Model Development - Predict Script

[RobotPsychologist]

The predict script should:

• load trained models from 0_meal_identification/meal_identification/models
• load data from 0_meal_identification/meal_identification/data/processed
• Evaluate trained models with sktime metrics
• Generate reports of metrics for a particular experiment 0_meal_identification/meal_identification/reports

Examples:

• Example predict script: https://github.com/manifoldailearning/Complete-MLOps-Bootcamp-v2/blob/962209facd70514febf28c5b0e77f4ea00fdd894/Packaging-ML-Model/packaging-ml-model/prediction_model/predict.py
• Deep learning example but similar functionality: https://github.com/mrdbourke/pytorch-deep-learning/blob/6aa8b4c9f868c6fc6113badce81b38cda71dfd30/going_modular/going_modular/predictions.py

[safiyamak]

interested

[liya-zhu]

hi

[RobotPsychologist]

@safiyamak @liya-zhu I provided you with a bit more context to get started.

[safiyamak]

Base our metrics off of the main three classes listed here: https://www.sktime.net/en/stable/api_reference/performance_metrics.html#time-series-segmentation