Comet ML

---

Comet is a machine learning (ML) platform that enables data scientists and machine learning engineers to track, compare, monitor, visualize, and optimize machine learning models from training to production monitoring. Its core features include experiment management, artifact management, model registry, and model production monitoring (MPM).

Machine learning practitioners can use Comet’s platform on the cloud, virtual private cloud (VPC), or on-premises with dozens of installation flavors. Comet’s extensible and fully-customizable platform allows ML practitioners to use their preferred workflow, infrastructure, and tools. By just adding two lines of code, Comet automatically logs hyperparameters, metrics, code, and more.

To promote knowledge sharing and reproducibility in the ML community, Comet offers free access to open source projects and academia.

History

Comet was founded by Gideon Mendels and Nimrod Lahav in 2017. Its headquarters is currently based in New York, with offices in Tel Aviv, Israel. Comet completed the Alexa Accelerator Program and received funding from the Amazon Alexa fund in 2017.

In 2018, Comet received a $2.3 million in seed funding led by Trilogy Equity Partners, alongside Two Sigma Ventures, Founders Co-Op, Fathom Capital, Techstars Ventures and several angel investors. As part of the funding, Yuval Neeman, Partner at Trilogy Equity Partners, joined Comet’s Board of Directors along with Matt Jacobus, Two Sigma’s Former CTO and a Venture Partner as a board observer. Comet received additional funding of $4.5 million from Trilogy Equity Partners, Two Sigma Ventures, and Founder’s Co-Op in 2020.

Motivated by a 500% growth in the previous year, Comet raised $13 million in 2021 in Series A funding led by Scale Venture Partners and its existing investors, Trilogy Equity Partners and Two Sigma Ventures. Andy Vitus of Scale Venture Partners joined Comet’s Board of Directors in conjunction with the funding. Comet also announced that it had acquired Stakion, a model production monitoring company based in the U.K. for an undisclosed amount. The acquisition was intended to support Comet’s newest product, Model Production Monitoring, a solution that helps organizations monitor the quality of their model after deployment. The Stakion team joined Comet post-acquisition and will continue to innovate production monitoring.

Six months after raising their Series A funding, Comet raised $50 million in Series B funding led by OpenView. Today, Comet has trained millions of models across various industries. Over 450 enterprises, startups, and academic teams use Comet’s platform to support their machine learning projects. Comet’s customers include major market leaders and Fortune 500 companies such as Uber, Zappos, Ancestry, Etsy, Boeing, and more.

Recognitions

Comet was named one of the 2022 Gartner Cool Vendors™ in AI Core Technologies. CRN®, a brand of The Channel Company, also recognized Comet in its 2022 Emerging Vendors list in the AI/Machine Learning Tools category.

Integrations

Comet integrates with several tools, frameworks, and platforms used by machine learning teams today. Comet’s extensible and fully customizable machine learning platform makes it easy for teams to incorporate their preferred tools. Current machine learning framework integrations include:

• Tensorflow
• Keras
• Pytorch
• HuggingFace
• Pytorch Lightning
• XGBoost
• LightGBM
• Ludwig
• MLFlow
• Prophet
• Pyspark
• scikit-learn
• Shap
• matplotlib
• Catalyst
• Spark NLP
• Ray

Comet also integrates with the following machine learning tools:

• Annoy
• Aquarium Learning
• Gitlab
• Gradio
• Metaflow
• Sagemaker
• TensorboardX
• Vertex AI
• YOLOv5
• Kubeflow
• Anomalib

References

This article "Comet ML" is from Wikipedia. The list of its authors can be seen in its historical and/or the page Edithistory:Comet ML. Articles copied from Draft Namespace on Wikipedia could be seen on the Draft Namespace of Wikipedia and not main one.

---