Parameter to specify model artifact name from config.pbtxt file when using .pth extension

AI & Data Science Deep Learning (Training & Inference) TensorRT
,
1

Description

I am having a pytorch artifact named model.pth but the Triton server is looking only for model.pt file which is by default here pytorch_backend/libtorch.cc at 4a971e6b6789310609ca84cf1c532084c1c6edc9 · triton-inference-server/pytorch_backend · GitHub

So, how do I override this config from config.pbtxt file so that I can point it to my .pth file. Which parameter to use in config.pbtxt file to do this ?

2

Hi,
We recommend you to raise this query in TRITON Inference Server Github instance issues section.

Thanks!