search

triton-inference-server / dali_backend

The Triton backend that allows running GPU-accelerated data pre-processing pipelines implemented in DALI's python API.
https://docs.nvidia.com/deeplearning/dali/user-guide/docs/index.html
MIT License
123 stars 29 forks source link

Fix model counting #227

Closed banasraf closed 7 months ago

banasraf commented 7 months ago

The instance counting was bugged which caused multiple issues in CI.

Instead of just fixing it, I changed it to counting loaded models, instead of created pipelines because it better fits the semantics of release_after_unload option and might save us from unnecessary memory release, when pipelines are temporarily created and destroyed.

dali-automaton commented 7 months ago

CI MESSAGE: [13030483]: BUILD STARTED

dali-automaton commented 7 months ago

CI MESSAGE: [13034380]: BUILD STARTED

dali-automaton commented 7 months ago

CI MESSAGE: [13030483]: BUILD PASSED