William Irwin-Harris

I'm also receiving this error when using a predict request with the `tensorflow/serving:latest` (2.9.1) Docker image. Observed this error for two different NLP models on TF hub. Fixed by switching...

> I'm facing the same issue and solved it with @williamih 's solution! This issue should be fixed in the future release. > > > I'm also receiving this error...

Thanks @bradgrantham-lunarg - sorry, was just trying to save time parsing through a lot of logs. There are `gfxrecon` lines, but (at least to me) I'm not seeing anything obviously...

I can also confirm that a trace file is being created, although I'd think it doesn't contain much, since the app doesn't finish starting up

A few further findings: 1. The `unassisted` memory tracking option works correctly on Quest; however, the performance impact is large and unfortunately is too slow to use the app normally....

Hi @panos-lunarg Thanks for looking into this - I'd missed that setting when looking at the docs originally, so it's good to know. Oddly, when I tried to reproduce the...

Thanks @panos-lunarg Logs here: [logs-gfxrecon-infinite-restores.txt](https://github.com/LunarG/gfxreconstruct/files/11702692/logs-gfxrecon-infinite-restores.txt)

That xHook library looks interesting - thanks for the suggestion @panos-lunarg. I have a couple of other thoughts/questions: 1. It would be fairly easy to modify our Vulkan code to...

Thanks for the info - makes sense. With `debug.gfxrecon.capture_android_trigger`, I've been setting it to `false` before starting the app, then setting it to `true` once the app is running. However,...

Thanks @panos-lunarg - I'll try the xHook patch out! I was thinking of using a trimmed capture with `unassisted` memory tracking as I thought that might reduce some of the...