mlops_framework.md

wandb

1. experiement tracking(By a tracking dashboard services using git)
   1. not easy to log custom metrics (keras model with F1 score and sentence prediction)
   2. still need to figure out how to build your own tracking server.
2. data versioning

mlflow

Intro - 1

Intro - 2

8.9k+ stars

1. experiment tracking(By a tracking dashboard services local services)
   1. is there a hub for collaboration?
2. project management(different environment (conda or docker)) - reprodutivity
3. deployment(unified model format, serving even RestAPI)
   1. there is a model serving functionality(can integrate in gcp?)

DVC

zenml

pixlake

can we do data versioning in pixlake?

daily stuff - yes

how about fix dataset?

v1/v2/v3 and check the difference?