Missing Simulator in SDK 2.1 – Clarification on Tutorial Compatibility

Technical Support
1

I’ve been following the tutorial Classification using Accelerator and noticed that the provided Python script imports:

Python

from memryx import NeuralCompiler, Simulator

Show more lines

However, in SDK 2.1 (installed via pip and confirmed with mx_nc --version), the Simulator class is no longer available. Attempting to run the script results in:

ImportError: cannot import name 'Simulator' from 'memryx'

Environment details:

  • Hardware: Orange Pi 5 Max

  • OS: Debian-based (similar to Ubuntu 22.04)

  • SDK installed via pip using:

    pip3 install --extra-index-url https://developer.memryx.com/pip memryx

  • SDK version: 2.1.0

Questions:

  1. Has Simulator been deprecated in SDK 2.1? If so, what is the recommended approach for running inference in Python?

  2. Are there plans to introduce a Python API for hardware inference (similar to NeuralAccelerator mentioned in some older examples)?

  3. Should tutorials be updated to use CLI tools like mx_infer or mx_bench for inference instead?

Any guidance or updated examples for SDK 2.1 would be greatly appreciated. Thanks for your support!

2

Hello @swells ,

  1. Has Simulator been deprecated in SDK 2.1? If so, what is the recommended approach for running inference in Python?

    Simulator API is not deprecated in 2.1 version, MemryX simulator is mainly supported on x86 based systems

    OrangePi is RK3588 ARM based system - MemryX simulator is not supported on ARM systems. please refer to platform support table here for more details : Install Software — MemryX Developer Hub

  2. Are there plans to introduce a Python API for hardware inference (similar to NeuralAccelerator mentioned in some older examples)?

    MemryX runtime package includes both Python and C++ APIs to run inference on our accelarator : Accelerator — MemryX Developer Hub

  3. Should tutorials be updated to use CLI tools like mx_infer or mx_bench for inference instead?

    All of our tutorial pages include a Download and Run section with ready-to-use scripts that can be executed directly from the CLI.
    In addition, you can find several real-time application examples here:MemryX Examples.
    Please follow the README inside each example folder for step-by-step instructions on running them.