Best Practices for Training Custom AI Models in Squirro

Training with Squirro
1

Hello

I’m currently exploring Squirro’s AI training capabilities and would like to optimize the process of training custom models for better data insights. While the documentation provides an overview, I’m looking for best practices on refining training datasets, improving accuracy, and avoiding common pitfalls. :upside_down_face:

Additionally, I’d like to understand how to effectively leverage feedback loops to continuously improve model performance within Squirro’s ecosystem. Are there specific strategies or configurations that have worked well for others? :thinking: Checked Training with Squirro - The Squirro Forum Machine Learning guide for reference.

Would love to hear experiences from the community and any insights on making AI training more efficient within Squirro! :slightly_smiling_face:

Thank you ! :slightly_smiling_face:

2

Hey!

Great to hear that you’re exploring Squirro’s training capabilities. To optimize your process of training custom models for better data insights, I would recommend starting with Squirro AI Studio – the interface designed to simplify the creation and training of ML models. It follows the usual process of creating a candidate set, establishing a ground truth set, model training, validation, and publishing. You could also explore Squirro’s AutoML capabilities: you can use templates that can train multiple models with automated hyperparameter tuning, and more.

Feel free to reach out if you have more questions or need further assistance!