Improving AI in Math with Google’s Gemma and Microsoft’s Orcamath dataset??

Improving AI in Math with Google’s Gemma and Microsoft’s Orcamath dataset??

Have you ever wanted to finetune LLMs locally to make them better problem solvers?


Last month, I worked on enhancing the mathematical problem-solving abilities of Google's new open-source large language models, Gemma 2B and 7B, using Microsoft's latest dataset, Orcamath.


?? Project Details:

  • Goal: Enhance the ability of Gemma models to solve complex math problems more effectively.
  • Tools: Applied Gemma 2B and 7B models with the Orcamath dataset.
  • Results: The models now solve math problems with greater performance.


???? What I Did:

I fine-tuned the training settings of these models to better tackle the challenging math problems presented by Orcamath. This effort not only improved their performance but also shed light on potential uses of AI in educational tools.


You can view the entire project on my GitHub: Gemma Orcamath Finetuning Project


I'm keen to connect with anyone interested in AI, especially in how it can be used to improve education and other fields.


#AI #DeepLearning #MachineLearning #DataScience #Education #Google #Microsoft #OpenSource

Daniel Prince

Product Manager

11 个月

Hey Neil, that's great can i kindly know the overview of your system config. Gemma "allegedly" is focusing on mobile devices. i was not able to deploy the 2b-IT model on my local even with a graphics card!

回复
Aman Kumar

???? ???? I Publishing you @ Forbes, Yahoo, Vogue, Business Insider And More I Monday To Friday Posting About A New AI Tool I Help You Grow On LinkedIn I Connect on WhatsApp: +91 8235569237

11 个月

Absolutely fascinating topic! It's amazing to see how AI is pushing the boundaries of creativity. You have an amazing profile. Please add me to your network?Neil de la Fuente :)

回复

要查看或添加评论,请登录

Neil de la Fuente的更多文章

社区洞察

其他会员也浏览了