I Taught an Open Source AI to Think Like OpenAI o3 (here’s how)
Ai Austin
0:00 / 0:00
I Taught an Open Source AI to Think Like OpenAI o3 (here’s how)
5 417 просмотров · 1 год назад
Ai Austin
21 тыс. подписчиков
5 417 просмотров · 1 год назад
Learn how to finetune an open source language model to create thought tokens before responding. This allows us to increase the intelligence of a language model, without increasing hardware requirements to chat with these local open source language models.
In this video I will go through:
Creating a dataset to train your thinking model.
Finetuning your own custom reasoning model in a Google Colab Notebook.
Creating a custom Ollama model from your model to run locally with maximum efficiency.
Code a Python script to interact with your custom model to mimic the functionality of ChatGPT's o1 & o3 models.
BECOME A PRO MEMBER: https://buymeacoffee.com/aiaustin
MY DISCORD SERVER: / discord
Google Colab Notebook for this video:
https://colab.research.google.com/dri...
#ai #artificialintelligence #tech