regression.neuclir22-ru-dt-splade.yaml Maven / Gradle / Ivy

Go to download
---
corpus: neuclir22-ru-en-splade
corpus_path: collections/multilingual/neuclir22-ru-en-splade/

index_path: indexes/lucene-index.neuclir22-ru-en-splade
collection_class: JsonVectorCollection
generator_class: DefaultLuceneDocumentGenerator
index_threads: 8
index_options: -impact -pretokenized -storeRaw
index_stats:
  documents: 4627541
  documents (non-empty): 4627541

metrics:
  - metric: MAP
    command: bin/trec_eval
    params: -c -m map
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: true
  - metric: nDCG@20
    command: bin/trec_eval
    params: -c -m ndcg_cut.20
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: false
  - metric: J@20
    command: python -m pyserini.eval.trec_eval
    params: -c -m judged.20
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: false
  - metric: Recall@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: "[NeuCLIR 2022 (Russian): title (original English queries)](https://neuclir.github.io/)"
    id: title
    path: topics.neuclir22-en.splade.original-title.txt.gz
    qrel: qrels.neuclir22-ru.txt
  - name: "[NeuCLIR 2022 (Russian): desc (original English queries)](https://neuclir.github.io/)"
    id: desc
    path: topics.neuclir22-en.splade.original-desc.txt.gz
    qrel: qrels.neuclir22-ru.txt
  - name: "[NeuCLIR 2022 (Russian): desc+title (original English queries)](https://neuclir.github.io/)"
    id: desc_title
    path: topics.neuclir22-en.splade.original-desc_title.txt.gz
    qrel: qrels.neuclir22-ru.txt

models:
  - name: splade
    display: SPLADE
    params: -impact -pretokenized
    results:
      MAP:
        - 0.3745
        - 0.3294
        - 0.3559
      nDCG@20:
        - 0.4865
        - 0.4193
        - 0.4573
      J@20:
        - 0.3741
        - 0.3702
        - 0.3838
      Recall@1000:
        - 0.8538
        - 0.8376
        - 0.8513
  - name: splade+rm3
    display: +RM3
    params: -impact -pretokenized -rm3 -collection JsonVectorCollection
    results:
      MAP:
        - 0.3151
        - 0.2717
        - 0.3026
      nDCG@20:
        - 0.4154
        - 0.3552
        - 0.4000
      J@20:
        - 0.3469
        - 0.3219
        - 0.3518
      Recall@1000:
        - 0.8169
        - 0.7529
        - 0.7704
  - name: splade+rocchio
    display: +Rocchio
    params: -impact -pretokenized -rocchio -collection JsonVectorCollection
    results:
      MAP:
        - 0.3909
        - 0.3378
        - 0.3659
      nDCG@20:
        - 0.4836
        - 0.4243
        - 0.4604
      J@20:
        - 0.3816
        - 0.3680
        - 0.3846
      Recall@1000:
        - 0.8686
        - 0.8238
        - 0.8544