-
Do you get the F0.5 score file based on official evaluation code? I try to run it and it is very slow.
-
https://github.com/singnet/language-learning/issues/67
-
Number of words in corpus after filtering with `min_word_count`:
1. Save cleaned word count to learn_&_test_stats.txt.
2. Add to corpus_stats.txt?
-
Currently, the statistics view is always refering to the state of the VC the moment the statistics button was pressed. When the VC changes, the statistics view needs to be closed and then reopened to …
Akron updated
5 years ago
-
#### Description
When building an LsiModel in distributed mode, one of the workers gets "stuck" while orthonormalizing the action matrix. This stalls the whole process of building the model, as the…
-
The pretrained MLM+TLM model achieves < 30% Pearson correlation with human scores. With CBOW at 60% and BERT at 86%, this score seems low for the STS-B task.
I am not sure if there is a mistake in …
-
Hello. I've successfully prepared the data:
```
jon@jon:~/PycharmProjects/news-translit-nmt-master/experiments$ bash prepare-data.sh
Preparing ThEn devset
Preparing EnTh devset
Preparing PeEn d…
-
Hi, I just installed `marian-nmt/marian` and ran the code ` ./run-me.sh` as default in traning-basics, I got error:
```
[2019-05-16 06:07:44] Error: CUDA error 2 'out of memory' - /tmp/marian/src…
-
I think we should add "ve" to the ENGLISH_STOP_WORDS constant.
The default tokenization splits ``I've"`` into ``['i', 've']``. ``have`` is in the stopwords, but not ``ve``.
Why do I care? I extrac…
-