DylanJHJ/APRIL / logs /autoqrel-test.out
DylanJHJ's picture
download
raw
16.2 kB
=== Generating auto-qrel files for trec-dl-2019 ===
Generating: pool-40-systems-top10-rerank-judge
[direct] queries=43 judged_docs=2187 relevant=1874 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge/trec-dl-2019/autollmqrel.direct.txt
[thresholding@0.5] queries=43 judged_docs=2187 relevant=1874 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge/trec-dl-2019/autollmqrel.thresholding@0.5.txt
[rank@10] queries=43 judged_docs=2187 relevant=430 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge/trec-dl-2019/autollmqrel.rank@10.txt
[largest_gap@1] queries=43 judged_docs=2187 relevant=1097 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge/trec-dl-2019/autollmqrel.largest_gap@1.txt
[quantile_binary@0.75] queries=43 judged_docs=2187 relevant=1322 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge/trec-dl-2019/autollmqrel.quantile_binary@0.75.txt
[quantile_bucket] queries=41 judged_docs=2069 relevant=1422 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge/trec-dl-2019/autollmqrel.quantile_bucket.txt
[optimal_per_topic] queries=43 judged_docs=2187 relevant=1823 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge/trec-dl-2019/autollmqrel.optimal_per_topic.txt
[optimal_global] queries=43 judged_docs=2187 relevant=1874 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge/trec-dl-2019/autollmqrel.optimal_global.txt
[optimal_precision] queries=43 judged_docs=2187 relevant=1080 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge/trec-dl-2019/autollmqrel.optimal_precision.txt
[optimal_recall] queries=43 judged_docs=2187 relevant=1867 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge/trec-dl-2019/autollmqrel.optimal_recall.txt
Generating: pool-40-systems-top10-rerank-judge_expr
[direct] queries=43 judged_docs=2187 relevant=1615 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge_expr/trec-dl-2019/autollmqrel.direct.txt
[thresholding@0.5] queries=43 judged_docs=2187 relevant=1615 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge_expr/trec-dl-2019/autollmqrel.thresholding@0.5.txt
[rank@10] queries=43 judged_docs=2187 relevant=430 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge_expr/trec-dl-2019/autollmqrel.rank@10.txt
[largest_gap@1] queries=43 judged_docs=2187 relevant=1510 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge_expr/trec-dl-2019/autollmqrel.largest_gap@1.txt
[quantile_binary@0.75] queries=43 judged_docs=2187 relevant=562 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge_expr/trec-dl-2019/autollmqrel.quantile_binary@0.75.txt
[quantile_bucket] queries=43 judged_docs=2187 relevant=1949 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge_expr/trec-dl-2019/autollmqrel.quantile_bucket.txt
[optimal_per_topic] queries=43 judged_docs=2187 relevant=1753 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge_expr/trec-dl-2019/autollmqrel.optimal_per_topic.txt
[optimal_global] queries=43 judged_docs=2187 relevant=1782 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge_expr/trec-dl-2019/autollmqrel.optimal_global.txt
[optimal_precision] queries=43 judged_docs=2187 relevant=411 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge_expr/trec-dl-2019/autollmqrel.optimal_precision.txt
[optimal_recall] queries=43 judged_docs=2187 relevant=1819 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-judge_expr/trec-dl-2019/autollmqrel.optimal_recall.txt
Generating: pool-40-systems-top10-rerank-point
[direct] queries=43 judged_docs=2187 relevant=1611 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-point/trec-dl-2019/autollmqrel.direct.txt
[thresholding@0.5] queries=43 judged_docs=2187 relevant=1617 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-point/trec-dl-2019/autollmqrel.thresholding@0.5.txt
[rank@10] queries=43 judged_docs=2187 relevant=430 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-point/trec-dl-2019/autollmqrel.rank@10.txt
[largest_gap@1] queries=43 judged_docs=2187 relevant=1611 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-point/trec-dl-2019/autollmqrel.largest_gap@1.txt
[quantile_binary@0.75] queries=43 judged_docs=2187 relevant=580 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-point/trec-dl-2019/autollmqrel.quantile_binary@0.75.txt
[quantile_bucket] queries=43 judged_docs=2187 relevant=1946 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-point/trec-dl-2019/autollmqrel.quantile_bucket.txt
[optimal_per_topic] queries=43 judged_docs=2187 relevant=1697 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-point/trec-dl-2019/autollmqrel.optimal_per_topic.txt
[optimal_global] queries=43 judged_docs=2187 relevant=1767 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-point/trec-dl-2019/autollmqrel.optimal_global.txt
[optimal_precision] queries=43 judged_docs=2187 relevant=427 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-point/trec-dl-2019/autollmqrel.optimal_precision.txt
[optimal_recall] queries=43 judged_docs=2187 relevant=1806 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-point/trec-dl-2019/autollmqrel.optimal_recall.txt
Generating: pool-40-systems-top10-rerank-rankgpt
[direct] queries=43 judged_docs=2187 relevant=43 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-rankgpt/trec-dl-2019/autollmqrel.direct.txt
[thresholding@0.5] queries=43 judged_docs=2187 relevant=86 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-rankgpt/trec-dl-2019/autollmqrel.thresholding@0.5.txt
[rank@10] queries=43 judged_docs=2187 relevant=430 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-rankgpt/trec-dl-2019/autollmqrel.rank@10.txt
[largest_gap@1] queries=43 judged_docs=2187 relevant=43 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-rankgpt/trec-dl-2019/autollmqrel.largest_gap@1.txt
[quantile_binary@0.75] queries=43 judged_docs=2187 relevant=562 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-rankgpt/trec-dl-2019/autollmqrel.quantile_binary@0.75.txt
[quantile_bucket] queries=43 judged_docs=2187 relevant=1949 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-rankgpt/trec-dl-2019/autollmqrel.quantile_bucket.txt
[optimal_per_topic] queries=43 judged_docs=2187 relevant=1638 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-rankgpt/trec-dl-2019/autollmqrel.optimal_per_topic.txt
[optimal_global] queries=43 judged_docs=2187 relevant=2096 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-rankgpt/trec-dl-2019/autollmqrel.optimal_global.txt
[optimal_precision] queries=43 judged_docs=2187 relevant=419 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-rankgpt/trec-dl-2019/autollmqrel.optimal_precision.txt
[optimal_recall] queries=43 judged_docs=2187 relevant=1771 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-rankgpt/trec-dl-2019/autollmqrel.optimal_recall.txt
Generating: pool-40-systems-top10-rerank-setmaxheaptopk
[direct] queries=43 judged_docs=2187 relevant=43 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-setmaxheaptopk/trec-dl-2019/autollmqrel.direct.txt
[thresholding@0.5] queries=43 judged_docs=2187 relevant=86 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-setmaxheaptopk/trec-dl-2019/autollmqrel.thresholding@0.5.txt
[rank@10] queries=43 judged_docs=2187 relevant=430 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-setmaxheaptopk/trec-dl-2019/autollmqrel.rank@10.txt
[largest_gap@1] queries=43 judged_docs=2187 relevant=43 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-setmaxheaptopk/trec-dl-2019/autollmqrel.largest_gap@1.txt
[quantile_binary@0.75] queries=43 judged_docs=2187 relevant=562 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-setmaxheaptopk/trec-dl-2019/autollmqrel.quantile_binary@0.75.txt
[quantile_bucket] queries=43 judged_docs=2187 relevant=1949 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-setmaxheaptopk/trec-dl-2019/autollmqrel.quantile_bucket.txt
[optimal_per_topic] queries=43 judged_docs=2187 relevant=1799 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-setmaxheaptopk/trec-dl-2019/autollmqrel.optimal_per_topic.txt
[optimal_global] queries=43 judged_docs=2187 relevant=2168 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-setmaxheaptopk/trec-dl-2019/autollmqrel.optimal_global.txt
[optimal_precision] queries=43 judged_docs=2187 relevant=408 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-setmaxheaptopk/trec-dl-2019/autollmqrel.optimal_precision.txt
[optimal_recall] queries=43 judged_docs=2187 relevant=1921 -> /users/judylan1/APRIL/qrel-analysis/autoqrels/pool-40-systems-top10-rerank-setmaxheaptopk/trec-dl-2019/autollmqrel.optimal_recall.txt
=== nDCG@10 — human qrel baseline ===
run nDCG@10
bm25 0.5058
splade-v3 0.7270
nomicai-modernbert-embed 0.6960
qwen3-embed-600m 0.6693
colbert-small 0.7213
bm25-rerank-judge 0.6488
bm25-rerank-judge_expr 0.7018
bm25-rerank-point 0.6298
bm25-rerank-rankgpt 0.7442
bm25-rerank-setmaxheaptopk 0.7074
bm25-rerank-rankfirst 0.7284
bm25-rerank-rankzephyr 0.7378
splade-v3-rerank-judge 0.6625
splade-v3-rerank-judge_expr 0.7180
splade-v3-rerank-point 0.6417
splade-v3-rerank-rankgpt 0.7676
splade-v3-rerank-setmaxheaptopk 0.7495
splade-v3-rerank-rankfirst 0.7655
splade-v3-rerank-rankzephyr 0.7748
nomicai-modernbert-embed-rerank-judge 0.6567
nomicai-modernbert-embed-rerank-judge_expr 0.6999
nomicai-modernbert-embed-rerank-point 0.6479
nomicai-modernbert-embed-rerank-rankgpt 0.7809
nomicai-modernbert-embed-rerank-setmaxheaptopk 0.7330
nomicai-modernbert-embed-rerank-rankfirst 0.7551
nomicai-modernbert-embed-rerank-rankzephyr 0.7738
qwen3-embed-600m-rerank-judge 0.6533
qwen3-embed-600m-rerank-judge_expr 0.7074
qwen3-embed-600m-rerank-point 0.6457
qwen3-embed-600m-rerank-rankgpt 0.7498
qwen3-embed-600m-rerank-setmaxheaptopk 0.7142
qwen3-embed-600m-rerank-rankfirst 0.7383
qwen3-embed-600m-rerank-rankzephyr 0.7567
colbert-small-rerank-judge 0.6643
colbert-small-rerank-judge_expr 0.7153
colbert-small-rerank-point 0.6452
colbert-small-rerank-rankgpt 0.7656
colbert-small-rerank-setmaxheaptopk 0.7422
colbert-small-rerank-rankfirst 0.7717
colbert-small-rerank-rankzephyr 0.7683
=== nDCG@10 — auto-qrels ===
Writing: pool-40-systems-top10-rerank-judge.autollmqrel.direct.txt
Writing: pool-40-systems-top10-rerank-judge.autollmqrel.largest_gap@1.txt
Writing: pool-40-systems-top10-rerank-judge.autollmqrel.optimal_global.txt
Writing: pool-40-systems-top10-rerank-judge.autollmqrel.optimal_per_topic.txt
Writing: pool-40-systems-top10-rerank-judge.autollmqrel.optimal_precision.txt
Writing: pool-40-systems-top10-rerank-judge.autollmqrel.optimal_recall.txt
Writing: pool-40-systems-top10-rerank-judge.autollmqrel.quantile_binary@0.75.txt
Writing: pool-40-systems-top10-rerank-judge.autollmqrel.quantile_bucket.txt
Writing: pool-40-systems-top10-rerank-judge.autollmqrel.rank@10.txt
Writing: pool-40-systems-top10-rerank-judge.autollmqrel.thresholding@0.5.txt
Writing: pool-40-systems-top10-rerank-judge_expr.autollmqrel.direct.txt
Writing: pool-40-systems-top10-rerank-judge_expr.autollmqrel.largest_gap@1.txt
Writing: pool-40-systems-top10-rerank-judge_expr.autollmqrel.optimal_global.txt
Writing: pool-40-systems-top10-rerank-judge_expr.autollmqrel.optimal_per_topic.txt
Writing: pool-40-systems-top10-rerank-judge_expr.autollmqrel.optimal_precision.txt
Writing: pool-40-systems-top10-rerank-judge_expr.autollmqrel.optimal_recall.txt
Writing: pool-40-systems-top10-rerank-judge_expr.autollmqrel.quantile_binary@0.75.txt
Writing: pool-40-systems-top10-rerank-judge_expr.autollmqrel.quantile_bucket.txt
Writing: pool-40-systems-top10-rerank-judge_expr.autollmqrel.rank@10.txt
Writing: pool-40-systems-top10-rerank-judge_expr.autollmqrel.thresholding@0.5.txt
Writing: pool-40-systems-top10-rerank-point.autollmqrel.direct.txt
Writing: pool-40-systems-top10-rerank-point.autollmqrel.largest_gap@1.txt
Writing: pool-40-systems-top10-rerank-point.autollmqrel.optimal_global.txt
Writing: pool-40-systems-top10-rerank-point.autollmqrel.optimal_per_topic.txt
Writing: pool-40-systems-top10-rerank-point.autollmqrel.optimal_precision.txt
Writing: pool-40-systems-top10-rerank-point.autollmqrel.optimal_recall.txt
Writing: pool-40-systems-top10-rerank-point.autollmqrel.quantile_binary@0.75.txt
Writing: pool-40-systems-top10-rerank-point.autollmqrel.quantile_bucket.txt
Writing: pool-40-systems-top10-rerank-point.autollmqrel.rank@10.txt
Writing: pool-40-systems-top10-rerank-point.autollmqrel.thresholding@0.5.txt
Writing: pool-40-systems-top10-rerank-rankgpt.autollmqrel.direct.txt
Writing: pool-40-systems-top10-rerank-rankgpt.autollmqrel.largest_gap@1.txt
Writing: pool-40-systems-top10-rerank-rankgpt.autollmqrel.optimal_global.txt
Writing: pool-40-systems-top10-rerank-rankgpt.autollmqrel.optimal_per_topic.txt
Writing: pool-40-systems-top10-rerank-rankgpt.autollmqrel.optimal_precision.txt
Writing: pool-40-systems-top10-rerank-rankgpt.autollmqrel.optimal_recall.txt
Writing: pool-40-systems-top10-rerank-rankgpt.autollmqrel.quantile_binary@0.75.txt
Writing: pool-40-systems-top10-rerank-rankgpt.autollmqrel.quantile_bucket.txt
Writing: pool-40-systems-top10-rerank-rankgpt.autollmqrel.rank@10.txt
Writing: pool-40-systems-top10-rerank-rankgpt.autollmqrel.thresholding@0.5.txt
Writing: pool-40-systems-top10-rerank-setmaxheaptopk.autollmqrel.direct.txt
Writing: pool-40-systems-top10-rerank-setmaxheaptopk.autollmqrel.largest_gap@1.txt
Writing: pool-40-systems-top10-rerank-setmaxheaptopk.autollmqrel.optimal_global.txt
Writing: pool-40-systems-top10-rerank-setmaxheaptopk.autollmqrel.optimal_per_topic.txt
Writing: pool-40-systems-top10-rerank-setmaxheaptopk.autollmqrel.optimal_precision.txt
Writing: pool-40-systems-top10-rerank-setmaxheaptopk.autollmqrel.optimal_recall.txt
Writing: pool-40-systems-top10-rerank-setmaxheaptopk.autollmqrel.quantile_binary@0.75.txt
Writing: pool-40-systems-top10-rerank-setmaxheaptopk.autollmqrel.quantile_bucket.txt
Writing: pool-40-systems-top10-rerank-setmaxheaptopk.autollmqrel.rank@10.txt
Writing: pool-40-systems-top10-rerank-setmaxheaptopk.autollmqrel.thresholding@0.5.txt

Xet Storage Details

Size:
16.2 kB
·
Xet hash:
5b3765e33b0860c3a4902e9b9041ba81cc6c17176f43bd2a1afadb042e72ba7b

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.