Reports

Monitor and share key insights from your model training metrics

Compare multiple metrics on a single chart to better understand model performance. Highlight trends by dividing training runs into groups. Reports give you a customizable analysis of specific metrics in one place.

Request free trial See Reports documentation

The problem

When it’s hard to see what’s happening inside your models, it’s hard to move fast

Do any of these sound familiar?

“What’s a quick way to check the most important metrics for my long training jobs?”

“How can I easily track trends and changes in key metrics over time?”

“Is there a shareable way to analyze insights and make decisions about our results so we can collaborate without the hassle?”

“How can I get a concise summary of our key findings at my fingertips? Retaining stakeholder trust would be so much easier!”

“How do I create customized views of the same data for technical and non-technical stakeholders — and save a ton of confusion?”

“Is there a simple way to document my work so I can come back and continue it later?”

The solution

The answer to all these questions? Reports

Here’s a breakdown of how Reports make your training more productive.

Tailored views

Choose what data and runs to display in your UI

Tailor the UI to your needs by creating reports showing only the runs you monitor every day. Or the ones you look at when something goes wrong.

Create new widgets in your reports with the simple click of a button.
Make your reports easier to analyze by organizing widgets into titled, collapsible segments.
Add descriptions, explanations, and context to your report using markdown to make sure your colleagues understand what they’re looking at.

Custom plots

Gain a deeper understanding of your model metrics

Get better insights into your model performance by customizing your widgets.

Compare multiple metrics on a single chart.
Find patterns and trends in large-scale projects faster by dividing runs into groups.
Create more meaningful metrics by defining your own mathematical expressions based on existing data — all inside Neptune.

Collaboration

Get everyone on the same page. Literally.

Making collaborative decisions. Presenting results to stakeholders. Sharing progress with your team. Communicating key insights is easier when you can display your most important metrics on a single screen.

Use version control to track exactly who made changes and when.
Add runs from different projects to evaluate model performance across your organization.
Share reports with a persistent link.

Our customers

Reports take the clunky out of your communication

Just ask these companies

Ronen Ben-David Algorithms Team Lead at HP

We primarily use Neptune for training monitoring, particularly for loss tracking, which is crucial to decide whether to stop training if it’s not converging properly. It’s also invaluable for comparing experiments and presenting key insights through an intuitive dashboard to our managers and product owners.

Fabio Fonseca Data Scientist at Marcura

The killer feature in Neptune is custom dashboards. Without this, I wouldn’t be able to communicate my simulations to Developers, the Analytics team, and business stakeholders without any hassle. Neptune gives our Data Scientists the piece of mind that their best results won’t be lost and that communication will be a breeze.

Łukasz Grad Chief Data Scientist at ReSpo.Vision

I like dashboards because we need to monitor several metrics. So you code the dashboard once, have those styles, and easily see everything on one screen. Then, any other person can view the same thing, so that’s pretty nice. For all our running pipelines, we use those Neptune dashboards.

Demo