libonnx

What it is

LibONNX is a lightweight and portable C99 library that enables efficient ONNX model inference on embedded systems, offering hardware acceleration support to enhance performance in constrained environments.

LibONNX stands out as a lightweight and portable solution specifically designed for embedded systems, offering hardware acceleration without the overhead of larger frameworks.

When it fits — and when it doesn’t

Known weaknesses

Each one links to the public issue, changelog or docs page it came from, where we have one.

1. 01

   Limited language support

   LibONNX is primarily a C99 library, which restricts its usability to developers familiar with or willing to use C99.

2. 02

   Complex setup for hardware acceleration

   Configuring the library to leverage hardware acceleration can be complex and may require deep knowledge of both the hardware and the library's configuration options.

3. 03

   Performance may vary on non-targeted hardware

   While optimized for embedded systems, performance might suffer when used in environments that differ significantly from those for which it was designed.

4. 04

   Smaller community and slower updates

   Being an open-source project focused on a niche area (embedded ONNX inference), the community may be smaller, leading to slower development cycles and fewer contributions compared to larger frameworks.

What it does well

• Lightweight and portable C99 library for ONNX inference.

• Hardware acceleration support for enhanced performance.

• Suitable for embedded systems with limited resources.

Pricing

Free tier

YesOpen source — free to use

Paid from

$0

Model

Flat rate

Enterprise plan

No

Pricing not independently confirmed — check the vendor’s page.