does anyone have an idea why it didn't work ? 

Thanks, 

Lichao