---
title: "Discussion on \"LLaMA 3.2 90B VRAM: How Much Memory Does Fine-tuning Need?\" | Hashnode"
description: "Discussion on \"LLaMA 3.2 90B VRAM: How Much Memory Does Fine-tuning Need?\". Optimize LLaMA 3.2 90B model fine-tuning with cloud-based GPU Instances from Novita AI. Cost-effective solution for memory-intensive tasks."
url: "https://hashnode.com/posts/llama-32-90b-vram-how-much-memory-does-fine-tuning-need/679c4acc54f065484595e23c"
---

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

Discussion

# [LLaMA 3.2 90B VRAM: How Much Memory Does Fine-tuning Need?](https://novita.hashnode.dev/llama-32-90b-vram-how-much-memory-does-fine-tuning-need?utm_source=hashnode&utm_medium=feed)

Optimize LLaMA 3.2 90B model fine-tuning with cloud-based GPU Instances from Novita AI. Cost-effective solution for memory-intensive tasks.

Jan 31, 2025 · 7 min read · novita.hashnode.dev  
Read the article

-   [#llama3](https://hashnode.com/tag/llama3)
-   [#finetune](https://hashnode.com/tag/finetune)

## Written by

[NovitaAI](https://hashnode.com/@NovitaAI)

@NovitaAI

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

Deploy AI models effortlessly with our simple API. Build and scale on the most affordable, reliable GPU cloud.

## Responses

Markdown supported · ⌘↵ to post

No responses yet.
