claudios commited on
Commit
2270990
·
verified ·
1 Parent(s): c75b047

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +44 -0
README.md ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ arxiv: 2001.00059
4
+ pipeline_tag: fill-mask
5
+ tags:
6
+ - code
7
+ - cubert
8
+ ---
9
+
10
+ # CuBERT: Learning and Evaluating Contextual Embedding of Source Code
11
+
12
+ ## Overview
13
+ This model is the unofficial HuggingFace version of "[CuBERT](https://github.com/google-research/google-research/tree/master/cubert)". In particular, this version comes from [gs://cubert/20210711_Java/pre_trained_model_epochs_2__length_1024](https://console.cloud.google.com/storage/browser/cubert/20210711_Java/pre_trained_model_epochs_2__length_1024). It was trained 2021-07-11 for 2 epochs with a 1024 token context window on the Java BigQuery dataset. I manually converted the Tensorflow checkpoint to PyTorch and have uploaded it here. The [tokenizer](https://github.com/google-research/google-research/blob/master/cubert/python_tokenizer.py) has not been converted yet. All credit goes to Aditya Kanade, Petros Maniatis, Gogul Balakrishnan, and Kensen Shi.
14
+
15
+ The other versions are available here:
16
+
17
+ [cubert-20210711-Python-512](https://huggingface.co/claudios/cubert-20210711-Python-512/)
18
+
19
+ [cubert-20210711-Python-1024](https://huggingface.co/claudios/cubert-20210711-Python-1024/)
20
+
21
+ [cubert-20210711-Python-2048](https://huggingface.co/claudios/cubert-20210711-Python-2048/)
22
+
23
+ [cubert-20210711-Java-512](https://huggingface.co/claudios/cubert-20210711-Java-512/)
24
+
25
+ [cubert-20210711-Java-1024](https://huggingface.co/claudios/cubert-20210711-Java-1024/)
26
+
27
+ [cubert-20210711-Java-2048](https://huggingface.co/claudios/cubert-20210711-Java-2048/)
28
+
29
+
30
+ Citation:
31
+ ```bibtex
32
+ @inproceedings{cubert,
33
+ author = {Aditya Kanade and
34
+ Petros Maniatis and
35
+ Gogul Balakrishnan and
36
+ Kensen Shi},
37
+ title = {Learning and evaluating contextual embedding of source code},
38
+ booktitle = {Proceedings of the 37th International Conference on Machine Learning,
39
+ {ICML} 2020, 12-18 July 2020},
40
+ series = {Proceedings of Machine Learning Research},
41
+ publisher = {{PMLR}},
42
+ year = {2020},
43
+ }
44
+ ```