# \#benchmark

**URL:** https://discourse.ossbase.org/tag/benchmark/37.md

[Latest](https://discourse.ossbase.org/latest.md) · [Categories](https://discourse.ossbase.org/categories.md) · [Tags](https://discourse.ossbase.org/tags.md)

---

## [GPU Efficiency in VLAI Model Training](https://discourse.ossbase.org/t/gpu-efficiency-in-vlai-model-training/738)

<div class="topic-metadata">

**Author:** [@cedric](https://discourse.ossbase.org/u/cedric)\
**Replies:** 0\
**Last updated:** [December 12, 2025, 12:03pm UTC](https://discourse.ossbase.org/t/gpu-efficiency-in-vlai-model-training/738 "2025-12-12T12:03:58Z")

</div>

Earlier this year, we introduced our VLAI Severity Classification model. Since then, we’ve been training it throughout the year on our own hardware, right here in Luxembourg at CIRCL. With 2025 coming to a close, we want…