Closed Eugen2525 closed 4 years ago
Hi @Eugen2525, There are at least two problems when creating solutions for established tasks on new languages: datasets and tokenization. What tasks are you interested in? We do have two multi-language models you can take a look at: https://demo.deeppavlov.ai/#/mu/textqa
Dear @yoptar Thanks for a prompt reply, I wanted to create chat bot just like the default one provided which runs on the Cambridge restaurant dataset.
So, I am interested what steps should I take to make a chatbot for Korean? I know for sure I need a dataset like the Cambridge restaurant dataset for the training of the bot, but what are some other things needed to be done?
Could you please elaborate...
basic config uses embeddings for English
Nothing with Russian? What's the quickest way to build a VK and FB bot?
I am not sure if you mean if there is a Russian language support (which I saw there is), but for Korean, I just translated datasets and it worked.
For the quick prototyping you can select any dataset and just try and see with the mentioned approach. This is what I did at least
Hi,
I am considering creating language support for the Korean language, and I am interested in what are the requirements, steps I need to take?
Could you please elaborate on the steps and also on the feasibility of doing so?
Thanks! ``