agijsberts

Out of curiosity, any reason why it works currently for `std::vector` and not for `std::list`? The only mention I found of `vector` is the special treatment of stl sequence types...

Strange, to be sure I just triple checked and I only get an error on the list operation ``` l.emplace(l.begin(), 2) ``` and **not** on the equivalent vector operation ```...

Thanks a lot for getting to the bottom of this. And sorry, I should've specified that I was running this on Linux. > It's actually a bit smarter: implicit conversion...

I cannot guarantee that it's the same issue, but I see a lot of `Failed to materialize symbols` errors in cppyy's test suite. For example, I get similar errors when...

Given that I was baffled that nobody seemed to have reported such a basic problem earlier, I investigated this issue further and, not surprisingly, it is a bit more complicated...

Pinning `huggingface-hub==0.21.4` is a sensible stopgap in my opinion. I'll prepare a PR right away. Regarding the `model_kwargs` and without going into too much detail: a significant chunk of `DGMR.__init__`...

I have just pushed the possible `__init__` simplification in my fork https://github.com/agijsberts/skillful_nowcasting/tree/refactor_huggingface-hub. This already includes tests to make sure that serialization-deserialization returns an identical object for `DGMR`, `Discriminator`, etc. Unfortunately...