Linxiao ZENG

> Do you have an example of a sequence that cannot be decoded with this mechanism ? I've tried a few and the joiner marker will join words even when...

@michelleqyhqyh Could you try running it with `--max_generator_batches 0`?

Since you are providing an external alignment file, that suggests you want to use that as external knowledge to guide the model training, a.k.a guided alignment training. Then you may...

Hello @ongzexuan, If you want to contribute to this, you can start by looking at the [code of how the vocabulary fields are built](https://github.com/OpenNMT/OpenNMT-py/blob/master/onmt/inputters/fields.py) and [what the field is](https://pytorch.org/text/stable/data.html#fields) (especially...

Ref: section 5 of the original paper for [Jointly Learning to Align and Translate with Transformer Models](https://arxiv.org/abs/1909.02074).

You should be able to find all repo/scripts necessary to replicate their experiment in the paper. Especially the link listed under the Experiments section 5.1.1 for stuff related to alignment...

> Even if we keep it, there are some major changes to implement because Fields are deprecated. > My suggestion is to switch completely to Dataloader from pytorch. SHoud be...

Hello @texttheater, Your provided tutorial link is not correct, but I assume you refer to the [quickstart](https://github.com/OpenNMT/OpenNMT-py/blob/master/docs/source/quickstart.md) as you mentioned "step". In the quickstart, you train a LSTM model by...

Hello @SoYoungCho , I haven't used that feature, but it should work after briefly read the code. As it runs in silence without logging, I recommend you to print the...

I want to use the `case_markup` feature with `sentencepiece` in order to decrease the vocab duplication caused by the case, not sure what's the best practice to do it. As...