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

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

index_path: indexes/lucene-index.neuclir22-zh-en-splade
collection_class: JsonVectorCollection
generator_class: DefaultLuceneDocumentGenerator
index_threads: 8
index_options: -impact -pretokenized -storeRaw
index_stats:
  documents: 3179206
  documents (non-empty): 3179206

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 (Chinese): title (original English queries)](https://neuclir.github.io/)"
    id: title
    path: topics.neuclir22-en.splade.original-title.txt.gz
    qrel: qrels.neuclir22-zh.txt
  - name: "[NeuCLIR 2022 (Chinese): desc (original English queries)](https://neuclir.github.io/)"
    id: desc
    path: topics.neuclir22-en.splade.original-desc.txt.gz
    qrel: qrels.neuclir22-zh.txt
  - name: "[NeuCLIR 2022 (Chinese): 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-zh.txt

models:
  - name: splade
    display: SPLADE
    params: -impact -pretokenized
    results:
      MAP:
        - 0.3068
        - 0.3108
        - 0.3034
      nDCG@20:
        - 0.4233
        - 0.4299
        - 0.4236
      J@20:
        - 0.3851
        - 0.3825
        - 0.3961
      Recall@1000:
        - 0.7997
        - 0.7597
        - 0.7922
  - name: splade+rm3
    display: +RM3
    params: -impact -pretokenized -rm3 -collection JsonVectorCollection
    results:
      MAP:
        - 0.2843
        - 0.2631
        - 0.2820
      nDCG@20:
        - 0.3816
        - 0.3496
        - 0.3770
      J@20:
        - 0.3575
        - 0.3531
        - 0.3675
      Recall@1000:
        - 0.7546
        - 0.6969
        - 0.7481
  - name: splade+rocchio
    display: +Rocchio
    params: -impact -pretokenized -rocchio -collection JsonVectorCollection
    results:
      MAP:
        - 0.3151
        - 0.3108
        - 0.3092
      nDCG@20:
        - 0.4204
        - 0.4142
        - 0.4206
      J@20:
        - 0.3860
        - 0.3825
        - 0.3969
      Recall@1000:
        - 0.8038
        - 0.7623
        - 0.8067