By leveraging ONNX, Semidynamics hardware is ready to run thousands of models available in the ONNX format in common repositories like HuggingFace, or the ONNX Model Zoo. Using the SMD ONNX-RT, you’ll be able to integrate any of these models out of the box in Semidynamics hardware. Additionally, Semidynamics is actively researching and developing the AI space and have therefore curated a small set of AI models that we use as technology demonstrators. We provide these in ONNX format in our model zoo available below.

  • Llama V2

    Source Hugging Face
    Category LLM
    Version Chat v2
    Parameters 7B
    Data type fp16
  • Yolo V10

    Source GitHub
    Category Object Detection
    Version v10-N (Nano)
    Parameters 2.3M
    Data type fp16
  • ViT (Vision Transformer) - fp16

    Source GitHub
    Category Image Classification
    Version Base patch16 224
    Parameters 86M
    Data type fp16
  • ViT (Vision Transformer) - fp32

    Source GitHub
    Category Image Classification
    Version Base patch16 224
    Parameters 86M
    Data type fp32
  • MobileNet

    Source GitHub
    Category Image Classification
    Version v2 1.0
    Parameters 400K
    Data type fp16
  • AlexNet

    Source GitHub
    Category CNN
    Version v12
    Parameters 60M
    Data type fp32