David van Leeuwen

I am not sure it succeeded, but it might be that version 0.3.6 is available now. It turned out I already had the new testing infrastructure configured. On Wed, Mar...

This is a general Julia problem, I think. If packages `A` and `B` both define function `f` which isn't in Base, then these definitions might clash. I think you can...

I think what is returned is `binary`, so I have a little function to make kaldi.io a little easier to work with: ```python from kaldi.base import io def construct(cls, fname):...

This is a bummer --- I get a similar error `ReadOnlyMemoryError()` in a similar setup, using PyJulia in a `multiprocessing.Process` instance. Has anyone found a way around this, or should...

Small update: when I use `context = multiprocessing.get_context("spawn")` or `"forkserver"`, I am able to run a python-jl process without crashing. In a basic application I can run the python-jl process...

So... despite that we input a token sequence `input_ids` in the `generate()` function, the length of this is irrelevant in the encoder-decoder model, and the `max_new_tokens` in `generate()` only refers...

Ah thanks, @gante---I do appreciate the difficulty of choosing sensible parameter/variable names, the number of times I am refactoring names back and forth in my own code is quite scary!

Well, even in pure python the multiprocessing communication is difficult (but that is off topic). In combination with `asyncio` the Queues become quite complex, every kind of multiprocessing and `asyncio`...

For what it worth, in my case (for a completely unrelated project) I had to increase the memory for the JVM in `grade.properties` ``` org.gradle.jvmargs=-Xmx4096m -Dfile.encoding=UTF-8 ```

> I would be helped if there is a way to produce a model file that works with recognize.py from the original base model and a checkpoint. OK, I think...