Huge shoutout to @stefan027 for always being generous with your solutions your consistency in sharing is a true inspiration and such a big win for the communit... See more

IntroductionThank you to Zindi and Lelapa for hosting a competition on such an important topic. Congratulations to @yvan_carre whose private LB score is on a di... See more

https://youtu.be/8-WfVyUJMyc Thank you all! I learned a lot.

Enjoy! https://www.youtube.com/watch?v=DDYxCWEwlV8https://www.kaggle.com/code/fiveflowerstarfish/lelapa-2nd-place-data-synthesis https://www.kaggle.com/code/fiv... See more

Could someone please provide some clarity around the code review process and the determination of the final LB for this challenge? It says on the Info page that... See more

https://www.youtube.com/watch?v=Ri0rpzLaPRU

are we allow to use LLM models other than InkubaLM-0.4B ?

https://www.youtube.com/watch?v=To-nKbhYXQ8 the score is great but there is a catch :P

🧠 Quick Tip Drop for all Buzuzu-Mavi participants!Data preprocessing is EVERYTHING when working with African languages. Here's what to focus on this week:✅ Fix... See more

I just noticed that XNLI task has three states: true, false or neither. Please checkout the updated notebook https://www.kaggle.com/code/fiveflowerstarfish/zind... See more

Hi guys. Hope you are enjoying this challenge. I would appreciate if anybody can share their approach or hints. Currently, my model is struggling with sentiment... See more

Can someone from @zindi or the organisers please clearly state whether external data is permitted in this challenge or not? The competition rules on the Info pa... See more

when the models have trained on the all tasks.

https://www.youtube.com/watch?v=Co4gaZpFx7wPlease subscribe to my channel if you it helpful. Thanks!

I am getting this error when I am trying to submit. Please could you explain why / fix it. Thanks!

Are we going to use 1 model to fine tune all 3 dataset or what ?

https://www.youtube.com/watch?v=tAujNysSiFM

1. Can we use any open-source LLM as a teacher model to train a student model?2. Can we use any student model or are there any limitations on that?

Hi every one, I hope you're enjoying this challenge. But It seems that there is a bug in the scoring script.I went through the eval script. (https://github.com/... See more

Or i should use one model for all three tasks? Thanks.

I need help for those with successful submissions. I don't know which labels to use for the sentiment analysis.I tried == > 0 : Negative, 1 : Neutral , 2 : P... See more

ValueError: Cannot handle batch sizes > 1 if no padding token is defined.The error persists despite running , tokenizer.pad_token = tokenizer.eos_tokenmodel.... See more

Is it ok to fine tune the model for just swahili and hausa languages for this challenge specifically?, or would that defeat the models purpose of the 5 language... See more

I was looking at demo code submitted by lelapa team. How can they use 0,1,2 to calculate the log softmax ? I checked the vocabulary of lelapa and 0,1,2 does not... See more

Can we use open-source models for the training or finetuning other than InkubaLM?