aIbrahiim opened a new pull request, #41200:
URL: https://github.com/apache/beam/pull/41200
Fixes: #38782
Inference Python Dataflow benchmarks were failing in a few places
Image jobs still used the OpenImage input list and those files are gone, so
switch them to the ImageNet validation list
Table row streaming hit NOT_FOUND on preflight because the Pub/Sub
subscription was missing. Create the topic and subscription before submit
Captioning died on workers with cannot create weak reference to tuple.
Return a small model/processor object from load_model instead, and pin
torchvision for the GPU image
Streaming cancel was also messy. Only cancel and clean up Pub/Sub in
streaming, and wait for the cancel to finish before teardown
GPU captioning / OD / rightfit were still on the broken worker_accelerator
flag. Use the experiments accelerator + tensor_rt + Python 3.12, and bump disk
/ cut workers a bit
<!--
Thank you for your contribution! Follow this checklist to help us
incorporate your contribution quickly and easily:
- [ ] If this contribution is large, please file an Apache [Individual
Contributor License Agreement](https://www.apache.org/licenses/icla.pdf).
- [ ] Replace these instructions with a description of your change and why
you have made it.
- [ ] Mention the appropriate issue in your description (for example:
`addresses #123` or `fixes #456`).
- [ ] Update `CHANGES.md` with noteworthy changes.
See [CONTRIBUTING.md]() for more details.
-->
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
