# Hugging Face — Convert Transformers to ONNX with Hugging Face Optimum

- Company: Hugging Face (huggingface.co)
- Announced: 2022-06-22
- Category: not stated
- Coverage: not counted
- Announcement: no
- Group: routine
- Source: https://huggingface.co/blog/convert-transformers-to-onnx
- Record: https://forck.live/items/2186-convert-transformers-to-onnx-with-hugging-face-optimum
- Subject: Platform
- Models affected: ALBERT, BART, BERT, DistilBERT, ELECTRA, GPT Neo, GPT-J, GPT-2, RoBERTa, T5, ViT, XLM, distilbert-base-uncased-finetuned-sst-2-english

To convert Hugging Face Transformers models to the ONNX format using the Optimum library, covering three methods: low-level torch.onnx, mid-level transformers.onnx, and high-level Optimum API.

## Evidence

Verbatim from https://huggingface.co/blog/convert-transformers-to-onnx:

> In this guide, you'll learn about: What is ONNX? What is Hugging Face Optimum? What Transformers architectures are supported? How can I convert a Transformers model (BERT) to ONNX?

---

Record: https://forck.live/items/2186-convert-transformers-to-onnx-with-hugging-face-optimum
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md