Jaremy Creechley
Jaremy Creechley
Oh, the compiler runs much faster when it's not using Rosetta! ❤️
Bummer, looks like adding the C flag `-mmacosx-version-min=10.3` or similar might help with the older MacOS versions. Some combination of: - https://stackoverflow.com/questions/2923502/what-does-macosx-version-min-imply - https://micro-manager.org/Building_Universal_Binaries_on_Mac_OS_X
> In regards to 1. I'm not sure it should be solved via conditional compilation: can lead to nasty and obscure application bugs, it also does not really make possible...
Ok, this is pretty easy. I can pull it from my code if needed. It's just the MPack ID and byte specifier `::float-32`.
It seems like a good way to handle it, but I'm not sure I follow all the details. Would this impose a performance hit on every user of the library?...
> Here CairoBackend etc. are just objects that contain the fields currently living in the variant branches. The signatures of the backend specific draw* procedures just gets an additional typedesc[Cairo/TikZ/...Backend]...
Ran into this too, running the `ex02_handwritten_digits_recognition.nim` example. I use MsgPack a lot, and so tried `msgpack4nim`. It blew up, but after reading this I dove into a bit more....
A note on the above, MsgPack does pretty well in size compared to pure JSON. The exported msgpack file from above is ~22MB (or 16MB when bzipped), or when converted...
I'm running into what looks to be incomplete saving of a trained model. Saving a fully trained `DemoNet` model (with 90+% accuracy) using the previously described `msgpack4nim` method then reloading...