---
title: "Discussion on \"Serving LLM's can be faster than you think !!\" | Hashnode"
description: "Discussion on \"Serving LLM's can be faster than you think !!\". This blog talks about how to enable speculative decoding to increase LLM token speed on existing infra"
url: "https://hashnode.com/posts/serving-llm-s-can-be-faster-than-you-think/69e1e54a9eff9dc9260ca9fc"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Serving LLM's can be faster than you think !!](https://pranavv.hashnode.dev/serving-llm-s-can-be-faster-than-you-think?utm_source=hashnode&utm_medium=feed)

This blog talks about how to enable speculative decoding to increase LLM token speed on existing infra

Apr 17 · 7 min read · pranavv.hashnode.dev  
Read the article

-   [#vllm](https://hashnode.com/tag/vllm)
-   [#gpu](https://hashnode.com/tag/gpu)
-   [#llm](https://hashnode.com/tag/llm)
-   [#ollama](https://hashnode.com/tag/ollama)
-   [#token](https://hashnode.com/tag/token)
-   [#speed](https://hashnode.com/tag/speed)

## Written by

[Pranav Prajapati](https://hashnode.com/@pranavv)

@pranavv · Nashik, India

  
[View profile](https://hashnode.com/@pranavv)

Data Science and Machine Learning engineer

Machine learning and Data Science | Open Source contributor | Python Developer

## Responses

Markdown supported · ⌘↵ to post

No responses yet.