Release date: February 18, 2021
- Added support for replicating DPR retrieval results (Karpukhin et al. 2020).
- Improved support for TCT-ColBERT (Lin et al. 2020).
- Added interactive retrieval support for dense and hybrid techniques, example with TCT-ColBERT.
- Added integration test cases and replication guides for both TCT-ColBERT and DPR.
- Added replication guide for KILT BM25 baselines.
- Added support for accessing qrels from standard test collections.
- Improved query iterator for standard test collections.
- Improved built-in support for MS MARCO doc/passage evaluation and
trec_eval
inpyserini.eval
. - Updated documentation for spaCy and entity linking.
- Standardized dependencies, especially important for
transformers
,torch
,tensorflow
, andfaiss-cpu
.
Sorted by number of commits:
- Xueguang Ma (MXueguang)
- Jimmy Lin (lintool)
- Ronak Pradeep (ronakice)
- Kai Sun (KaiSun314)
- Larry Li (larryli1999)
- Jiarui Zhang (jrzhang12)
- Xinyu Mavis Liu (x389liu)
- Yuxuan Ji (yuxuan-ji)
- Marko Arezina (mrkarezina)
- Emily Ye (yemiliey)
Sorted by number of commits, according to GitHub:
- Jimmy Lin (lintool)
- Xueguang Ma (MXueguang)
- Johnson Han (x65han)
- Yuqi Liu (yuki617)
- Stephanie Hu (stephaniewhoo)
- Chris Kamphuis (Chriskamphuis)
- Zeynep Akkalyoncu Yilmaz (zeynepakkalyoncu)
- Xinyu Mavis Liu (x389liu)
- Pepijn Boers (PepijnBoers)
- Marko Arezina (mrkarezina)
- Ronak Pradeep (ronakice)
- Qing Guo (qguo96)
- Tommaso Teofili (tteofili)
- Hang Cui (HangCui0510)
- Dahlia Chehata (Dahlia-Chehata)
- Rodrigo Nogueira (rodrigonogueira4)
- Larry Li (larryli1999)
- Jiarui Zhang (jrzhang12)
- Kai Sun (KaiSun314)
- Tim Hatch (thatch)
- Yue Zhang (nsndimt)
- Alireza Mirzaeiyan (amirzaeiyan)
- Rakeeb Hossain (rakeeb123)
- Jerry Huang (jhuang265)
- Jeffrey Chen (JeffreyCA)
- Adam Yang (adamyy)
- Hector (Xinhai) Wei (HEC2018)
- Emily Ye (yemiliey)
- Yuxuan Ji (yuxuan-ji)