Response Quality control

[chiranshu14]

Hi,

This is one of the best implementations of a gpt2 bot so far. I tried many, but this is the fastest of all.

I'm currently testing the medium-cpu model for a "Quick response" generation for chat application. It works, but the responses seem to be reddit-inspired. They contain slang language like 'gib' instead of give etc. Can get violent at times. How can I make it "professional". So that it only gives good responses to be used in business/professional applications?

Thanks!!

[polakowo]

The best you can do is to fine tune it on your custom dataset, even a few texts (<100mb) can do a trick. This bot reflects an average discussion on Reddit, and business is not an integral part of it. You can also experiment with the temperature and other params that control creativity, and decrease it to get “normality”. Or penalize shorter responses. Or include a ranker or two to filter out responses that would lead to shorter discussions on Reddit. There are a lot of ways, each having its own pros and cons.