roboflow / inference

A fast, easy-to-use, production-ready inference server for computer vision supporting deployment of many popular model architectures and fine-tuned models.
https://inference.roboflow.com
Other
1.15k stars 85 forks source link

Need to skip Yoloact inst-seg tests for the parallel GPU inference server #384

Closed bigbitbus closed 1 month ago

bigbitbus commented 2 months ago

Search before asking

Bug

When running the tests on Nvidia T4 via GH actions, such errors are encountered: https://github.com/roboflow/inference/actions/runs/8992522142

This is because the parallel GPU inference docker does not support Yoloact Inst Seg.

This code fixes the issue by skipping yoloact tests:

https://github.com/roboflow/inference/blob/2aa0127f5a6a59b7a641a2182875fedb1696ebce/tests/inference/integration_tests/batch_regression_test.py#L202

https://github.com/roboflow/inference/blob/2aa0127f5a6a59b7a641a2182875fedb1696ebce/tests/inference/integration_tests/regression_test.py#L409

After this, the GH action passes (example link below): https://github.com/roboflow/inference/actions/runs/9023206470

PR to make the test change: https://github.com/roboflow/inference/pull/379

Environment

Nvidia GPU container

Minimal Reproducible Example

https://github.com/roboflow/inference/actions/runs/8992522142

Additional

No response

Are you willing to submit a PR?

Yes, already submitted #379

bigbitbus commented 2 months ago

@PawelPeczek-Roboflow fixing the GPU test

PawelPeczek-Roboflow commented 2 months ago

👍

grzegorz-roboflow commented 1 month ago

Closing as this change was already added within #379