sizeless - Make ML reproducible and safe

TL;DR: sizeless makes it easy for ML practitioners to get their models from prototyping to production in days. Previously, highly trained ML engineers spent weeks trying to manually run, test and improve their models. With sizeless, engineers can focus on building great models - we take care of the rest and make sure that your models are safe and robust, all of this in just a few clicks.

Our Team 💪

We are Cornelius and Roger and want to make machine learning easy and safe. Machine learning is great but extremely hard to use, let alone reproduce! We are super excited to announce sizeless - a platform to automatically scale, benchmark and improve your ML models.

The Problem ❌

Machine learning is very fast-paced. Keeping up with every state-of-the-art model, reproducing results, and benchmarking your model on different data sets is a tedious and manual process. This takes weeks - even for highly trained engineers.

We encountered this problem often during our PhDs: Cornelius in machine learning/robotics at ETH Zurich and Roger in astrophysics at Cambridge.

Our Solution ✅

We automatically run your ML models at scale and perform benchmarks. This allows for reproducibility, transparency, and safety across different models and datasets. We help you flag and understand failure modes to speed up your model development cycle.

Can I use it? 🧐

Yes, please do so by uploading your ML model and data set to run and benchmark your model.  For now, we are focusing on computer vision but will soon expand to other fields, e.g. ML, reinforcement learning, LLMs, and many more! Tell us which one you would love to see and use on our platform!

Our mission 👀

We make machine learning research accessible, reproducible, and transparent for everyone! We want to help you focus on what you enjoy doing most - building great and robust ML models. sizeless is built by and for ML engineers and practitioners.

Asks 🙏: Intros to ML teams who run, test, and deploy ML models

Feedback: We’d love it if you could try it out and provide feedback (sizeless.co)!

Questions: Shoot us an email (info@sizeless.co) with feedback or questions.

Network: Know any ML engineers deploying ML models? Please connect us!