For those who has attempted bahasa Melayu like I did in chatgpt, I'm sure you encountered disappointment or just laugh it off in amusement. There is limited dataset in Malay and therefore the result was dismally bad, mostly translated from Indonesian.

Let's not think about mixing languages around in the style of Singlish or Manglish first. We are having problem with bahasa baku itself. It should be Jepun, not Jepang. It should be ayat, not kalimat.

I am now trying to find more Malay language users to help out in the open source community to help improve it, and make sure our bahasa is represented properly. You are cordially invited contribute in Malay language, regardless of your background, as long as you read and write the language. Feel free to spread the message around so that we have better representation in the multilingual language models.

https://txt.cohere.com/aya-multilingual/
https://www.linkedin.com/feed/update/urn:li...=member_desktop

Join our dedicated Discord server at https://discord.gg/FNtZ9Vab2J?flyingspaghet...=txt.cohere.com to be part of this global effort for the multilingual project. Once it is completed it will be open sourced to all, so that we can use it to propel research progress in the field.

Of course, polyglots of other languages are welcomed to contribute as well! Here is the youtube tutorial regarding the project, everyone is welcomed to participate and contribute to this open source effort https://www.youtube.com/watch?v=Z39YO3kOwqU

There will be exciting prizes to be won for top contributors on the project sprint days, stay tuned for more information on the sprint program!

This post has been edited by HL_O: Jun 21 2023, 01:51 AM