regression.msmarco-v2-doc-segmented.yaml Maven / Gradle / Ivy

Go to download
---
corpus: msmarco-v2-doc-segmented
corpus_path: collections/msmarco/msmarco_v2_doc_segmented/

index_path: indexes/lucene-inverted.msmarco-v2-doc-segmented/
collection_class: MsMarcoV2DocCollection
generator_class: DefaultLuceneDocumentGenerator
index_threads: 24
index_options: -storeRaw
index_stats:
  documents: 124131414
  documents (non-empty): 124131414
  total terms: 24780918039

metrics:
  - metric: MAP@100
    command: bin/trec_eval
    params: -c -M 100 -m map
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: true
  - metric: MRR@100
    command: bin/trec_eval
    params: -c -M 100 -m recip_rank
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: true
  - metric: R@100
    command: bin/trec_eval
    params: -c -m recall.100
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: false
  - metric: R@1000
    command: bin/trec_eval
    params: -c -m recall.1000
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: false

topic_reader: TsvInt
topics:
  - name: "[MS MARCO V2 Doc: Dev](https://microsoft.github.io/msmarco/TREC-Deep-Learning.html)"
    id: dev
    path: topics.msmarco-v2-doc.dev.txt
    qrel: qrels.msmarco-v2-doc.dev.txt
  - name: "[MS MARCO V2 Doc: Dev2](https://microsoft.github.io/msmarco/TREC-Deep-Learning.html)"
    id: dev2
    path: topics.msmarco-v2-doc.dev2.txt
    qrel: qrels.msmarco-v2-doc.dev2.txt

models:
  - name: bm25-default
    display: BM25 (default)
    params: -bm25 -hits 10000 -selectMaxPassage -selectMaxPassage.delimiter "#" -selectMaxPassage.hits 1000
    results:
      MAP@100:
        - 0.1875
        - 0.1903
      MRR@100:
        - 0.1896
        - 0.1930
      R@100:
        - 0.6555
        - 0.6629
      R@1000:
        - 0.8542
        - 0.8549
# PRF regressions are no longer maintained for sparse judgments to reduce running times.
# (commenting out instead of removing; in case these numbers are needed, just uncomment and rerun.)
#
#  - name: bm25-default+rm3
#    display: +RM3
#    params: -bm25 -rm3 -collection MsMarcoV2DocCollection -hits 10000 -selectMaxPassage -selectMaxPassage.delimiter "#" -selectMaxPassage.hits 1000
#    results:
#      MAP@100:
#        - 0.1644
#        - 0.1681
#      MRR@100:
#        - 0.1660
#        - 0.1702
#      R@100:
#        - 0.6556
#        - 0.6566
#      R@1000:
#        - 0.8608
#        - 0.8639
#  - name: bm25-default+rocchio
#    display: +Rocchio
#    params: -bm25 -rocchio -collection MsMarcoV2DocCollection -hits 10000 -selectMaxPassage -selectMaxPassage.delimiter "#" -selectMaxPassage.hits 1000
#    results:
#      MAP@100:
#        - 0.1663
#        - 0.1690
#      MRR@100:
#        - 0.1681
#        - 0.1710
#      R@100:
#        - 0.6535
#        - 0.6618
#      R@1000:
#        - 0.8657
#        - 0.8642