Skip to content
Shop
Search icon
Shop
AI SOFTWARE

Voyager® SDK Model Zoo: AI Models & Pipelines for Computer Vision

 

Model Zoo

The Voyager SDK comes with a Model Zoo, a catalog of state-of-the-art AI models and turnkey pipelines for real-world use cases including image classification, object detection, segmentation, keypoint detection, face recognition and other Computer Vision tasks. This freely accessible catalog is continuously updated and improved with support for the latest models, covering a broad range of market verticals and use cases. Importantly, developers can easily modify any of the offered models to work with their own datasets or make them fit better to their application requirements. In addition to access via the web, access to the Model Zoo is seamlessly integrated within the SDK enabling effortless importing of Model Zoo content to the application under development.

Go to our Model Zoo
Try Metis Today
Voyager SDK

Under the hood

The VoyagerTM SDK builds on industry-standard APIs and open-source frameworks, enhanced with advanced capabilities by Axelera® R&D team. Voyager’s Machine Learning compiler, built on the Apache TVM compiler framework, automates the compilation and optimization of models for the Metis AI Processing Unit. The compiler inputs models pretrained in industry-standard frameworks such as PyTorch, and outputs code tuned for Metis hardware. During compilation, the compiler quantizes the model using proprietary, state-of-the-art algorithms and partitions the model for optimal execution on Metis AIPU. Without manual intervention, the compiler generates code with an accuracy practically indistinguishable from the original model. Similarly, pipelines use the open source framework GStreamer and whilst the vast majority of users need not understand any of the internals of generated pipelines, the open nature of the stack allows expert users to customize the generated code for a specific use case.

Applications deployed with Voyager onto the Metis AIPU have the highest performance and energy efficiency, while retaining equivalent accuracy of the original FP32 model.

Read on
AI at the Edge: A fast, accurate and effortless journey with Voyager SDK
February 27, 2023




 Development machine: Python project with native framework support for PyTorch/Tensor Flow

Input form dataloader   Normalize   Scale & crop   Convert to tensor   Model   NMS   Calculate metric and visualise stream
 
  ↓ Deploy to the Edge  
  Inference host CPU: C++ optimized code with streamer and dma-buf for efficient buffer sharing and synchronization between computing elements  
  Input from camera or socket (h264 decode)   YUV-to-RGB   Normalize   Scale & crop   Convert to tensor   Model   NMS   C++/Python callback   Visualize   Output to display or socket  


  ↑ Optimize at the Edge   ↑   ↑   ↑      
  Host Media Accelerator:
Input from camera, YUV-to-RGB convert, scale and crop		         Axelera AI:
Optimized Model		   Customer application
code		   Host GPU:
OpenGL kernel