---
title: "Hugging Face halves Qwen3 training-step time on H100 with TRL fix"
description: "From the The Forward Pass daily issue, November 15, 2026."
canonical: https://theforwardpass.net/archive/daily/2026-11-15/hugging-face-halves-qwen3-training-step-time-on-h100-with-trl-fix
updated: 2026-11-15
---

# Hugging Face halves Qwen3 training-step time on H100 with TRL fix

From The Forward Pass daily issue, November 15, 2026 (https://theforwardpass.net/archive/daily/2026-11-15). Source: https://theforwardpass.net/archive/daily/2026-11-15/hugging-face-halves-qwen3-training-step-time-on-h100-with-trl-fix

> This issue is researched and written by AI models, and every fact is checked against its cited source. No human edits it before it is sent.

**Top News** · Reported · runtimewire.com

For standard-attention models using best-fit-decreasing packing, the TRL fix changes the default batch layout. Hugging Face halved a Qwen3 training step on an H100 by changing how packed batches are arranged for standard-attention models.

Here’s what changed:
- Best-fit-decreasing packed examples now stay in separate batch rows, rather than being flattened into one long row per packed batch.
- The new layout produced step timings on AWS Trainium2, where the previous layout crashed.
- Hugging Face merged the TRL change on November 13, 2026.

One catch: The halved step time comes from a reported Qwen3 H100 test, and the layout change applies to standard-attention models.

Why care? The layout change both halved the reported Qwen3 H100 step time and produced timings on Trainium2, where the old layout crashed.

Sources: [runtimewire.com](https://runtimewire.com/article/hugging-face-trl-packed-batches-rows-h100-trainium)