ML Model Labs is a comprehensive environment for developing, experimenting with, and evaluating machine learning models. It provides a structured workspace for data scientists and researchers to conduct experiments, train models, and analyze results across various ML tasks and datasets.
- each experiment dir has several (potentially) mlflow experiments
- the granularity of the data is several datasets per filename experiment. Each filename experiment has several mlflow experiments
- Log models to Mlflow
- Check if metrics work in a distributed enviroment
fabric run --accelerator=cuda --devices=1 --main-address=localhost --main-port=29400 ./experiments/thesis/run.py