|
<?xml version="1.0" encoding="UTF-8"?> |
|
<feed xmlns="http://www.w3.org/2005/Atom"> |
|
<link href="http://arxiv.org/api/query?search_query%3D%26id_list%3D1910.01108%26start%3D0%26max_results%3D1" rel="self" type="application/atom+xml"/> |
|
<title type="html">ArXiv Query: search_query=&id_list=1910.01108&start=0&max_results=1</title> |
|
<id>http://arxiv.org/api/GrfVMsjRJ5QfXT4borMr9RuzWhk</id> |
|
<updated>2023-06-09T00:00:00-04:00</updated> |
|
<opensearch:totalResults xmlns:opensearch="http://a9.com/-/spec/opensearch/1.1/">1</opensearch:totalResults> |
|
<opensearch:startIndex xmlns:opensearch="http://a9.com/-/spec/opensearch/1.1/">0</opensearch:startIndex> |
|
<opensearch:itemsPerPage xmlns:opensearch="http://a9.com/-/spec/opensearch/1.1/">1</opensearch:itemsPerPage> |
|
<entry> |
|
<id>http://arxiv.org/abs/1910.01108v4</id> |
|
<updated>2020-03-01T02:57:50Z</updated> |
|
<published>2019-10-02T17:56:28Z</published> |
|
<title>DistilBERT, a distilled version of BERT: smaller, faster, cheaper and |
|
lighter</title> |
|
<summary> As Transfer Learning from large-scale pre-trained models becomes more |
|
prevalent in Natural Language Processing (NLP), operating these large models in |
|
on-the-edge and/or under constrained computational training or inference |
|
budgets remains challenging. In this work, we propose a method to pre-train a |
|
smaller general-purpose language representation model, called DistilBERT, which |
|
can then be fine-tuned with good performances on a wide range of tasks like its |
|
larger counterparts. While most prior work investigated the use of distillation |
|
for building task-specific models, we leverage knowledge distillation during |
|
the pre-training phase and show that it is possible to reduce the size of a |
|
BERT model by 40%, while retaining 97% of its language understanding |
|
capabilities and being 60% faster. To leverage the inductive biases learned by |
|
larger models during pre-training, we introduce a triple loss combining |
|
language modeling, distillation and cosine-distance losses. Our smaller, faster |
|
and lighter model is cheaper to pre-train and we demonstrate its capabilities |
|
for on-device computations in a proof-of-concept experiment and a comparative |
|
on-device study. |
|
</summary> |
|
<author> |
|
<name>Victor Sanh</name> |
|
</author> |
|
<author> |
|
<name>Lysandre Debut</name> |
|
</author> |
|
<author> |
|
<name>Julien Chaumond</name> |
|
</author> |
|
<author> |
|
<name>Thomas Wolf</name> |
|
</author> |
|
<arxiv:comment xmlns:arxiv="http://arxiv.org/schemas/atom">February 2020 - Revision: fix bug in evaluation metrics, updated |
|
metrics, argumentation unchanged. 5 pages, 1 figure, 4 tables. Accepted at |
|
the 5th Workshop on Energy Efficient Machine Learning and Cognitive Computing |
|
- NeurIPS 2019</arxiv:comment> |
|
<link href="http://arxiv.org/abs/1910.01108v4" rel="alternate" type="text/html"/> |
|
<link title="pdf" href="http://arxiv.org/pdf/1910.01108v4" rel="related" type="application/pdf"/> |
|
<arxiv:primary_category xmlns:arxiv="http://arxiv.org/schemas/atom" term="cs.CL" scheme="http://arxiv.org/schemas/atom"/> |
|
<category term="cs.CL" scheme="http://arxiv.org/schemas/atom"/> |
|
</entry> |
|
</feed> |