experiments-distilbert_tasb.md

Pyserini: Reproducing DistilBERT KD TASB Results

This guide provides instructions to reproduce the DistilBERT KD TASB dense retrieval model on the MS MARCO passage ranking task, described in the following paper:

Sebastian Hofstätter, Sheng-Chieh Lin, Jheng-Hong Yang, Jimmy Lin, Allan Hanbury. Efficiently Teaching an Effective Dense Retriever with Balanced Topic Aware Sampling. SIGIR 2021.

Note that we often observe minor differences in scores between different computing environments (e.g., Linux vs. macOS). However, the differences usually appear in the fifth digit after the decimal point, and do not appear to be a cause for concern from a reproducibility perspective. Thus, while the scoring script provides results to much higher precision, we have intentionally rounded to four digits after the decimal point.

Dense retrieval, with brute-force index:

python -m pyserini.search.faiss \
  --index msmarco-v1-passage.distilbert-dot-tas_b-b256 \
  --topics msmarco-passage-dev-subset \
  --encoded-queries distilbert_tas_b-msmarco-passage-dev-subset \
  --output runs/run.msmarco-passage.distilbert-dot-tas_b-b256.tsv \
  --output-format msmarco \
  --batch-size 512 --threads 16

Replace --encoded-queries with --encoder sebastian-hofstaetter/distilbert-dot-tas_b-b256-msmarco for on-the-fly query encoding.

To evaluate:

python -m pyserini.eval.msmarco_passage_eval msmarco-passage-dev-subset \
  runs/run.msmarco-passage.distilbert-dot-tas_b-b256.tsv

Results:

#####################
MRR @10: 0.3444
QueriesRanked: 6980
#####################

We can also use the official TREC evaluation tool trec_eval to compute other metrics than MRR@10. For that we first need to convert runs and qrels files to the TREC format:

python -m pyserini.eval.convert_msmarco_run_to_trec_run \
  --input runs/run.msmarco-passage.distilbert-dot-tas_b-b256.tsv \
  --output runs/run.msmarco-passage.distilbert-dot-tas_b-b256.trec

python -m pyserini.eval.trec_eval -c -mrecall.1000 -mmap msmarco-passage-dev-subset \
  runs/run.msmarco-passage.distilbert-dot-tas_b-b256.trec

Results:

map                     all     0.3515
recall_1000             all     0.9771

Reproduction Log*

• Results reproduced by @lintool on 2021-05-28 (commit 102ed2)
• Results reproduced by @lintool on 2022-12-23 (commit 0c495c)
• Results reproduced by @lintool on 2023-01-10 (commit 7dafc4)
• Results reproduced by @lintool on 2024-10-07 (commit 3f7609)