John Bauer

When you say you updated to the latest dev branch, I'm not sure I believe that, because the lines around line 56 of pretrain.py should look like: ``` try: data...

can we get the bugfix for `nan` merged for mLSTM? even the unreleased 2.0.0 still hits `nan` when mLSTM is used as a dropin replacement for an LSTM layer in...

Thanks for looking into this! I am currently on vacation (which is why I haven't written back sooner about the features table) but I will give it a try this...

Thanks for taking care of this! I realized I had neglected to respond to the new form. It looks quite nice. It look like it allows for features incompatible with...

The primary reason for the temp dir is to make the results not permanently stick around. You can provide the `--save_output` flag to the `run_...` scripts and it should turn...

Hmm, I wouldn't say it was necessary to close the issue. There are definitely some actionable items here, including finding a way to get rid of the perl dependency. It...

Well, I can't promise this will fix everything in terms of training on Windows, but I converted the `conllu_to_text.pl` script to python. It works exactly the same on several datasets...

Alright, I can't promise that each individual NER, Sentiment, etc dataset is properly using utf-8 encoding on Windows, but the UD-based ones certainly are. If any NER models don't rebuild...

Thanks, this is a good observation. So what I'm hearing is that we need some way to pass `local_files_only` to the code path(s) that load the transformers, right? But probably...

Are you comfortable using branches? We made the `local_files_only` branch so that `download_method=None` now doesn't download from HF either, in addition to not downloading Stanza models. Only caveat is the...