# Hugging Face Transformers v4.2.0 — Release & Upgrade Intelligence
- **Product:** Hugging Face Transformers (https://huggingface.co)
- **Version:** v4.2.0
- **Released:** Jan 13, 2021
- **Impact Score:** 7/10
- **Breaking Changes:** Source flagged
- **Canonical Analysis:** https://changelogs.us/transformers/updates/v4.2.0/

## Executive Summary (TL;DR)
Hugging Face Transformers v4.2.0. Source highlight: v4.2.0: LED from AllenAI, encoder-decoder templates, fast imports

## Breaking Changes & Deprecations
- Breaking changes There are no breaking changes between the previous version and this one.

## New Features
- Add flags to return scores, hidden states and / or attention weights in GenerationMixin #9150 (@SBrandeis)
- [Seq2Seq Templates] Add embedding scale to templates #9342 (@patrickvonplaten)
- [Seq2Seq Templates] Add forgotten imports to templates #9346 (@patrickvonplaten)

## Improvements
- Improve BERT-like models performance with better self attention #9124 (@jplu)
- Improve LayoutLM #9476 (@NielsRogge)
- Improve documentation coverage for Bertweet #9379 (@Qbiwan)
- Improve documentation coverage for Herbert #9428 (@Qbiwan)
- Improve documentation coverage for Phobert #9427 (@Qbiwan)

## Bug Fixes
- [LED Test] fix common inputs pt for flaky pt-tf led test #9459 (@SBrandeis, @patrickvonplaten)
- [TF Led] Fix flaky TF Led test #9513 (@patrickvonplaten)
- add tests for the new sharded ddp fairscale integration #9177 (@stas00)
- Added TF CTRL Sequence Classification #9151 (@spatil6)
- [trainer] apex fixes and tests #9180 (@stas00)
- Fix link to old NER fine-tuning script #9182 (@mrm8488)

## Upgrade Command
```bash
python -m pip install 'transformers==4.2.0'
```