Toolkit for allowing inference and serving with PyTorch on SageMaker. Dockerfiles used for building SageMaker Pytorch Containers are at https://github.com/aws/deep-learning-containers.
Apache License 2.0
131
stars
70
forks
source link
[Question] Using model.mar with built-in handler script #133
What did you find confusing? Please describe. Hi, I've recently used a torchserve export utility provided by the MMDetection library. This uses the torch model archiver to package the following files into a
model.mar
:Is it possible to use this
model.mar
directly with SM TorchServe without needing to pull out the handler script and reformat it?Additional context I've added mmdet into my
requirements.txt
so the dependencies are not a problem.The
mmdet_handler.py
script looks like this: