Is there any way I can give custom fairseq config while using fairseq_converter in ctranslate2

Support
, ,
1

Hi,
I am trying to export an fairseq MT model using ctranslate2. The problem is I have given a custom architecture which is creating a key error while using ctranslate2.

While using CT2 convertor for OpeNMT-tf there is a way in which we can give the config file. But that does not work while in case of CT2 fairseq convertor.

Any kind of help would be greatly appreciated

Thank You