David van Leeuwen

icon indicating an email [email protected]

icon company @SpraakLab icon location Nijmegen icon location Educated in experimental physics. Currently researcher in speech technology. Interested in the Julia programming language.

Results 14 issues of David van Leeuwen

OOV words with small LM

Hello, We have an application with a very small vocabulary (~100 words). With an almost trivial bigram model (as kenlm seems not to be able to make a unigram model),...

Update documentation

1 comment

Also figure out a way to use a proper documentation environment, rather than just the [README](https://github.com/davidavdav/NamedArrays.jl/blob/master/README.md), which is pretty outdated.

Functionality outside Base should be removed / deferred

With Julia 1.0, some library stuff has moved out of julia Base. `NamedArrays` should probably lazy-load support for this stuff only when it is used, for instance with [Requires](https://github.com/MikeInnes/Requires.jl). First...

Understand if it is possible to use own checkpoint from training as model file

1 comment

Hello, I've been busy with the default fairseq `examples/speech_recognition/infer.py` and also this repo's `recognize.py`, to see if it is possible to run inference using a model we made ourselves by...