Prophet VS MLflow

Prophet is a forecasting model made by Facebook.

I like FB Prophet and LinkedIn's GreyKite

Prophet

I like Meta's Prophet. It's a flexible model that I feel is easy to understand and explain to non-technical stakeholders. Implementations for R and python. It can be as simple as a univariate series, or accept exogenous explanatory variables (categorical & continuous). Seasonality is part of the model, out of the box. It also has decent default parameters if you're not looking to do a lot of tuning. Lastly, the documentation is quite thorough and approachable. https://facebook.github.io/prophet/

Besides building, testing and deploying, Kraken CI is also a pretty nice tool to build an MLOps pipeline. In this article, it will be shown how to leverage Kraken CI to build a CI workflow for machine learning using MLflow.

For MLFlow users you can now serve models directly in MLFlow using MLServer and if you're a Kubernetes user you should definitely check out Seldon Core - an open source tool that deploys models to Kubernetes (it uses MLServer under the covers).

But as you can imagine, tracking each experiment with Git can become a hassle. We’d like to automate the logging process of each run. The same as for large file versioning, many tools emerged in recent years for experiment logging, such as W&B, MLflow, TensorBoard, and the list goes on. In this case, I believe that it doesn’t matter with which hammer you choose to hit the nail, as long as you punch it through.

Its a bug with mlflow -> https://github.com/mlflow/mlflow/issues/3755 Keep the server on, open another terminal export MLFLOW_TRACKING_URI env variable, if on windows set the env variable.....should work.

mlflow for logging and visualizing ML model experiments

So, if you even want to use MLFlow to track your experiments, run the pipeline on Airflow, and then deploy a model to a Neptune Model Registry, ZenML will facilitate this MLOps Stack for you. This decision can be made jointly by the data scientists and engineers. As ZenML is a framework, custom pieces of the puzzle can also be added here to accommodate legacy infrastructure.

Consideration Number #2: Consider using model life cycle development and management platforms like MLflow, DVC, Weights & Biases, or SageMaker Studio. And Ray, Ray Tune, Ray Train (formerly Ray SGD), PyTorch and TensorFlow for distributed, compute-intensive and deep learning ML workloads.