Intel Pretrained Models

AI model by Intel

Accelerate AI development with pre-optimized foundational models for Intel hardware.

A collection of foundational and optimized AI models for computer vision, NLP, and speech, designed to accelerate development on Intel hardware. These models provide a customizable starting point for businesses to build and deploy high-performance AI solutions.

Model specs

The full spec card (context, pricing per token, modalities and more) on AntSpace. Sign up free to see the full spec card.

Languages
English

Category: Machine Learning

Sizes and variants
Foundational, fine-tuned, and optimized models
Model Types
Foundational, Fine-tuned, Optimized
Target Hardware
Intel CPUs, GPUs, and Accelerators
Supported Frameworks
OpenVINO Toolkit and popular AI frameworks
Primary Domains
Computer Vision, NLP, Speech Recognition
Access Method
Intel AI Software Catalog and Developer Zone
  • Foundational models for computer vision, NLP, and speech
  • Optimized for Intel CPUs, GPUs, and AI accelerators
  • Seamless integration with OpenVINO toolkit
  • Open source starting points for custom fine-tuning

Overview

Intel Pretrained Models are a collection of foundational and optimized AI models designed for developers and data scientists to accelerate the AI pipeline. These models provide a starting point for AI capabilities, allowing businesses to customize and fine-tune them for specific use cases in computer vision, natural language processing, and speech recognition. They are specifically optimized for Intel hardware infrastructure, including Xeon processors and Gaudi accelerators, to ensure high performance during inference. By using these pre-optimized models, organizations can simplify their AI workflow and reduce the need for specialized data science expertise. The models are designed to work seamlessly with the Intel Distribution of OpenVINO Toolkit for efficient deployment across multiple platforms.

Model Development and Training

  • Foundational Models: Intel provides open source foundational models that serve as a starting point for AI capabilities, allowing businesses to customize them for specific needs.
  • Fine-tuned Models Tuning: The catalog includes models that are already fine-tuned to fit specific needs of AI solutions more easily and quickly.
  • Hardware Optimization: Models are optimized to enhance performance when paired with Intel hardware infrastructure.
  • Accelerated Development: Prebuilt solution elements and models help aid and accelerate the model selection, development, and training phases.

Supported AI Domains

  • Computer Vision Support: The catalog includes models designed for computer vision tasks.
  • NLP Capabilities: Natural language processing models are available for tasks involving text analysis and generation.
  • Speech Recognition: The collection features models for automatic speech recognition tasks.

Deployment and Optimization

  • OpenVINO Integration: Models are designed to work with the Intel Distribution of OpenVINO Toolkit for efficient inference and deployment.
  • Workflow Simplification: Pretrained models help simplify the AI workflow and increase deployment flexibility for the capabilities being enabled.
  • Scalable Deployment: The software tools assist in making the deployment process efficient, scalable, and manageable.

Use cases

  • Developers use foundational models as a starting point to customize AI capabilities for specific business needs without building from scratch.
  • Data scientists fine-tune pretrained vision models to accurately detect specific patterns in image and video datasets.
  • Organizations deploy optimized NLP models on Intel Xeon processors to achieve high-performance inference for text analysis tasks.
  • Engineers integrate pretrained speech recognition models into AI pipelines to accelerate the development of voice-enabled applications.

Specifications

Sizes and variants
Foundational, fine-tuned, and optimized models
Model Types
Foundational, Fine-tuned, Optimized
Target Hardware
Intel CPUs, GPUs, and Accelerators
Supported Frameworks
OpenVINO Toolkit and popular AI frameworks
Primary Domains
Computer Vision, NLP, Speech Recognition
Access Method
Intel AI Software Catalog and Developer Zone

How it works

  1. Model Selection: Developers browse the Intel AI Software Catalog to find foundational or fine-tuned models that match their specific use case, such as vision or speech.
  2. Customization and Fine-tuning: The selected foundational models are customized and fine-tuned using specific business data to meet the unique requirements of the application.
  3. Optimization: Models are optimized using Intel software tools and frameworks to leverage specific hardware features like Intel Deep Learning Boost.
  4. Deployment and Scaling: The final models are deployed across Intel hardware infrastructure, including CPUs and GPUs, often using the OpenVINO toolkit for efficient inference.

FAQ

What are Intel Pretrained Models?

Intel Pretrained Models are open source foundational models that provide a starting point for AI capabilities, allowing businesses to customize and fine-tune them for specific needs rather than building from scratch.

What types of AI tasks do these models support?

The catalog covers various domains including computer vision, natural language processing (NLP), and automatic speech recognition.

How are these models optimized for performance?

These models are optimized for Intel hardware infrastructure to ensure high performance during inference and deployment, and they integrate with tools like the OpenVINO toolkit.

How do these models accelerate time to market?

By providing ready-to-use and pre-optimized starting points, these models help simplify the AI workflow, accelerate innovation, and reduce the time required for model selection and training.

Added by AntSpace from public sources (intel.com).

More AI models by Intel