Atreya Shankar

@ilkhem just wanted to chip in and say that fuzzy matching works in general for me. I tested the example with `smith2020anexamplecitation` being matched by `sm20` and it also works....

@thomasflynn918 I can try to help with this. > If I type M-x c-r-l at \ref{ I do see an autocomplete menu. That would mean that `company-reftex` works in general,...

> I have facing the same issue with my implementation even with the 2 letter, I guess company-complete is not triggering company-reftex-label ? I have assigned a keybinding to company-complete....

Unfortunately I use spacemacs so I can't provide just a few lines responsible in my config. However, my `company-diag` output shows: ``` company-backends: ((:separate company-reftex-labels company-reftex-citations) (:separate company-auctex-symbols company-auctex-environments company-capf...

Awesome, glad to hear it worked. Yes `company-backends` is IMO a complicated object and it requires some tweaking to get it to cooperate with other backends.

Hmm I am not sure why this happens, but good to know you found a solution that at least works. To diagnose the problem better, could you provide the following...

My bad, I should have mentioned I just needed everything except the completions; which is what makes the output so long. But in any case, these outputs show that everything...

I am working on a similar sequence tagging task for argument candidate identification. Essentially BERT or ALBERT would perform the encoding aspect of the raw input. Then, you would need...

I am testing albert with non-crf layers as decoders for a sequence labelling task where one classifies tokens into argumentation candidates. The repository is very rough at the moment but...

Just as a note, I used `bert-for-tf2` successfully in a sequence labelling task. This is detailed in the repository [here](https://github.com/atreyasha/sentiment-argument-mining/). The relevant information on the training and pre-processing can be...