Patrick Huy

@emjotde memory mapping models and reusing workspaces sounds interesting. Is my understanding correct that this is currently not directly supported by `marian-decoder`?

Currently we are just running the provided `marian-server` as is and talk to it from a different process. With amun we used a python based server based on the included...

Thank you, that sounds great! I believe that the `marian-decoder` also has the same memory characteristics however and also requires more ram than `amun` I can verify tomorrow in case...

It's a "nematus" type RNN model with ~90k vocabulary size. The model file is ~600MB. I compiled 1.9.0 with a more recent version of Intel MKL, could that make a...

Yes. The GPU is also used and the process shows up in nvidia-smi

Looks very good! Marian 1.9.1 (https://github.com/marian-nmt/marian-dev/commit/adba021a5e6fee65870d16eae9d88319b07fa9bb) ``` ps aux | grep marian-server && free -h mt 16 0.8 5.2 7572348 853528 ? Sl 10:43 0:03 /marian/marian-server -c model/config.yml -w 256...

I think I can make using marian-dev work (for a while) - I've requested an approval but am rather confident that it will be possible We have discussed the logo...

Note: as a workaround I tried deactivating the customizer by overriding the Bean like this: ``` @ManagementContextConfiguration(ManagementContextType.CHILD) class ServletManagementChildContextConfiguration { @Bean @Primary @ConditionalOnClass(name = "org.apache.catalina.valves.AccessLogValve") public WebServerFactoryCustomizer tomcatAccessLogCustomizer() { return...