# Hugging Face — Fit More and Train Faster With ZeRO via DeepSpeed and FairScale

- Company: Hugging Face (huggingface.co)
- Announced: 2021-01-19
- Category: capability-change
- Coverage: not counted
- Announcement: yes
- Group: announcements
- Source: https://huggingface.co/blog/zero-deepspeed-fairscale
- Record: https://forck.live/items/2271-fit-more-and-train-faster-with-zero-via-deepspeed-and-fairscale
- Subject: Platform

Hugging Face Trainer now supports DeepSpeed and FairScale ZeRO features for training large models faster and with more memory efficiency.

## Evidence

Verbatim from https://huggingface.co/blog/zero-deepspeed-fairscale:

> If you use the Hugging Face Trainer, as of transformers v4.2.0 you have the experimental support for DeepSpeed's and FairScale's ZeRO features.

---

Record: https://forck.live/items/2271-fit-more-and-train-faster-with-zero-via-deepspeed-and-fairscale
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md