paradigm
dagster
paradigm | dagster | |
---|---|---|
9 | 46 | |
36 | 10,215 | |
- | 2.1% | |
7.6 | 10.0 | |
11 months ago | 5 days ago | |
Python | Python | |
Apache License 2.0 | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
paradigm
-
Deploying speech recognition models at scale
I built Paradigm from scratch to deploy any model at scale. It deploys the model on Kubernetes with load balancers. If you run into any issues, I'm happy to guide you on how to use it.
-
Which is the best MLOps tool for getting started?
I started with paradigm. I got a deeper understanding about argo workflows through it as well. Helps to get a proper grab of industry standards from the beginning.
- What are some open-source ML pipeline managers that are easy to use?
- I use this OS tool to deploy LLMs on Kubernetes.
-
Serving Scikit-Learn model on EC2 instance and Scaling
For scalability, it should be on Kubernetes. This is the best solution I have come across. You can deploy the model as a service with a LoadBalancer. You can refer to Kubernetes services or use a tool such as this one that handles building the service for you.
-
Who wants to run ML pipelines on Kubernetes? This might be the simplest tool for the job.
I came across this tool today and checked it out, I feel this can get the job done very quickly without so many complex features. It is also very small in size, so does not take up a lot of space in the cluster as well.
-
[P] I found the simplest tool to run ML pipelines on Kubernetes. Github link in comments.
Link - https://github.com/ParadigmAI/paradigm It seems to be a pretty new project. But this has high usability.
-
Airflow + Slurm for ML Training Pipelines?
Prefect is a good choice, But I wanted a much simpler tool. Hence, I built a barebone workflow controller here.
dagster
- Experience with Dagster.io?
-
Dagster tutorials
My recommendation is to continue on with the tutorial, then look at one of the larger example projects especially the ones named “project_”, and you should understand most of it. Of what you don't understand and you're curious about, look into the relevant concept page for the functions in the docs.
-
The Dagster Master Plan
I found this example that helped me - https://github.com/dagster-io/dagster/tree/master/examples/project_fully_featured/project_fully_featured
-
What are some open-source ML pipeline managers that are easy to use?
I would recommend the following: - https://www.mage.ai/ - https://dagster.io/ - https://www.prefect.io/ - https://metaflow.org/ - https://zenml.io/home
-
The Why and How of Dagster User Code Deployment Automation
In Helm terms: there are 2 charts, namely the system: dagster/dagster (values.yaml), and the user code: dagster/dagster-user-deployments (values.yaml). Note that you have to set dagster-user-deployments.enabled: true in the dagster/dagster values-yaml to enable this.
-
Best Orchestration Tool to run dbt projects?
Dagster seemed really cool when I looked into it as an alternative to airflow. I especially like the software defined assets and built-in lineage which I haven't seen in any other tool. However it seems it does not support RBAC which is a pretty big issue if you want a self-service type of architecture, see https://github.com/dagster-io/dagster/issues/2219. It does seem like it's available in their hosted version, but I wanted to run it myself on k8s.
-
dbt Cloud Alternatives?
Dagster? https://dagster.io
-
What's the best thing/library you learned this year ?
One that I haven't seen on here yet: dagster
- Anyone have an example of a project where a handful of the more popular Python tools are used? (E.g. airbyte, airflow, dbt, and pandas)
- Can we take a moment to appreciate how much of dataengineering is open source?
What are some alternatives?
Kedro - Kedro is a toolbox for production-ready data science. It uses software engineering best practices to help you create data engineering and data science pipelines that are reproducible, maintainable, and modular.
Prefect - The easiest way to build, run, and monitor data pipelines at scale.
flecs - A fast entity component system (ECS) for C & C++
Airflow - Apache Airflow - A platform to programmatically author, schedule, and monitor workflows
Mage - 🧙 The modern replacement for Airflow. Mage is an open-source data pipeline tool for transforming and integrating data. https://github.com/mage-ai/mage-ai
aws-sfn-resume-from-any-state - Resume failed state machines midstream and skip all previously succeded steps.
airbyte - The leading data integration platform for ETL / ELT data pipelines from APIs, databases & files to data warehouses, data lakes & data lakehouses. Both self-hosted and Cloud-hosted.
zenml - ZenML 🙏: Build portable, production-ready MLOps pipelines. https://zenml.io.
MLflow - Open source platform for the machine learning lifecycle
wenet - Production First and Production Ready End-to-End Speech Recognition Toolkit
meltano