conditions: # BM25 - name: bm25.ar eval_key: miracl-v1.0-ar command: python -m pyserini.search.lucene --language ar --topics miracl-v1.0-ar-${split} --index miracl-v1.0-ar --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.4434 R@100: 0.8562 - split: dev scores: - nDCG@10: 0.4809 R@100: 0.8885 - name: bm25.bn eval_key: miracl-v1.0-bn command: python -m pyserini.search.lucene --language bn --topics miracl-v1.0-bn-${split} --index miracl-v1.0-bn --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.5122 R@100: 0.8934 - split: dev scores: - nDCG@10: 0.5079 R@100: 0.9088 - name: bm25.en eval_key: miracl-v1.0-en command: python -m pyserini.search.lucene --language en --topics miracl-v1.0-en-${split} --index miracl-v1.0-en --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3415 R@100: 0.7928 - split: dev scores: - nDCG@10: 0.3506 R@100: 0.8190 - name: bm25.es eval_key: miracl-v1.0-es command: python -m pyserini.search.lucene --language es --topics miracl-v1.0-es-${split} --index miracl-v1.0-es --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3030 R@100: 0.7020 - split: dev scores: - nDCG@10: 0.3193 R@100: 0.7018 - name: bm25.fa eval_key: miracl-v1.0-fa command: python -m pyserini.search.lucene --language fa --topics miracl-v1.0-fa-${split} --index miracl-v1.0-fa --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3270 R@100: 0.7139 - split: dev scores: - nDCG@10: 0.3334 R@100: 0.7306 - name: bm25.fi eval_key: miracl-v1.0-fi command: python -m pyserini.search.lucene --language fi --topics miracl-v1.0-fi-${split} --index miracl-v1.0-fi --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.5106 R@100: 0.8471 - split: dev scores: - nDCG@10: 0.5513 R@100: 0.8910 - name: bm25.fr eval_key: miracl-v1.0-fr command: python -m pyserini.search.lucene --language fr --topics miracl-v1.0-fr-${split} --index miracl-v1.0-fr --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.2152 R@100: 0.6601 - split: dev scores: - nDCG@10: 0.1832 R@100: 0.6528 - name: bm25.hi eval_key: miracl-v1.0-hi command: python -m pyserini.search.lucene --language hi --topics miracl-v1.0-hi-${split} --index miracl-v1.0-hi --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.4745 R@100: 0.9016 - split: dev scores: - nDCG@10: 0.4578 R@100: 0.8679 - name: bm25.id eval_key: miracl-v1.0-id command: python -m pyserini.search.lucene --language id --topics miracl-v1.0-id-${split} --index miracl-v1.0-id --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.4844 R@100: 0.9234 - split: dev scores: - nDCG@10: 0.4486 R@100: 0.9041 - name: bm25.ja eval_key: miracl-v1.0-ja command: python -m pyserini.search.lucene --language ja --topics miracl-v1.0-ja-${split} --index miracl-v1.0-ja --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3796 R@100: 0.8225 - split: dev scores: - nDCG@10: 0.3689 R@100: 0.8048 - name: bm25.ko eval_key: miracl-v1.0-ko command: python -m pyserini.search.lucene --language ko --topics miracl-v1.0-ko-${split} --index miracl-v1.0-ko --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.4279 R@100: 0.7572 - split: dev scores: - nDCG@10: 0.4190 R@100: 0.7831 - name: bm25.ru eval_key: miracl-v1.0-ru command: python -m pyserini.search.lucene --language ru --topics miracl-v1.0-ru-${split} --index miracl-v1.0-ru --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3153 R@100: 0.6464 - split: dev scores: - nDCG@10: 0.3342 R@100: 0.6614 - name: bm25.sw eval_key: miracl-v1.0-sw command: python -m pyserini.search.lucene --language sw --topics miracl-v1.0-sw-${split} --index miracl-v1.0-sw --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3356 R@100: 0.6499 - split: dev scores: - nDCG@10: 0.3826 R@100: 0.7008 - name: bm25.te eval_key: miracl-v1.0-te command: python -m pyserini.search.lucene --language te --topics miracl-v1.0-te-${split} --index miracl-v1.0-te --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.4814 R@100: 0.8077 - split: dev scores: - nDCG@10: 0.4942 R@100: 0.8307 - name: bm25.th eval_key: miracl-v1.0-th command: python -m pyserini.search.lucene --language th --topics miracl-v1.0-th-${split} --index miracl-v1.0-th --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.4629 R@100: 0.8768 - split: dev scores: - nDCG@10: 0.4838 R@100: 0.8874 - name: bm25.zh eval_key: miracl-v1.0-zh command: python -m pyserini.search.lucene --language zh --topics miracl-v1.0-zh-${split} --index miracl-v1.0-zh --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: train scores: - nDCG@10: 0.2018 R@100: 0.5541 - split: dev scores: - nDCG@10: 0.1801 R@100: 0.5599 - name: bm25.de eval_key: miracl-v1.0-de command: python -m pyserini.search.lucene --language de --topics miracl-v1.0-de-${split} --index miracl-v1.0-de --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: dev scores: - nDCG@10: 0.2262 R@100: 0.5724 - name: bm25.yo eval_key: miracl-v1.0-yo command: python -m pyserini.search.lucene --pretokenized --topics miracl-v1.0-yo-${split} --index miracl-v1.0-yo --output $output --batch 128 --threads 16 --bm25 --hits 1000 splits: - split: dev scores: - nDCG@10: 0.4059 R@100: 0.7325 # mdpr-tied-pft-msmarco - name: mdpr-tied-pft-msmarco.ar eval_key: miracl-v1.0-ar command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-ar-${split} --index miracl-v1.0-ar-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.4653 R@100: 0.8293 - split: dev scores: - nDCG@10: 0.4993 R@100: 0.8407 - name: mdpr-tied-pft-msmarco.bn eval_key: miracl-v1.0-bn command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-bn-${split} --index miracl-v1.0-bn-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.4362 R@100: 0.8045 - split: dev scores: - nDCG@10: 0.4427 R@100: 0.8193 - name: mdpr-tied-pft-msmarco.en eval_key: miracl-v1.0-en command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-en-${split} --index miracl-v1.0-en-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3986 R@100: 0.7779 - split: dev scores: - nDCG@10: 0.3938 R@100: 0.7675 - name: mdpr-tied-pft-msmarco.es eval_key: miracl-v1.0-es command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-es-${split} --index miracl-v1.0-es-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.4637 R@100: 0.8654 - split: dev scores: - nDCG@10: 0.4777 R@100: 0.8643 - name: mdpr-tied-pft-msmarco.fa eval_key: miracl-v1.0-fa command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-fa-${split} --index miracl-v1.0-fa-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.4882 R@100: 0.9092 - split: dev scores: - nDCG@10: 0.4800 R@100: 0.8980 - name: mdpr-tied-pft-msmarco.fi eval_key: miracl-v1.0-fi command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-fi-${split} --index miracl-v1.0-fi-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.4426 R@100: 0.7611 - split: dev scores: - nDCG@10: 0.4721 R@100: 0.7877 - name: mdpr-tied-pft-msmarco.fr eval_key: miracl-v1.0-fr command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-fr-${split} --index miracl-v1.0-fr-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.4372 R@100: 0.9268 - split: dev scores: - nDCG@10: 0.4352 R@100: 0.9154 - name: mdpr-tied-pft-msmarco.hi eval_key: miracl-v1.0-hi command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-hi-${split} --index miracl-v1.0-hi-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3685 R@100: 0.7780 - split: dev scores: - nDCG@10: 0.3830 R@100: 0.7755 - name: mdpr-tied-pft-msmarco.id eval_key: miracl-v1.0-id command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-id-${split} --index miracl-v1.0-id-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.2549 R@100: 0.5610 - split: dev scores: - nDCG@10: 0.2719 R@100: 0.5734 - name: mdpr-tied-pft-msmarco.ja eval_key: miracl-v1.0-ja command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-ja-${split} --index miracl-v1.0-ja-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.4342 R@100: 0.8211 - split: dev scores: - nDCG@10: 0.4390 R@100: 0.8254 - name: mdpr-tied-pft-msmarco.ko eval_key: miracl-v1.0-ko command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-ko-${split} --index miracl-v1.0-ko-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.4147 R@100: 0.7699 - split: dev scores: - nDCG@10: 0.4189 R@100: 0.7369 - name: mdpr-tied-pft-msmarco.ru eval_key: miracl-v1.0-ru command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-ru-${split} --index miracl-v1.0-ru-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3812 R@100: 0.7854 - split: dev scores: - nDCG@10: 0.4073 R@100: 0.7972 - name: mdpr-tied-pft-msmarco.sw eval_key: miracl-v1.0-sw command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-sw-${split} --index miracl-v1.0-sw-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.2973 R@100: 0.5761 - split: dev scores: - nDCG@10: 0.2990 R@100: 0.6158 - name: mdpr-tied-pft-msmarco.te eval_key: miracl-v1.0-te command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-te-${split} --index miracl-v1.0-te-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3723 R@100: 0.7698 - split: dev scores: - nDCG@10: 0.3557 R@100: 0.7619 - name: mdpr-tied-pft-msmarco.th eval_key: miracl-v1.0-th command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-th-${split} --index miracl-v1.0-th-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3451 R@100: 0.6728 - split: dev scores: - nDCG@10: 0.3578 R@100: 0.6783 - name: mdpr-tied-pft-msmarco.zh eval_key: miracl-v1.0-zh command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-zh-${split} --index miracl-v1.0-zh-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.5040 R@100: 0.9355 - split: dev scores: - nDCG@10: 0.5116 R@100: 0.9436 - name: mdpr-tied-pft-msmarco.de eval_key: miracl-v1.0-de command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-de-${split} --index miracl-v1.0-de-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: dev scores: - nDCG@10: 0.4895 R@100: 0.8983 - name: mdpr-tied-pft-msmarco.yo eval_key: miracl-v1.0-yo command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco --topics miracl-v1.0-yo-${split} --index miracl-v1.0-yo-mdpr-tied-pft-msmarco --output $output --batch 128 --threads 16 --hits 1000 splits: - split: dev scores: - nDCG@10: 0.4439 R@100: 0.8403 # mdpr-tied-pft-msmarco-ft-all - name: mdpr-tied-pft-msmarco-ft-all.ar eval_key: miracl-v1.0-ar command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-ar-${split} --index miracl-v1.0-ar-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.6954 R@100: 0.8542 - split: dev scores: - nDCG@10: 0.5782 R@100: 0.7953 - name: mdpr-tied-pft-msmarco-ft-all.bn eval_key: miracl-v1.0-bn command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-bn-${split} --index miracl-v1.0-bn-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.6823 R@100: 0.8646 - split: dev scores: - nDCG@10: 0.5804 R@100: 0.8480 - name: mdpr-tied-pft-msmarco-ft-all.en eval_key: miracl-v1.0-en command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-en-${split} --index miracl-v1.0-en-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3491 R@100: 0.5678 - split: dev scores: - nDCG@10: 0.2813 R@100: 0.5083 - name: mdpr-tied-pft-msmarco-ft-all.es eval_key: miracl-v1.0-es command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-es-${split} --index miracl-v1.0-es-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.2488 R@100: 0.4799 - split: dev scores: - nDCG@10: 0.2509 R@100: 0.4706 - name: mdpr-tied-pft-msmarco-ft-all.fa eval_key: miracl-v1.0-fa command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-fa-${split} --index miracl-v1.0-fa-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3809 R@100: 0.6899 - split: dev scores: - nDCG@10: 0.3836 R@100: 0.6863 - name: mdpr-tied-pft-msmarco-ft-all.fi eval_key: miracl-v1.0-fi command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-fi-${split} --index miracl-v1.0-fi-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.7738 R@100: 0.9081 - split: dev scores: - nDCG@10: 0.5694 R@100: 0.7984 - name: mdpr-tied-pft-msmarco-ft-all.fr eval_key: miracl-v1.0-fr command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-fr-${split} --index miracl-v1.0-fr-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.2989 R@100: 0.6197 - split: dev scores: - nDCG@10: 0.3010 R@100: 0.6005 - name: mdpr-tied-pft-msmarco-ft-all.hi eval_key: miracl-v1.0-hi command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-hi-${split} --index miracl-v1.0-hi-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3336 R@100: 0.6388 - split: dev scores: - nDCG@10: 0.3286 R@100: 0.6371 - name: mdpr-tied-pft-msmarco-ft-all.id eval_key: miracl-v1.0-id command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-id-${split} --index miracl-v1.0-id-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3321 R@100: 0.5492 - split: dev scores: - nDCG@10: 0.3462 R@100: 0.5841 - name: mdpr-tied-pft-msmarco-ft-all.ja eval_key: miracl-v1.0-ja command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-ja-${split} --index miracl-v1.0-ja-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.6378 R@100: 0.7950 - split: dev scores: - nDCG@10: 0.4999 R@100: 0.7451 - name: mdpr-tied-pft-msmarco-ft-all.ko eval_key: miracl-v1.0-ko command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-ko-${split} --index miracl-v1.0-ko-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.5795 R@100: 0.7850 - split: dev scores: - nDCG@10: 0.4864 R@100: 0.7183 - name: mdpr-tied-pft-msmarco-ft-all.ru eval_key: miracl-v1.0-ru command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-ru-${split} --index miracl-v1.0-ru-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.6011 R@100: 0.8188 - split: dev scores: - nDCG@10: 0.3933 R@100: 0.6707 - name: mdpr-tied-pft-msmarco-ft-all.sw eval_key: miracl-v1.0-sw command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-sw-${split} --index miracl-v1.0-sw-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.8882 R@100: 0.9710 - split: dev scores: - nDCG@10: 0.6575 R@100: 0.8883 - name: mdpr-tied-pft-msmarco-ft-all.te eval_key: miracl-v1.0-te command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-te-${split} --index miracl-v1.0-te-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.8757 R@100: 0.9725 - split: dev scores: - nDCG@10: 0.7783 R@100: 0.9513 - name: mdpr-tied-pft-msmarco-ft-all.th eval_key: miracl-v1.0-th command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-th-${split} --index miracl-v1.0-th-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.7761 R@100: 0.9241 - split: dev scores: - nDCG@10: 0.5975 R@100: 0.8360 - name: mdpr-tied-pft-msmarco-ft-all.zh eval_key: miracl-v1.0-zh command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-zh-${split} --index miracl-v1.0-zh-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: train scores: - nDCG@10: 0.3446 R@100: 0.6608 - split: dev scores: - nDCG@10: 0.3575 R@100: 0.6725 - name: mdpr-tied-pft-msmarco-ft-all.de eval_key: miracl-v1.0-de command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-de-${split} --index miracl-v1.0-de-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: dev scores: - nDCG@10: 0.3219 R@100: 0.5990 - name: mdpr-tied-pft-msmarco-ft-all.yo eval_key: miracl-v1.0-yo command: python -m pyserini.search.faiss --encoder-class auto --encoder castorini/mdpr-tied-pft-msmarco-ft-all --topics miracl-v1.0-yo-${split} --index miracl-v1.0-yo-mdpr-tied-pft-msmarco-ft-all --output $output --batch 128 --threads 16 --hits 1000 splits: - split: dev scores: - nDCG@10: 0.5983 R@100: 0.8908 - name: bm25-mdpr-tied-pft-msmarco-hybrid.ar eval_key: miracl-v1.0-ar command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: - nDCG@10: 0.6259 R@100: 0.9173 - split: dev scores: - nDCG@10: 0.6729 R@100: 0.9405 - name: bm25-mdpr-tied-pft-msmarco-hybrid.bn eval_key: miracl-v1.0-bn command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: - nDCG@10: 0.6587 R@100: 0.9297 - split: dev scores: - nDCG@10: 0.6540 R@100: 0.9321 - name: bm25-mdpr-tied-pft-msmarco-hybrid.en eval_key: miracl-v1.0-en command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: - nDCG@10: 0.5347 R@100: 0.8772 - split: dev scores: - nDCG@10: 0.5488 R@100: 0.8815 - name: bm25-mdpr-tied-pft-msmarco-hybrid.es eval_key: miracl-v1.0-es command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: - nDCG@10: 0.6234 R@100: 0.9425 - split: dev scores: - nDCG@10: 0.6413 R@100: 0.9479 - name: bm25-mdpr-tied-pft-msmarco-hybrid.fa eval_key: miracl-v1.0-fa command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: - nDCG@10: 0.5890 R@100: 0.9433 - split: dev scores: - nDCG@10: 0.5935 R@100: 0.9374 - name: bm25-mdpr-tied-pft-msmarco-hybrid.fi eval_key: miracl-v1.0-fi command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: scores: - nDCG@10: 0.6164 R@100: 0.8506 - split: dev scores: - nDCG@10: 0.6716 R@100: 0.8949 - name: bm25-mdpr-tied-pft-msmarco-hybrid.fr eval_key: miracl-v1.0-fr command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: - nDCG@10: 0.5299 R@100: 0.9709 - split: dev scores: - nDCG@10: 0.5233 R@100: 0.9647 - name: bm25-mdpr-tied-pft-msmarco-hybrid.hi eval_key: miracl-v1.0-hi command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: - nDCG@10: 0.6217 R@100: 0.9059 - split: dev scores: - nDCG@10: 0.6157 R@100: 0.9115 - name: bm25-mdpr-tied-pft-msmarco-hybrid.id eval_key: miracl-v1.0-id command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: - nDCG@10: 0.4442 R@100: 0.7595 - split: dev scores: - nDCG@10: 0.4433 R@100: 0.7683 - name: bm25-mdpr-tied-pft-msmarco-hybrid.ja eval_key: miracl-v1.0-ja command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: - nDCG@10: 0.5795 R@100: 0.9082 - split: dev scores: - nDCG@10: 0.5757 R@100: 0.9036 - name: bm25-mdpr-tied-pft-msmarco-hybrid.ko eval_key: miracl-v1.0-ko command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: - nDCG@10: 0.5758 R@100: 0.8744 - split: dev scores: - nDCG@10: 0.6086 R@100: 0.8997 - name: bm25-mdpr-tied-pft-msmarco-hybrid.ru eval_key: miracl-v1.0-ru command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: - nDCG@10: 0.4921 R@100: 0.8494 - split: dev scores: - nDCG@10: 0.5323 R@100: 0.8738 - name: bm25-mdpr-tied-pft-msmarco-hybrid.sw eval_key: miracl-v1.0-sw command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: - nDCG@10: 0.4100 R@100: 0.6987 - split: dev scores: - nDCG@10: 0.4457 R@100: 0.7254 - name: bm25-mdpr-tied-pft-msmarco-hybrid.te eval_key: miracl-v1.0-te command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: - nDCG@10: 0.6000 R@100: 0.8717 - split: dev scores: - nDCG@10: 0.6021 R@100: 0.8569 - name: bm25-mdpr-tied-pft-msmarco-hybrid.th eval_key: miracl-v1.0-th command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: - nDCG@10: 0.5669 R@100: 0.8195 - split: dev scores: - nDCG@10: 0.5990 R@100: 0.8228 - name: bm25-mdpr-tied-pft-msmarco-hybrid.zh eval_key: miracl-v1.0-zh command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: train scores: - nDCG@10: 0.5209 R@100: 0.9576 - split: dev scores: - nDCG@10: 0.5254 R@100: 0.9587 - name: bm25-mdpr-tied-pft-msmarco-hybrid.de eval_key: miracl-v1.0-de command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: dev scores: - nDCG@10: 0.5643 R@100: 0.9482 - name: bm25-mdpr-tied-pft-msmarco-hybrid.yo eval_key: miracl-v1.0-yo command: python -m pyserini.fusion --runs ${bm25_output} ${mdpr_output} --output ${output} --method interpolation --alpha 0.5 --depth 1000 --k 1000 splits: - split: dev scores: - nDCG@10: 0.6114 R@100: 0.9496