Pytorch model ingest via Torchscript

  • The OctoML Platform now supports the upload of PyTorch models that have been exported to Torchscript.

  • Users can package and accelerate these models using any available engine, including TVM, ONNX-RT, and PyTorch-RT.

  • You can find an end-to-end tutorial for optimizing a PyTorch model, including conversion to Torchscript on our developer docs page.

SDK updates

  • We have renamed "octomize" to "accelerate" in our latest SDK version. Be sure to make the necessary updates when you upgrade.

Did this answer your question?