ModelForge is a machine learning model building competition organized as part of Dakshh, the annual technical fest of Heritage Institute of Technology, Kolkata. The challenge is designed to test participants’ ability to analyze system-level data, engineer meaningful features, and build robust models.ModelForge will be conducted in two stages. The preliminary round will be hosted online on Kaggle, where teams will build and submit their models using the provided dataset. Based on leaderboard performance, the top 8 teams will qualify for the final round at the Heritage Institute of Technology campus, where they will further test and refine their models on a new dataset.Whether you are passionate about machine learning, data science, or system analytics, ModelForge offers an exciting opportunity to apply your skills in a competitive and real-world-inspired challenge.
Guidelines:
Team Size: Each team must consist of 2–3 members only.
Platform: The preliminary round will be conducted on Kaggle.
Dataset Release: Training and test datasets for the preliminary round will be released on 8 March.
Submission Deadline: Participants can submit their predictions on Kaggle until 11 March, 08:00 PM.
Multiple Submissions: Teams are allowed to submit multiple predictions before the deadline; the best leaderboard score will be considered.
Shortlisting: The top 8 teams from the preliminary leaderboard will qualify for the final round.
Final Round Venue: The final round will take place offline at Heritage Institute of Technology during Dakshh on 14 March.
Final Dataset: Shortlisted teams will be given a new test dataset during the final round.
Model Updates: Teams may retrain or modify their models during the final round before submission.
Final Submission: Final predictions must be submitted on Kaggle within the allotted time.
Winning Criteria: The top 3 teams on the final leaderboard will be declared winners.
Rules:
Original Work: All models and code must be developed by the participating team. Plagiarism will lead to disqualification.
No External Data: Only the dataset provided in the competition may be used.
Fair Play: Any attempt to manipulate results or gain unfair advantage will result in disqualification.
Model Explanation: Finalist teams may be asked to explain their model and approach.
Judging Decision: Decisions of the judges and organizers at Heritage Institute of Technology will be final.
