Primary competition visual

Lelapa AI Buzuzu-Mavi Challenge

Helping Africa
$1 300 USD
Completed (11 months ago)
Natural Language Processing
Sentiment Analysis
Machine Translation
492 joined
118 active
Starti
Jan 09, 25
Closei
Apr 06, 25
Reveali
Apr 07, 25
User avatar
zeroppl
Questions about restrictions on data and methods
Data Ā· 2 Feb 2025, 18:39 Ā· 6

Hi,

I'm requesting for more info on what training data is permitted? Are we limited to only what's provided or can we use other openly available datasets, including the Inkuba-Mono dataset?

Also, for methods... are there any restrictions?

Discussion 6 answers

Hi,

For training data, the competition rules typically specify whether external datasets are allowed. You should check the competition guidelines and FAQs for clarification. If the use of additional datasets like the Inkuba-Mono dataset is not explicitly mentioned, I recommend reaching out to the competition organizers via the forum or the official rules section to confirm whether it is permitted.

Hope this helps, and good luck with your submission! planet clicker

3 Feb 2025, 07:31
Upvotes 0
User avatar
WoWoGG

In rules, I find it 'You may use only the datasets provided for this challenge.'

and I'm confused. Is InkubaLM the only model we can focus on?or other pretrained models are allowed? Cause the challenge title is 'How can a focused version of InkubaLM for Swahili and Hausa be achieved through model compression techniques?', but in rules 'You may use pretrained models as long as they are openly available to everyone.'

3 Feb 2025, 07:38
Upvotes 0

Regarding the model, yes, InkubaLM is the only model you have to use for the challenge

Can we use external dataset to improve model performance?

User avatar
hark99
Self-employed

It is clearly stated that "You may use pretrained models as long as they are openly available to everyone." They didn't confine only to one model, @Zindi, your assistance is also required in this regard.