Fine Tuning Mistral v3.0 With Custom Data

Video is about Fine tuning Mistra v3 model with custom data.
Join Skool Community for $129:
www.skool.com/data-society-42...
Have questions or ideas, meet similar people?
join the discord : / discord
Don't fall behind the AI revolution, I can help integrate machine learning/AI into your company.
mosleh587084.typeform.com/to/...
Mistral Fine tuning: github.com/mistralai/mistral-...
Colab Notebook : colab.research.google.com/dri...
Mistral v3 is a new model that came out it has many benifits.
The Mistral v3.0 model brings significant advancements in AI technology with its new architectural features, including Sliding Window Attention and Grouped Query Attention (GQA), which enhance long-sequence processing and speed up inference. It includes improved instruction-tuned models for better chat interactions and supports Flash Attention 2 for faster execution.
The model also offers quantization to reduce memory usage, making it highly efficient. Available on the Hugging Face platform, Mistral v3.0 is optimized for diverse applications, ensuring robust performance and scalability, particularly through its partnership with Microsoft Azure.
All in all Mistral v3 makes a good LLM for Automated AI Agents, embeddings and other useful machine learning tasks.
Time stamps
0:00 Intro
1:10 Downloading Model
1:56 Preparing Data
4:08 Training Parameters
4:47 How to solve GPU memory problem
5:50 Inference

[@farazfitness]

And what if my data is not in that format because I have a few law judgements and it's not possible to format the data in that way

[@IR240474]

Thank Mosleh for a great video. Just subbed and watching your previous videos. Is there any chance you could share your colab with us, it's to save time typing it myself! hehe... Not saying I am lazy, just that I miss some lines, and I am not sure where to find where you got your code from. Thanks again for showing me a fast way to train a model. Take care.

[@moslehmahamud]

Hi,
Not worries just shared the colab there.

[@IR240474]

@@moslehmahamud You are very kind Mosleh! Going to check out the colab now, thanks again and i wish you 100,000 subs!

[@RabeeQasem]

thank you

[@user-ty2gg8vv6m]

Thanks for the video. But what models with what configuration could be trained with free tier gpu..? Maybe phi3 mini?

[@moslehmahamud]

I'll take a look good idea, but colab is the cheapest alternative in the market right now

[@user-ty2gg8vv6m]

@@moslehmahamud hmm, okay, thx!