Markus Hennerbichler
Markus Hennerbichler
This is also relevant to my use case, it looks like this got prioritised for the 2.3.0 release?
Thanks for looking at this. Here is a full example that you can run with `pytest`. For demonstration purpose I use the identity function as model :`torch.nn.Identity()` ```python import asyncio...
I am happy to contribute on this, just wanted to check if this isn't checked already and if the team considers it a good idea.
@Tabrizian I have found the root cause and it isn't just affecting ensemble models. https://github.com/triton-inference-server/core/pull/341. but potentially all sequence models that use implicit state and are accessed with the gRPC...
I have repro Python script which should work if you have pytorch and tritonserver installed: https://gist.github.com/HennerM/6ae28bc0360aae96bc5feb422aa1375c or if you already have the model repository setup that Jamie mentioned above then...
> Just to clarify - is the reproducible behavior the same prior to 23.10? Or are there different symptoms and triggers? We only have reproduced similar behaviour, that is multiple...
Thanks a lot. Definitely appreciate the help here. On (1) the review and test help will be good, as I struggled in the past to get the test setup working...
Sorry I am on UK hours, here is an updated script: https://gist.github.com/HennerM/cde5037442a2efa4238246cc5ccb7b51 I think I didn't leave enough time for the triton server to start, let me know if you...
I am having the same issue. Is there any particular reason openetelemetry-api is not compatible with >=7.0
Thanks @marcalff , I am just getting the CLA for my organisation through, could take a day or so