Rebuilding a Model from Your Best Experiment Settings
- Predictive AI
- H2O Hydrogen Torch
This short lesson shows how to build on your best-performing experiment in H2O Hydrogen Torch rather than starting from scratch. It walks through locating the experiment with the top mAP score in the experiments table, opening the three-dot menu, and selecting New experiment to carry its settings into a fresh configuration you can then tune further. It closes with a practical challenge for learners with Adobe LMS access: an assignment to keep adjusting hyperparameters and try to beat the current best experiment score.
Key takeaways
- How selecting New experiment from your best-performing run carries its settings into a fresh configuration
- Why starting from your best experiment's settings is faster than tuning hyperparameters from scratch again
- What the 'Beat the Best Experiment Score' assignment asks learners to do with their tuned settings
Start owning your AI