# Hugging Face Transformers v5.3.0 — Release & Upgrade Intelligence
- **Product:** Hugging Face Transformers (https://huggingface.co)
- **Version:** v5.3.0
- **Released:** Mar 04, 2026
- **Impact Score:** 7/10
- **Breaking Changes:** Source flagged
- **Canonical Analysis:** https://changelogs.us/transformers/updates/v5.3.0/

## Executive Summary (TL;DR)
Hugging Face Transformers v5.3.0. Source highlight: <img width="1080" height="1080" alt="image" src="https://github.com/user-attachments/assets/33603f42-5435-421a-9641-baf72faacb22" />

## Breaking Changes & Deprecations
- Breaking changes Tensor parallelism (TP) support for dense and MoE decoder-only models has been fixed and stabilized, requiring users to update their TP configurations and conversion mappings accordingly.

## New Features
- Add eurobert (#39455) by @ArthurZucker in #39455
- Add VibeVoice ASR (#43625) by @ebezzam in #43625
- [Model] Add PP-DocLayoutV2 Model Support (#43018) by @zhang-prog in #43018
- Add OLMo Hybrid model (#43358) by @yanhong-lbh in #43358
- Add ModernVBERT models (#42504) by @paultltc in #42504
- Add ModernVBERT models (#42504) by @paultltc in #42504

## Improvements
- perf: Optimize SynthID logits processor batch index construction (#44172)
- Improve `has_similar_generate_outputs` assertions (#44166)

## Bug Fixes
- 🚨 fix + tests dense & MoE TP all reduce (decoder only) (#43722)
- :rotating_light: [`Ernie 4.5 VL Moe`] Fix up namings to vllm/sglang convention (#44299)
- [vllm + v5 fix] handle TokenizersBackend fallback properly for v5 (#44255)
- [higgs-audio-v2] fix sampling (#44386) by @eustlb in [#44386]
- fix(flaky): idefics generate cache flake (#44180) by @tarekziade in [#44180]
- Fix generation integration tests (#44225) by @zucchini-nlp in [#44225]

## Upgrade Command
```bash
python -m pip install 'transformers==5.3.0'
```