Before you dive into 2024, have a look at what we released in Neptune in the last 3 months.
New 🎉
1. MLflow plugin
The new Neptune-MLflow integration allows you to send your metadata to Neptune while using the MLflow logging code.
It should be especially handy when you already use MLflow in some of your projects but you’d like to enhance the tracking with Neptune’s functionality.
2. Documentation updates
To make your life easier, we’re constantly improving our documentation.
There’s a new API index available. This page lists all functions, parameters, and constants exposed by the Neptune Python API.
You can also have a look at the overview of best practices to make sure you’re getting the most out of Neptune.
And we published a bunch of new guides and tutorials:
- Overwrite logged metadata
- Add a new field to existing runs
- Delete metadata from a run
- Widgets and supported field types
- …and more. See all the documentation news here.
By the way, we enabled the docs feedback form in the footer of each page. So if you have any notes or suggestions, go for it.
You have a feature request?
Improvements 🤓
Python client library
- We introduced a callback mechanism that you can enable in case the synchronization is not working properly. (neptune 1.8.0)
- You can now programmatically empty the project trash. (neptune 1.8.0)
- We reworked disk usage checking and introduced two related environment variables: NEPTUNE_MAX_DISK_USAGE and NEPTUNE_RAISE_ERROR_ON_DISK_USAGE_EXCEEDED. (neptune 1.8.3)
- When fetching runs with Project.fetch_runs_table(), we removed the limit where a maximum of 10 000 runs could be returned. We also improved the performance, making the fetching up to twice as fast as before. (Also applies to model-fetching methods: fetch_models_table() and fetch_model_versions_table()). (neptune 1.8.6)
Integrations
- We added support for the description argument in tf.summary.image(). Changed single images to be logged as a separate field. (neptune-tensorboard 1.0.1)
- We added OmegaConfig support. Replaced some dataset classes and modules. (kedro-neptune 0.3.0)
Web application
- When modifying a search query in the runs table, you can now edit the operator and value without changing the other parts. By the way, here’s a nice explainer on searching and filtering the runs table.
- You can now click on a username in the Owner column (in the runs table) to instantly create a filter for runs created by that account.
- Project owners can now set storage limits for projects individually.
Fixes 🔨
Python client library
- We fixed the behavior of synchronization callbacks. (neptune 1.8.5)
- We changed the behavior of tracking system metrics so that no monitoring namespace is created if all related options are disabled. This fixes an issue with extra fields being created especially if resuming a run multiple times. (neptune 1.8.6)
Web application
- We fixed an issue where a field named “type” could not be displayed as a runs table column.
See other minor improvements and fixes in the changelog.
On the roadmap 🔜
We’re planning a significant release for February. You can expect:
- Reports
- Run groups
- Various UI improvements (including custom and persistent run colors)
Keep an eye on our roadmap here.
More about Product Updates December ’23: MLflow Plugin, New Docs Tutorials, and More
Check out our product resources and related articles below:
Related article
Customizing LLM Output: Post-Processing Techniques
Related article
Deep Learning Optimization Algorithms
Related article
Track and Visualize Information From Your Pipelines: neptune.ai + ZenML Integration
Related article
Product Updates September ’23: Scatter Plots, Airflow Integration, and More
