PraisonAI Train: Train ANY AI Model with just 1 COMMAND

Discover how to streamline your AI model training process with just one command! Easiest Way to Train ANY Model & Save to Ollama. In this video, I'll demonstrate how you can fine-tune and save models using HuggingFace and Ollama seamlessly with PraisonAI Train. Whether you're an AI enthusiast or a professional, this tutorial will make your model training process faster and more efficient. 🚀

🔹 Key Takeaways:
Learn how to use praisonai train to fine-tune your model.
Integrate with Hugging Face and Ollama effortlessly.
Step-by-step guidance on exporting tokens, downloading necessary tools, and saving your models.
Tips for optimizing the training process using Linux.

🔹 Benefits:
Simplify complex AI training workflows.
Save time with efficient commands.
Enhance your AI development skills.

🔗 Links:

🔹 Steps Covered:
Exporting your Hugging Face token.
Downloading and setting up Ollama.
Integrating and saving models with Hugging Face and Ollama.
Fine-tuning the model using a single command.
Stay tuned till the end for some pro tips on optimizing your setup! Don't forget to like, subscribe, and hit the bell icon to get notified about more AI tutorials and tips. 😊

⏰ Timestamps:
0:00 - Introduction & Excitement
1:02 - Creating the Environment & Installing Packages
1:44 - Detailed Step-by-Step Guide
2:21 - Finalizing Setup and Training Model
2:58 - Adding Compatibility and Future Updates
4:41 - Uploading and Using the Model

#praisonai #train #llama3

[danpezarro]

I have been searching forever for a way to fine-tune in a single line that is compatible with Mac. Can you please make that addition? Love the content, thank you!

[NLPprompter]

bro just create the most useful tool... imagine you are in a situation where you need small model to be done tuned really fast to test something, doing this.... with already preset options would made it really simple. Thank you sir 🫡🫡

[redbaron]

Yes please add MLX compatibility as well! Awesome job as always! Thank you!

[APCMasteryPath]

Many thanks for your videos, they are quite simply and easy to follow.

I think it might be worthy to make a separate video to explain the environment preparation for unsloth, installing all dependencies and then the whole finetuning operation.

Keep up the good work.

Cheers.

[harikantipudi]

@mervin - What are the versions of CUDA, NVIDIA Driver and Torch . I am getting version compatable issues

I am using Driver version : 510.47.03 and CUDA Version - 11.6 and TeslaV100-SXM2

[arampamuk]

Hey Mervin Please add MLX for Apple Silicon

[yeahwearegood]

GREAT WORK

Make a video for Mac too

[AlvaroDalCanton]

Hi have you ever thought about adding to crewai and autogen also the Agency Swarm agent framework by Vrsen?

[sabamhr]

Awesome! Please add MLX for apple as well. Thank you

[garyb]

@mervinpraison where do you download praisonai from? also this is ubuntu, how about on windows laptop ?

[YisusStaloneZ]

Nice! Can you make another video for continued pretraining?

[osgoh]

Great tutorial and easy to follow! However, I found the message "WARNING: praisonai 0.0.52 does not provide the extra 'train'" when running pip install "praisonai[train]" . When do training by giving the command "praisonai train an error occurs with the message "File not found: train" It seems "train" file or module is missing. Anything missing here Mervin? Thanks.

[quantanovabusiness]

hi Mervin do you have the script changes to train using more than 1 of the sae gpu?

[sps]

Doing a Mac tutorial would be much appreciated

[siddani]

could you please add it for apple? this is awesome.

[GigaSwap]

Would be awesome to see How to do This on Mac

[Derick]

How do you do this without making it public

[shiki]

Why is this so damn inconsistent, Sometimes it decides to use 15 gb of ram others its using my entire 32 and causing the cloud VM to crash. Ridiculous

[charutapaliwal]

I am getting error message [WinError 2] The system cannot find the file specified when i am running the final execution command praisonai train --model --dataset yahma/alpaca-cleaned --hf abbbbdsds/llama3.1-instruct . @mervin praision or anyone, can you help here !!