# New Release: Ultralytics v8.4.62

**URL:** <https://community.ultralytics.com/t/new-release-ultralytics-v8-4-62/2044>\
**Category:** Discussion\
**Tags:** releases, announcements, ultralytics-official\
**Created:** [June 8, 2026, 2:26pm UTC](https://community.ultralytics.com/t/new-release-ultralytics-v8-4-62/2044 "2026-06-08T14:26:23Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![glenn-jocher](https://sea1.discourse-cdn.com/flex001/user_avatar/community.ultralytics.com/glenn-jocher/32/83_2.png) [@glenn-jocher](https://community.ultralytics.com/u/glenn-jocher)\
**Post date:** [June 8, 2026, 2:26pm UTC](https://community.ultralytics.com/t/new-release-ultralytics-v8-4-62/2044/1 "2026-06-08T14:26:23Z")

</div>

# 🚀 Ultralytics v8.4.62 Released

## 🌟 Summary

We’ve released **Ultralytics v8.4.62** , a high-value reliability update focused on one of the most important parts of training: **making sure your trained models are actually saved** 💾

The headline fix prevents successful training runs from ending without checkpoints due to EMA `NaN/Inf` issues during save checks, especially in some **AdamW + AMP** workflows. This release also includes a substantial **Ultralytics Platform** docs refresh, improved test stability, cleaner external documentation links, and leaner CI.

If you train **Ultralytics YOLO** models regularly, this is a strongly recommended upgrade ✅

## 🔥 Key Highlights

### 💾 Major training fix: checkpoints are no longer discarded

The most important change in this release comes from [PR #24731 by @glenn-jocher](https://github.com/ultralytics/ultralytics/pull/24731), with author profile [@glenn-jocher](https://github.com/glenn-jocher).

This fixes a bug where **valid training runs could complete successfully but still fail to save any checkpoint**.

This was especially relevant for some setups using **AdamW + AMP** , where validation could corrupt live EMA weights and trigger repeated warnings like `Skipping checkpoint save... EMA contains NaN/Inf`.

### ✅ Safer validation during AMP training

With the fix in [PR #24731 by @glenn-jocher](https://github.com/ultralytics/ultralytics/pull/24731), validation still benefits from mixed precision speedups, but it no longer permanently modifies or “poisons” the live EMA model.

That means:

- validation no longer mutates the live EMA in place
- finiteness checks now happen on the original `fp32` EMA
- overflow is safely clamped during checkpoint serialization instead of skipping the save

### 🧪 New coverage for overflow-safe checkpoint saving

This release also adds test coverage to ensure models with large-but-finite EMA weights are still saved correctly, helping protect against regressions in future updates.

## 📘 Documentation Improvements

### 📚 Big Ultralytics Platform docs refresh

The documentation update in [PR #24726 by @glenn-jocher](https://github.com/ultralytics/ultralytics/pull/24726), from [@glenn-jocher](https://github.com/glenn-jocher), significantly improves accuracy and API coverage across the [Ultralytics Platform documentation](https://docs.ultralytics.com/platform).

Highlights include:

- corrected UI labels and workflows
- expanded Platform API reference
- clearer docs for datasets, annotations, training, deployment, billing, teams, and integrations
- newly documented API capabilities such as dataset embeddings, class management, GPU availability, and import flows

If you use the [Ultralytics Platform](https://platform.ultralytics.com), this should make day-to-day workflows much easier to follow 🙌

### 🔗 Fixed broken COCO evaluation links

In [PR #24723 by @glenn-jocher](https://github.com/ultralytics/ultralytics/pull/24723), from [@glenn-jocher](https://github.com/glenn-jocher), outdated COCO evaluation links were replaced with the canonical [COCO upload instructions](https://cocodataset.org/#upload), and old docs author profile links were corrected.

## ⚙ Improvements and Maintenance

### 🧪 Less flaky data-related tests

[PR #24724 by @glenn-jocher](https://github.com/ultralytics/ultralytics/pull/24724), authored by [@glenn-jocher](https://github.com/glenn-jocher), reduces unnecessary downloads in tests and reuses cached assets when possible.

This should improve CI dependability and reduce false failures.

### ⚡ Leaner CI workflows

Two updates helped streamline CI:

- [PR #24725 by @glenn-jocher](https://github.com/ultralytics/ultralytics/pull/24725), from [@glenn-jocher](https://github.com/glenn-jocher), reduces git clone size and speeds up docs publishing and selected test workflows
- [PR #24722 by @UltralyticsAssistant](https://github.com/ultralytics/ultralytics/pull/24722), from [@UltralyticsAssistant](https://github.com/UltralyticsAssistant), updates the Codecov GitHub Action from `v6` to `v7`

## 🎯 Why this release matters

Ultralytics v8.4.62 is not a major model-feature release, but it is a **very important stability release** 🛠

It helps by:

- preventing trained YOLO models from being lost at the end of training
- improving confidence in **AMP-based** training workflows
- reducing wasted compute from failed checkpoint saves
- making Platform docs more accurate and easier to use
- improving internal test and CI reliability

If you’ve ever seen a run finish well but end with **no checkpoint saved** , this release directly addresses that issue.

## 📦 What’s Changed

- [Fix broken COCO eval server and docs author links](https://github.com/ultralytics/ultralytics/pull/24723) by [@glenn-jocher](https://github.com/glenn-jocher)
- [Bump `codecov/codecov-action` from `v6` to `v7`](https://github.com/ultralytics/ultralytics/pull/24722) by [@UltralyticsAssistant](https://github.com/UltralyticsAssistant)
- [Fix flaky data asset tests](https://github.com/ultralytics/ultralytics/pull/24724) by [@glenn-jocher](https://github.com/glenn-jocher)
- [Slim CI git clones: blobless docs publish, shallow Conda tests, drop unused checkout](https://github.com/ultralytics/ultralytics/pull/24725) by [@glenn-jocher](https://github.com/glenn-jocher)
- [Improve Platform docs accuracy and API coverage](https://github.com/ultralytics/ultralytics/pull/24726) by [@glenn-jocher](https://github.com/glenn-jocher)
- [Prevent NaN/Inf EMA from discarding training checkpoints](https://github.com/ultralytics/ultralytics/pull/24731) by [@glenn-jocher](https://github.com/glenn-jocher)

You can also review the [full changelog for v8.4.62](https://github.com/ultralytics/ultralytics/compare/v8.4.61...v8.4.62) or browse the [release page for v8.4.62](https://github.com/ultralytics/ultralytics/releases/tag/v8.4.62).

## 🙌 Try it out

We recommend upgrading and giving it a spin, especially if you train with AMP or AdamW:

```bash
pip install -U ultralytics

```

If you’re starting a new project, we recommend **[Ultralytics YOLO26](https://platform.ultralytics.com/ultralytics/yolo26)** for the best current performance and end-to-end support across detect, segment, classify, pose, and OBB.

## 💬 Feedback

Please try **v8.4.62** and let us know how it works for your training and deployment workflows. Feedback, bug reports, and regression checks are always appreciated — they help us keep YOLO reliable for everyone.