Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Python backend initialize 'model_repository' arg breaking change

See original GitHub issue

Description

I am upgrading triton from 21.05 to 22.06 and it seems that the python backend now receives the current models model.py path instead of the current models folder as it previously did. This is a breaking change for all currently deployed models that use the python backend and relied on this. Are there any tips for handling this?

I also don’t see this mentioned in the release notes or any other issues, not sure if I missed it somewhere. Is there any reason a new arg wasn’t just added to args instead of changing the expected value of an existing arg?

Issue Analytics

State:
Created a year ago
Comments:5 (2 by maintainers)

Top GitHub Comments

1reaction

erfaneshraticommented, Jul 28, 2022

i think this PR can solve this issue: https://github.com/triton-inference-server/python_backend/pull/180

0reactions

krishung5commented, Jul 29, 2022

Closing this issue as the PR https://github.com/triton-inference-server/python_backend/pull/180 has been merged in.

Top Results From Across the Web

GitHub - triton-inference-server/python_backend

The goal of Python backend is to let you serve models written in Python by ... In the initialize function, you are given...

How we improved our Python backend start-up time - Medium

We moved the initialization code from init files into helpers and only called them when launching the web server. Break down big files...

Morepath Documentation

Morepath is a Python WSGI microframework. It uses routing, but the routing is to models. Morepath is model-driven.

Morepath Documentation - Read the Docs

Morepath is an Python WSGI microframework. It uses routing, but the routing is to models. Morepath is model-driven.

Plugin — OpenVINO™ documentation — Version(latest)

oneDNN backend for Intel CPU devices. NVIDIA cuDNN for NVIDIA GPUs. The responsibility of Inference Engine Plugin: Initializes a backend and throw exception ......