Alexander
Alexander
I got 0.73s\request in plain torchscript, basically a little less than in triton. And that doesn't include gpu to cpu memory copy
Personally I need the latest dependency fixes to avoid downgrading numpy, sot 0.2.2 is not really an option. There is a pre-release option in pypi which allows to post a...
I'm using poetry so building from sources will be a mess and broken env for me. If there were wheels at least, I could link to them.
Ok, thanks, but I would still suggest investigating the pre-release feature of pypi
Encountered this problem on Ubuntu and solved with this > sudo apt-get install libmysqlclient-dev
Seems like this can help > sudo apt-get install libmysqlclient-dev
same here with torch.jit.optimize_for_inference. 200mb turns to 6mb model which isn't loading
This ones aren't working on tg https://www.instagram.com/reel/CxMR9VJv2h3/?igshid=MzRlODBiNWFlZA== https://www.ddinstagram.com/reel/Cxj5HRoxIYh/?igshid=MzRlODBiNWFlZA== Seems it's working like a coinflip
After page reload it starts working. Strange. Does it need some time to fetch the video?
> > > Stop working.... > > > > > > Please provide more information like screenshot and the link you tried to use > > Choose any link from...