wehub-resource-sync
7 lines
296 B
Markdown
7 lines
296 B
Markdown
title: "ZeRO-Inference: 20X faster inference through weight quantization and KV cache offloading"
|
|
excerpt: ""
|
|
link: https://github.com/deepspeedai/DeepSpeedExamples/blob/master/inference/huggingface/zero_inference/README.md
|
|
date: 2023-09-12 00:09:00
|
|
tags: inference ZeRO quantization English
|
|
---
|