We release our ELMo representations trained on many languages which helps us win the CoNLL 2018 shared task on Universal Dependencies Parsing according to LAS.
We use the same hyperparameter settings as Peters et al. (2018) for the biLM and the character CNN. We train their parameters on a set of 20-million-words data randomly sampled from the raw text released by the shared task (wikidump + common crawl) for each language. We largely based ourselves on the code of AllenNLP, but made the following changes:
- We support unicode characters;
- We use the sample softmax technique to make training on large vocabulary feasible (Jean et al., 2015). However, we use a window of words surrounding the target word as negative samples and it shows better performance in our preliminary experiments.
The training of ELMo on one language takes roughly 3 days on an NVIDIA P100 GPU.
The Simplified Chinese model was trained on xinhua proportion of Chinese gigawords-v5.
- python 3.6
- pytorch 0.4
- other requirements from allennlp
First, after unzip the model, please change the "config_path"
field in ${lang}.model/config.json
to ${project_home}/configs/cnn_50_100_512_4096_sample.json
.
Then, prepare your input file in the conllu format, like
1 Sue Sue _ _ _ _ _ _ _
2 likes like _ _ _ _ _ _ _
3 coffee coffee _ _ _ _ _ _ _
4 and and _ _ _ _ _ _ _
5 Bill Bill _ _ _ _ _ _ _
6 tea tea _ _ _ _ _ _ _
Fileds should be separated by '\t'
. We only use the second column and space (' '
) is supported in
this field (for Vietnamese, a word can contains spaces).
Do remember tokenization!
When it's all set, run
python src/gen_elmo.py test \
--input_format conll \
--input /path/to/your/input \
--model /path/to/your/model \
--output_prefix /path/to/your/output \
--output_format hdf5 \
--output_layer -1
It will dump an hdf5 encoded dict
onto the disk, where the key is '\t'
separated
words in the sentence and the value is it's 3-layer averaged ELMo representation.
You can also dump the cnn encoded word with --output_layer 0
,
the first layer of the LsTM with --output_layer 1
and the second layer
of the LSTM with --output_layer 2
.
We are actively changing the interface to make it more adapted to the
AllenNLP ELMo and more programmatically friendly.
Please run
python src/biLM.py train -h
to get more details about the ELMo training. However, we
need to add that the training process is not very stable.
In some cases, we end up with a loss of nan
. We are actively working on that and hopefully
improve it in the future.
If our ELMo gave you nice improvements, please cite us.
@InProceedings{che-EtAl:2018:K18-2,
author = {Che, Wanxiang and Liu, Yijia and Wang, Yuxuan and Zheng, Bo and Liu, Ting},
title = {Towards Better {UD} Parsing: Deep Contextualized Word Embeddings, Ensemble, and Treebank Concatenation},
booktitle = {Proceedings of the {CoNLL} 2018 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies},
month = {October},
year = {2018},
address = {Brussels, Belgium},
publisher = {Association for Computational Linguistics},
pages = {55--64},
url = {http://www.aclweb.org/anthology/K18-2005}
}
- Bo Zheng <[email protected]>