Deepchecks Testing Package - Open source tool for testing & validating your model & data

Hi Product Hunt! 🙋‍♀️ I am Shir, co-founder and CTO of Deepchecks. Excited to share a bit about the journey of developing and launching an open-source package for testing machine learning models! At Deepchecks we’re seeing LOTS of models in production. While monitoring them is a must, we noticed that the tools for something much more basic was missing: How should these models be tested before they’re even deployed? 🤔 Like every startup - we started with experimentation. We named the first prototype the “Preemptive report”. Despite the REALLY compelling name 😅 the first people who saw it responded with: “How come this doesn’t exist yet??” At the same time, we realized that testing for problems with your data, model, and ML process, is relevant also much earlier in the research phase. The smiles on the faces of Data Scientists when running our package on a new batch of data, told us that this understanding is correct... As we are a company of tech-savvy engineers, it was clear that the way to go for a dev-tool used during research, was to build it open source, and we started working on the package towards its planned release. After iterations with hundreds of Data Scientists from the amazing community (to which we owe many thanks 🙏) and receiving great feedback and suggestions (for usability, additional ideas for checks, requests for integrations with tools, and many more), we’re happy to announce the deepchecks testing package here on Product Hunt 🚀 Show your support by giving us a star ⭐️ on github: https://www.github.com/deepchecks/deepchecks Our commitment is to keep this project open source, and to continue building a comprehensive ecosystem of tools, for “Continuous Validation” of ML pipelines, from research to production – for Tabular, CV, and NLP Data & Models. More links to jump right in: - Get Started - Join Our Slack Community for more updates and info about ML validation and testing - Open an Issue on Github for feature requests and feedback

👋 Hi all! 👨‍🦰 I’m Philip Tannor, I’m the CEO and one of the co-founders at Deepchecks. Thank you SO MUCH for the support you all have been giving us since we released this package, it really means the world to Shir, myself, and the rest of the team. 👨‍👩‍👧‍👦 We’ve always had a passion for the community-focused approach, but that’s not how Deepchecks started. After building out a (closed) solution for monitoring ML models in production, we kept looking for a way to provide value to a much wider community. It took us over a year of hard work, but we ended up realizing that we can provide the ML community with the (extensive) open-source package that you see here today while focusing our commercial efforts elsewhere (e.g. where other non-ML-practitioner users are involved, etc.). 🚀 The open-source package was released at the beginning of this year, and in the meantime, the feedback has been incredible. It’s really a privilege to get to work on a package that’s getting so much love from the professional community. If you haven’t yet, give it a try - should only take a few minutes. And can’t wait to get your feedback and your support (both here and on GitHub - https://github.com/deepchecks/deepchecks ). 👂 We’re constantly looking to add more checks & suites, as well as to support more types of data and frameworks. Since our roadmap is based primarily on community feedback, please let us know what you do or don’t like, as well as what you think we’re missing! Shir and I will also be here throughout the day to answer any questions that you may have.

🙏🙏🙏 Thank you all SO MUCH for the open source love. 🤩 If you like what we're doing, the best ways to support us are by starring our repo (https://github.com/deepchecks/de...) or by leaving feedback via our GitHub issues (such as opening a feature request or a bug report here - https://github.com/deepchecks/de...).

Congrats on the launch @ptannor @shirch !

Sounds pretty cool @shirch. Congratulations on the launch. Going to share it with ML team to try for upcoming stuff :)

Kudos team Deepchecks! It's a pleasure to see a repo with such great docs (: Just out of curiosity, how is this different than Great Expectations?

congrats for the launch team!

Thanks for sharing ! The testing package has been super helpful for my team!

Can confirm whenever I'm validating my models Deepchecks makes me feel warm and fuzzy inside.

Really cool product. I'm following you guys for a while :) good luck w/ the launch