etwk commited on
Commit
f2017c4
·
verified ·
1 Parent(s): a80c85b

update README

Browse files
Files changed (1) hide show
  1. README.md +28 -3
README.md CHANGED
@@ -1,9 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  The project contains indexes, datasets, checkpoints for RAG training and research.
2
 
3
  ## Sources
4
  - checkpoint: ColBERTv2 (https://downloads.cs.stanford.edu/nlp/data/colbert/colbertv2/colbertv2.0.tar.gz)
5
  - dataset: wiki_dpr (https://github.com/facebookresearch/DPR/blob/main/dpr/data/download_data.py)
6
- - indexes (TODO)
7
-
8
- * Didn't compress the index files to one archive because compress save less than 5% space and takes a long time.
9
 
 
1
+ ---
2
+ annotations_creators:
3
+ - no-annotation
4
+ language_creators:
5
+ - crowdsourced
6
+ language:
7
+ - en
8
+ license: cc-by-nc-4.0
9
+ multilinguality:
10
+ - multilingual
11
+ size_categories:
12
+ - 10M<n<100M
13
+ source_datasets:
14
+ - original
15
+ task_categories:
16
+ - fill-mask
17
+ - text-generation
18
+ task_ids:
19
+ - language-modeling
20
+ - masked-language-modeling
21
+ pretty_name: Wiki-DPR
22
+ tags:
23
+ - text-search
24
+ ---
25
+
26
  The project contains indexes, datasets, checkpoints for RAG training and research.
27
 
28
  ## Sources
29
  - checkpoint: ColBERTv2 (https://downloads.cs.stanford.edu/nlp/data/colbert/colbertv2/colbertv2.0.tar.gz)
30
  - dataset: wiki_dpr (https://github.com/facebookresearch/DPR/blob/main/dpr/data/download_data.py)
31
+ - indexes (https://github.com/ittia-research/check/tree/main/datasets/wiki_dpr)
32
+ - indexing config: `ColBERTConfig(nbits=2, doc_maxlen=220)`
33
+ - Didn't compress the index files to one archive because compress save less than 5% space and takes a long time.
34