Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- en
|
4 |
+
tags:
|
5 |
+
- retrieval
|
6 |
+
- document_expansion
|
7 |
+
datasets:
|
8 |
+
- irds:msmarco-passage
|
9 |
+
library_name: pyterrier
|
10 |
+
---
|
11 |
+
|
12 |
+
A Doc2Query model based on `t5-base` and trained on MS MARCO. This is a version of [the checkpoint released by the original authors](https://git.uwaterloo.ca/jimmylin/doc2query-data/raw/master/T5-passage/t5-base.zip), converted to pytorch format and ready for use in [`pyterrier_doc2query`](https://github.com/terrierteam/pyterrier_doc2query).
|
13 |
+
|
14 |
+
**Creating a transformer:**
|
15 |
+
|
16 |
+
```python
|
17 |
+
import pyterrier as pt
|
18 |
+
pt.init()
|
19 |
+
from pyterrier_doc2query import Doc2Query
|
20 |
+
doc2query = Doc2Query('macavaney/doc2query-t5-base-msmarco')
|
21 |
+
```
|
22 |
+
|
23 |
+
**Transforming documents**
|
24 |
+
|
25 |
+
```python
|
26 |
+
import pandas as pd
|
27 |
+
doc2query(pd.DataFrame([
|
28 |
+
{'docno': '0', 'text': 'Hello Terrier!'},
|
29 |
+
{'docno': '1', 'text': 'Doc2Query expands queries with potentially relevant queries.'},
|
30 |
+
]))
|
31 |
+
# docno text querygen
|
32 |
+
# 0 Hello Terrier! hello terrier what kind of dog is a terrier wh...
|
33 |
+
# 1 Doc2Query expands queries with potentially rel... can dodoc2query extend query query? what is do...
|
34 |
+
```
|
35 |
+
|
36 |
+
**Indexing transformed documents**
|
37 |
+
|
38 |
+
```python
|
39 |
+
doc2query.append = True # append querygen to text
|
40 |
+
indexer = pt.IterDictIndexer('./my_index', fields=['text'])
|
41 |
+
pipeline = doc2query >> indexer
|
42 |
+
pipeline.index([
|
43 |
+
{'docno': '0', 'text': 'Hello Terrier!'},
|
44 |
+
{'docno': '1', 'text': 'Doc2Query expands queries with potentially relevant queries.'},
|
45 |
+
])
|
46 |
+
```
|
47 |
+
|
48 |
+
**Expanding and indexing a dataset**
|
49 |
+
|
50 |
+
```python
|
51 |
+
dataset = pt.get_dataset('irds:vaswani')
|
52 |
+
pipeline.index(dataset.get_corpus_iter())
|
53 |
+
```
|
54 |
+
|
55 |
+
## References
|
56 |
+
|
57 |
+
- [Nogueira20]: Rodrigo Nogueira and Jimmy Lin. From doc2query to docTTTTTquery. https://cs.uwaterloo.ca/~jimmylin/publications/Nogueira_Lin_2019_docTTTTTquery-v2.pdf
|
58 |
+
- [Macdonald20]: Craig Macdonald, Nicola Tonellotto. Declarative Experimentation inInformation Retrieval using PyTerrier. Craig Macdonald and Nicola Tonellotto. In Proceedings of ICTIR 2020. https://arxiv.org/abs/2007.14271
|