Memory increase when decoding exception occurs

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

121 stars 28 forks source link

Memory increase when decoding exception occurs #147

Closed ToTheMonn closed 2 years ago

ToTheMonn commented 2 years ago

Restart pipeline cause memory leak .

szalpal commented 2 years ago

@KrisChou ,

thank you for the information. I'd gladly look into this issue. May I ask for more information about this? What kind of restart you have in mind? Is it reloading the model using TritonClient API? Any information you could give me would be very helpful finding and solving the bug.

ToTheMonn commented 2 years ago

@szalpal This bug has been confirmed by DALI issue that will be fixed in next release.