{"id":1699,"date":"2024-11-19T11:38:14","date_gmt":"2024-11-19T11:38:14","guid":{"rendered":"https:\/\/nas01.tallpaul.net\/wordpress\/?p=1699"},"modified":"2024-11-19T11:41:46","modified_gmt":"2024-11-19T11:41:46","slug":"future-of-ai-models","status":"publish","type":"post","link":"https:\/\/nas01.tallpaul.net\/wordpress\/2024\/11\/future-of-ai-models\/","title":{"rendered":"The Future of AI Models and Granite 3.0"},"content":{"rendered":"\n

In a recent podcast episode, industry experts discussed the future of AI models and their development. This conversation provided valuable insights into how the AI landscape is evolving and what we can expect in the near future. Below are six key takeaways, plus three actions on how AI on IBM Power<\/strong> can help transform your organization:<\/p>\n\n\n\n

1. Pre-trained models will be abundant. <\/h2>\n\n\n\n

The speakers agreed that in the near future, there will be many pre-trained models available for various problems.<\/p>\n\n\n\n

This means that companies won’t need to create new models from scratch but can instead focus on customization and fine-tuning these existing models. This trend is underway, with numerous open-source models being developed and shared by researchers and organizations worldwide.<\/p>\n\n\n\n

2. Customization and fine-tuning will be the focus. <\/h2>\n\n\n\n

With pre-trained models readily available, companies can now concentrate on customizing and fine-tuning these models to suit their specific needs. This process involves adapting the pre-trained models to work with their unique data sets and use cases. By doing so, companies can leverage the power of AI without having to invest significant resources in developing new models from scratch.<\/p>\n\n\n\n

3. NVIDIA’s role in the customization process <\/h2>\n\n\n\n

NVIDIA can position itself as a provider of tools and resources for this customization process. <\/p>\n\n\n\n

The company has already made significant strides in this area by offering platforms like CUDA, cuDNN, and NVIDIA TensorRT, which enable developers to optimize their AI models for better performance and efficiency. By focusing on providing these tools, NVIDIA can help companies streamline their customization process and make the most of pre-trained models.<\/p>\n\n\n\n

CUDA<\/h3>\n\n\n\n

A parallel programming model and direct API that GPUs use to communicate with machine learning deployments. CUDA is a standard feature of all NVIDIA GPUs, including GeForce, Quadro, and Tesla.<\/p>\n\n\n\n

cuDNN<\/h3>\n\n\n\n

An inference library that provides kernels for Tensor Cores to improve performance on compute-bound operations. cuDNN is an optional dependency for TensorRT, and is only used to speed up a few layers.<\/p>\n\n\n\n

NVIDIA TensorRT<\/h3>\n\n\n\n

A library that runs on CUDA cores to optimize deep learning models for faster inference on NVIDIA GPUs. TensorRT can be 4\u20135 times faster than other methods for many real-time services and embedded applications.<\/p>\n\n\n\n

4. AI models as high fashion items<\/h2>\n\n\n\n

The speakers also discussed the idea of AI models becoming more like high fashion items, with prestige and value in their own right. This concept suggests that AI models will not just be commodities but also have a certain level of cachet. Companies may compete to use the most advanced or prestigious models, much like how they might choose high-end fashion brands for their products.<\/p>\n\n\n\n

5. IBM’s Granite models as an example.<\/h2>\n\n\n\n

IBM’s Granite models were mentioned as an example of this trend. These models are designed to be highly performant and efficient, making them attractive options for companies looking to fine-tune their AI solutions. By being found on NVIDIA platforms, these models can reach a wider audience and further solidify their status as prestigious AI tools.<\/p>\n\n\n\n

Granite 3.0 <\/h3>\n\n\n\n

Small Language Models (SLMs) for Enterprise<\/h4>\n\n\n\n

Workhorse’ models<\/p>\n\n\n\n

    \n
  1. Granite-3.0-8B-Instruct <\/li>\n\n\n\n
  2. Granite-3.0-2B-Instruct<\/li>\n<\/ol>\n\n\n\n

    Inference-efficient Mixture of Experts (MoE)  <\/h4>\n\n\n\n

    Extremely efficient inference and low latency<\/p>\n\n\n\n

      \n
    1. Granite-3.0-3B-A800M<\/li>\n\n\n\n
    2. Granite-3.0-1B-A400M<\/li>\n<\/ol>\n\n\n\n

      Guardrail models   <\/h4>\n\n\n\n

      Advance safe and trustworthy AI<\/p>\n\n\n\n

        \n
      1. Granite-Guardian-3.0-8B<\/li>\n\n\n\n
      2. Granite-Guardian-3.0-2B<\/li>\n<\/ol>\n\n\n\n

        Trusted<\/h3>\n\n\n\n

        Trained on trusted and governed data relevant to enterprise domains.<\/strong><\/p>\n\n\n\n