New Generative AI Extension for Altair RapidMiner Studio (AI Studio) Now Available!

Options
David_A
edited June 18 in Help

We have a new Generative AI extension for Altair RapidMiner Studio (now called Altair AI Studio) and AI Hub. This extension is available to all licensed users at no additional cost. 

Download the Generative AI extension here: 

Find the user documentation here:  
 
https://docs.rapidminer.com/latest/studio/generative-ai/ 

Please be sure to follow the instructions in the documentation carefully. There are multiple dependencies that must fulfilled to make the extension work properly. 

The Generative Models extension (aka Generative AI extension) offers access to large language models (LLM) from Huggingface and OpenAI as well as finetuning of those models. It offers the following features around large language models (LLM): 

  • New use cases, including data enrichment, that go well beyond traditional machine learning tasks like classification and regression. 

  • Access to all LLMs from the 300,000+ models on Huggingface

  • Access to public models, including ChatGPT. 

  • Easy-to-use finetuning capabilities, including finetuning of the 175 billion parameter models from OpenAI. 

  • Finetuning of local models to secure privacy or IP. 

  • Ability to make generative AI a part of your analytics workflows for better deployments and orchestration. 

  • Innovative prompting language for flexible data injection. 

  • Zero shot classification for training-free classification tasks. 


Tagged:

Answers

  • Exciting news! The introduction of the new Generative AI Extension for Altair RapidMiner Studio, AI Studio, is a game-changer. This innovative addition is set to elevate data science and analytics capabilities, providing users with enhanced tools and possibilities. Looking forward to exploring the endless potential it brings to the table.
  • I am trying the extension, however, I am getting error "datasets" not found.  How do I address the issue? 
  • Ingo
    Options
    Hi, can you share which operator you have used that produced that error message? Merci :-)
  • Thanks for the extension! I have browsed "RapidMiner Document - Generative AI" and have defined a plan to try it (OpenAI, HuggingFace, fine-tuning, and RAG). Some questions:

    1. Does the extension support any mechanisms to evaluate a model or RAG? For example, how good is an English-to-Dutch translation or RAG via RAGAS?
    2. Does fine-tuning support QLoRA?
    3. There are many vector databases now. Why does RapidMiner select Qdrant and Milvus? What are the technical/business reasons behind the decision?
    4. Do these vector database operators send data to clouds or save data on-premise?

    Also thanks for the video of "No Way Back".

    Regards
    Frank