Aim 3.21 is shipped! Check out the latest improvements.

Aim 3.21 - Python 3.12 Support, Ability to delete whole experiment

In this tutorial we will show how to use Aim and MONAI for 3d Spleen segmentation on a medical dataset from http://medicaldecathlon.com/. This is a

3D Spleen Segmentation with MONAI and Aim

Explore step-by-step guide for LangChain Debugger, a tool designed for in-depth tracking and visualization of LangChain scripts. The guide features a hands-on example emphasizing its abilities to log LLMs prompts, tool inputs/outputs, and chain metadata, providing a clear understanding of its functionalities.

AI Observability with AimOS: A Deep Dive into the LangChain Debugger

Discover AI observability with AimOS's LlamaIndex Observer. Essential insights for efficient AI model tracking and analysis. Check out our guide for observability solutions in the world of AI.

AI Observability with AimOS: A Deep Dive into the LlamaIndex Observer App

Aim v1.3.5 is now available. Activity View on the Dashboard, X-axis alignment by epoch, Ordering runs both on Explore and on Dashboard. Check out!

Aim 1.3.5 — Activity View and X-axis alignment

Discover Aim 1.3.8's Exciting Features: Advanced Group Coloring & Enhanced Context Table. Streamlined AI metrics visualization for efficient analysis. Explore now!

Aim 1.3.8 — Enhanced Context Table and Advanced Group Coloring

Aim 2.4.0 is out! XGBoost integration, Confidence interval aggregation and lots of UI performance improvements!

It’s time to review the 2022. In a year, Aim has become from a cool open-source project into a full-blown product. We have shipped 12 new Aim versions..

Aim 2022 community report

A retrospective look at the past year! Significant improvements enhanced the functionality and usability of Aim for tracking machine learning experiments for both small and large scale projects.

Aim 2022; Product Recap

Excited to share with you the Aim 3.1! 😊 A huge milestone on our journey towards democratizing AI dev tools. Thanks to the awesome Aim community

Aim 3.1 — Images Tracker and Images Explorer

Hey team, Aim 3.10 is now available! What's new? A lot! Visualize terminal logs, M1 support, better autocomplete experience, integration with CatBoost and LightGBM.

Aim 3.10 — Visualize terminal logs, M1 support & better query autocomplete

In this blogpost I will share the improvements Aim 3.13. We had released 2 more versions (Aim 3.11 and 3.12) that we haven’t had the chance to talk about much. 

Aim 3.13–Figures Explorer and notifications on stalled runs

Hey team, Aim 3.14 is now available! We're excited to introduce Brand new Home Page, MXNet and FastAI loggers, new tooltip positioning modes and ability to edit tags on the runs table!

Aim 3.14 — Brand new Home Page!

Aim 3.15 is out! Callbacks, Logging & notifications, Line chart legends, Audios Explorer, Experiment page, Robust locking mechanism, PaddlePaddle & Optuna integrations.

Aim 3.15 — Callbacks, logging and notifications, line chart legends, experiment page, Audios Explorer

Hey community, Aim 3.19 is out!! It's packed with new enhancements, such as Artifacts support, New HTTP-based tracking server.

Aim 3.19 - New HTTP-based tracking server, Support of Artifacts logging

Aim 3.2 featuring Jupyter notebook integration is now available! 😊 We are on a mission to democratize AI dev tools. 

Aim 3.2 — Jupyter Notebook integration & Histograms

Hey team, Aim 3.20 is out! What's new? Most of our focus have been on Aim robustness and stability. 

Aim 3.20 - More stable and robust Aim

Introducing Aim 3.3: Audio & Text Tracking, Plotly & Colab Integration, Images visualization on run details page. Democratizing AI dev tools with new features. Explore the enhancements now!

Aim 3.3 — Audio & Text tracking, Plotly & Colab integrations

Hey team, Aim 3.4  featuring remote-tracking is now available! 😊 Aim is an open source AI experiment tracking tool.

Aim 3.4 – Remote Tracking Alpha, Sorting & deleting runs

Hey team, Aim 3.5 featuring TensorBoard logs support is now available!! We are on a mission to democratize MLOps tools. 

Aim 3.5 — TensorBoard logs support, Matplotlib integration & System Params logging

Hey team, Aim 3.6 featuring Mlflow logs converter, Chart export, PyTorch Ignite & Activeloop Hub integrations is now available!

Aim 3.6 — Chart export, PyTorch Ignite & Activeloop Hub integrations

Hey team, Aim 3.7 featuring Aim Docker Image is now available! We are on a mission to democratize MLOps tools.

Aim 3.7 — Revamped Run Single Page and Aim Docker Image

Aim 3.8 featuring extensible HuggingFace trainer callbacks is out ! We are on a mission to democratize AI dev tools. 

Aim 3.8 — DVC integration & Extensible HuggingFace callbacks

Discover Aim 3.9's latest features! Add notes on Single Run page, enhance Table component usability, and more. Read the article for full details!

Aim 3.9 — Notes on training runs and upgrade to Pytorch Lightning 1.6 Pytorch Lightning

Discover AimOS: Open-source modular observability for AI systems. Easily log, connect and observe any parts of your AI Systems from experiments to production to prompts to AI system monitoring

AimOS: Open-source modular observability for AI Systems

We have moved the Aim 4.0 to a new repo AimOS and reinstated Aim to its previous version.

Aim 4.0 moved to a new repo AimOS. Here is why?

The release of aimlflow sparked user curiosity, a tool that facilitates seamless integration of a powerful experiment tracking user...

Aim and MLflow — Choosing Experiment Tracker for Zero-Shot Cross-Lingual Transfer

Discover how to leverage Aim's context and subplot features for effortless comparison of validation and test metrics, enhancing your ML experiment analysis.

Aim basics: using context and subplots to compare validation and test metrics

Aim community is officially moving to Discord!! 🎉 Let's rock and build great things together!! The slack workspace is going to become deprecated by the end of 2022.

Aim Community is Moving to Discord!! 🎉

In this blog post, we show how to use Aim’s basic to highly advanced functionality in order to track your machine learning experiments with various

Aim from Zero to Hero

Using Aim’s Explorers along with your Wandb dashboard is easy. Aim allows you to convert Weights & Biases runs into the native format & explore them via Aim UI.

Aim tutorial for Weights and Biases users

Aim 2.2.0 featuring Hugging Face integration is out! Hugging Face integration has been one of the most requested features and we are excited to finally ship it.

Aim v2.2.0 — Hugging Face integration

Aim 2.3.0 allowing System Resource Usage optimization is out! Some highlights: Reverse grouping, Line Chart Smoothing, New aggregation modes. Check out! 

Aim v2.3.0 — System Resource Usage and Reverse Grouping

Explore Aim v2.6.0: No More Docker Requirement, Metric as x-axis. Enhance your analysis with alternate axes, improving metric correlations. Learn how!

Aim v2.6.0 — Docker requirement removed and Metric as x-axis

Aim v2.7.1 is out! We are on a journey to democratize MLOps tools. Thanks to the community for continuous support and feedback! Check out the updated

Aim v2.7.1 — Table export and Explore view bookmarks

🚀 Aim v3.16 is out! Users will be able to view all Run messages right in the UI. Integration of Aim with TensorBoard, Hugging Face Datasets, Acme, Stable-Baselines3. 

Aim 3.16 — Run messages in UI, TensorBoard real-time sync, integration with Hugging Face Datasets

Hey team, Aim 3.18 is now available! Discover what's new: Metric min/max/first values support in Runs Explorer, SQLAlchemy 2.0 Support 

Aim 3.18 — Metric min/max/first values support in Runs Explorer, SQLAlchemy 2.0 Support

The origins of Aim. In the fateful summer of 2020, our friend mahnerak – a researcher at a non-profit lab was hitting the limits of Tensorboard. He wasn’t

Aim’s foundations & why we’re building a Tensorboard alternative

Integration: Aim Logger with XGBoost Library. Learn how to seamlessly use Aim's capabilities in XGBoost experiments. Enhance analysis and insights today! 

An end-to-end example of Aim logger used with XGBoost library

We are excited to announce the release of aimlflow, an integration that helps to seamlessly run a powerful experiment tracking UI on MLflow logs! 🎉

Exploring MLflow experiments with a powerful UI

Explore how AimOS can boost your PyTorch Lightning experiments. This article provides a comprehensive guide with a practical example, emphasizing the integration of PyTorch Lightning into AimOS.

Exploring your PyTorch Lightning experiments with AimOS

We are thrilled to unveil aimlflow, a tool that allows for a smooth integration of a robust experiment tracking UI with MLflow logs! 🚀

How to integrate aimlflow with your remote MLflow

What is a random seed and how is it important? The random seed is a number for initializing the pseudorandom number generator. It can have...

How to tune hyperparams with fixed seeds using PyTorch Lightning and Aim

Explore YerevaNN's use of Aim: Comparing 100s of models for cross-lingual transfer in NLP. Reproducing studies, model selection, and insights revealed. Learn more!

How YerevaNN researchers use Aim to compare 100s of models

Combining Hugging Face and Aim to make machine learning experiments traceable, reproducible and easier to compare.

Hugging the Chaos: Connecting Datasets to Trainings with Hugging Face and Aim

In this article, we'll show you how to keep your language model up to date for question answering tasks. We'll tweak a pre-trained DistilBERT model and fine-tune it on the SQuAD question-answering dataset. 

Keep your model up to date for Question Answering 

As AI systems get increasingly complex, the ability to effectively debug and monitor them becomes crucial. Use Aim to easily trace complex AI systems.

LangChain + Aim: Building and Debugging AI Systems Made EASY!

Deploy Aim on Hugging Face Spaces using the Docker template. Aim empowers you to explore logs with interactive visualizations, easily compare training runs at scale and be on top of ML development insights.

Launching Aim on Hugging Face Spaces

What is Reproducibility in ML? What makes it challenging? Tools and platforms to address different aspects of reproducibility: Data Management, Experiment Tracking, Version Control.

Reproducibility in ML

You can now track your Prophet experiments with Aim! The recent Aim v3.16 release includes a built-in logger object for Prophet runs. 

Reproducible forecasting with Prophet and Aim

AimCallback for Hugging Face is designed to enhance your experiment logging and monitoring. It thoroughly records essential information, including hyperparameters, training, validation, and test time metrics like loss and accuracy.

Streamlining your Hugging Face Experiments with AimOS

Explore heart attack prediction using machine learning.This article breaks down key factors and insights from the Heart Attack Prediction dataset. With the help of Aim we keep track of model performance.

Aim 3.21 - Python 3.12 Support, Ability to delete whole experiment

Key highlights:

Python 3.12 support

Ability to delete whole experiment

Improvements in Large Objects Remote Tracking

Improvements in Running Aim UI on Jupyter Notebook

Learn More