I just finished the data science with python track on data camp and I am wondering if this competition is something fits my current standing to start with? if so, is there anything I can read to get me started? I am just looking for guidance not worried about the score/money currently. Thanks in advance
In my opinion this is a difficult competition to start as it is multimodal, you need to combine two modalities (text and image).
For me a good starting point is to go to the hugging face library and finetune a model like DistilBert just with the text. Then you may finetune a pretrained ResNet on images and combine both outputs with a mlp.
You can read a lot of useful information here https://arxiv.org/abs/2005.04790
And the winners code is also available
I really appreciate your time and answer, I am looking at these resources now
Personally, I think this competition could be tough for newbies. But, if you are interested, here are the prize-winning solutions and their papers:
#1 alfred lab: Code (not mentioned in the paper), Paper
#2 Niklas Muennighof: Code, Paper
#3 HateDetectron (my team): Code, Paper
#4 Kingsterdam: Code, Paper
#5 burebista: Code, Paper
Hi Everyone
I hope you all are fine and doing great!
Here, I am suggesting some resources (discussions) that are concerned with the individuals who are thinking or have started their journey of breaking out into the data science domain that might be helpful for them:
- The Future of Data Science with Python: Trends and Predictions for 2025 - #2 by coderfreak80 - Articles - Dataquest Community
- The Evolution of Data Science Literature: Books to Read in 2025 - #3 by Soham990
- Analytics Vidhya
Thanks!