# Hugging Face Transformers v4.52.1 — Release & Upgrade Intelligence
- **Product:** Hugging Face Transformers (https://huggingface.co)
- **Version:** v4.52.1
- **Released:** May 20, 2025
- **Impact Score:** 7/10
- **Breaking Changes:** Source flagged
- **Canonical Analysis:** https://changelogs.us/transformers/updates/v4.52.1/

## Executive Summary (TL;DR)
Hugging Face Transformers v4.52.1. Source highlight: The Qwen2.5-Omni model is a unified multiple modalities model proposed in Qwen2.5-Omni Technical Report from Qwen team, Alibaba Group.

## Breaking Changes & Deprecations
- Breaking changes * 🚨 rm already deprecated pad_to_max_length arg  by @itazap in #37617

## New Features
- SAM-HQ only adds ~0.5% additional parameters compared to SAM
- Add AutoRound quantization support  by @wenhuach21 in #37393
- Add "selecting a quantization method" doc  by @DerekLiu35 in #37159
- Add GGUF support to Gemma3 Text backbone  by @Isotr0py in #37424
- Add Fast Image Processor for Perceiver  by @rootonchair in #37176
- Add Fast Image Processor for Flava  by @rootonchair in #37135

## Improvements
- Like SAM, the model performs better with input 2D points and/or input bounding boxes
- enhance require_deterministic_for_xpu (#37437)

## Bug Fixes
- Fix broken add-fast-image-processor CLI  by @yonigozlan in #37499
- Fast image processor for VitMatte added and bug in slow version fixed  by @henrikm11 in #37616
- 🚨🚨🚨 Fix forward of Dinov2ForImageClassification for models with registers  by @psandovalsegura in #37836
- fix flex attn when optional args aren't passed  by @winglian in #37327
- fix llama4 training  by @hiyouga in #37319
- Fix deepspeed with quantization  by @Cyrilvallez in #37324

## Upgrade Command
```bash
python -m pip install 'transformers==4.52.1'
```