That’s a wrap on the Landslide Detection Challenge, and I’m proud to say we finished 4th place live, on stream!
When we started this journey, we aimed for Top 1 or Top 3 as the worst case scenario, and even though we just missed the podium, I couldn’t be happier with the outcome. This was our first-ever livestream series, and competing live while explaining everything in real time was both exciting and humbling.
The best part? We built everything together, from data prep to augmentations and final model tweaks.
Massive thank you to everyone who joined the Twitch streams, subscribed to the YouTube channel, or watched the replays. And if I helped you in anyway , let me know in the comments under this discussion.
You can always catch past streams and new projects here: 🔴 Twitch Schedule 📺 YouTube Channel
Now on to the next competition! 💪🔥
Thanks for your useful sharing, I wish you a high ranking in the next competitions. My goal is to try and break your top 1 ranking, but it is difficult. I can only accumulate points through each competition. Recently most of the zindi competitions are not open to all 😁
You can. Just win three solid competitions 🤣
Thank you for the kind words @3B, and huge congrats on winning such an interesting competition! Looking forward to that write-up soon. 😄 Best of luck trying to steal my number 1 spot, it won’t be easy though! 😅
Of course, the Eva model was a big help! Thank you for the livestream. Really learnt a lot🙏🙏
You should have tried the beit model as well. Actually my best sub was an ensemble of the beit model we used on stream and catboost!!!!
Wow I really missed out. I had over ensembled that might be the reason I didn't add the Beit model. Yolo, lightgbm, efficientvnet and Eva. 🤣
Thank you
You're welcome!
I wish I had the time to practice what I learnt on the stream on the competition. I will definitely try it out later. Thank you @Koleshjr. On to the next 💪💪
No worries brother. Thank you for the support in joining the streams and sharing the links. Meant a lot!
I like good things 😀 😀, that's why I join 😀 .