TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

A simple guide to fine-tuning Llama 2

280 点作者 samlhuillier将近 2 年前

8 条评论

nmitchko将近 2 年前
This is a pretty useless post. You could also follow the same 1000x tutorials about llama and use the already uploaded hugging face formats that are on hugging face...<p>Here are some actually useful links<p><a href="https:&#x2F;&#x2F;blog.ovhcloud.com&#x2F;fine-tuning-llama-2-models-using-a-single-gpu-qlora-and-ai-notebooks&#x2F;" rel="nofollow noreferrer">https:&#x2F;&#x2F;blog.ovhcloud.com&#x2F;fine-tuning-llama-2-models-using-a...</a><p><a href="https:&#x2F;&#x2F;huggingface.co&#x2F;meta-llama&#x2F;Llama-2-70b-hf" rel="nofollow noreferrer">https:&#x2F;&#x2F;huggingface.co&#x2F;meta-llama&#x2F;Llama-2-70b-hf</a><p><a href="https:&#x2F;&#x2F;huggingface.co&#x2F;meta-llama&#x2F;Llama-2-7b-hf" rel="nofollow noreferrer">https:&#x2F;&#x2F;huggingface.co&#x2F;meta-llama&#x2F;Llama-2-7b-hf</a>
评论 #36854599 未加载
评论 #36856324 未加载
评论 #36855838 未加载
评论 #36859243 未加载
评论 #36854384 未加载
zoogeny将近 2 年前
What I&#x27;d like to do is create a website where:<p>1. There is a list of open source fine-tuning datasets on millions of topics. Like, anime, lord of the rings, dnd, customer service responses, finance, code in many programming languages, children&#x27;s books, religions, philosophies, etc. I mean, on every topic imaginable sort of like a Wikipedia or Reddit of fine-tuning data sets.<p>2. Users can select one or more available datasets as well as upload their own private datasets<p>3. Users can turn-key fine-tune llama 2 or other pre-trained models<p>Right now, doing this kind of thing is way beyond the capability of the common user.
评论 #36855831 未加载
评论 #36855821 未加载
评论 #36856835 未加载
syntaxing将近 2 年前
Can someone share a good tutorial how to prepare the data? And for fine tuning, does a 3090 have enough VRAM? I want to do what the author mentioned by fine tuning the model on my personal data but I’m not sure how to prepare the data. I tried using vector search + LLM but I find the results very subpar when using a local LLM.
评论 #36853686 未加载
评论 #36854887 未加载
评论 #36853393 未加载
marcopicentini将近 2 年前
Anyone has calculate the break even point (as number of token per month) between self-hosted LLAMA and OpenAI GPT-3.5 API?
eachro将近 2 年前
I&#x27;ve veen a bit out of the loop on this area but would like to get back into it given how much has changed in the LLM landscape in the last 1-2 yrs. What models are small enough to play with on Collab? Or am I going to have to spin up my own gpu box on aws to be able to mess around with these models?
评论 #36854346 未加载
评论 #36855162 未加载
treprinum将近 2 年前
Is there any tutorial on how to use HuggingFace LLaMA 2-derived models? They don&#x27;t have checkpoint files of the original LLaMA and can&#x27;t be used by the Meta&#x27;s provided inference code, instead they use .bin files. I am only interested in Python code so no llama.cpp.
评论 #36854743 未加载
评论 #36855178 未加载
评论 #36857893 未加载
moneywoes将近 2 年前
Any fine tuning success stories? Or real world use cases
评论 #36858362 未加载
m00dy将近 2 年前
Which dataset would be good to fine-tune for developing sales assistant like chatbot ?
评论 #36855097 未加载