Skip to content

TREC-ToT/bench

Repository files navigation

Benchmarks for TREC-ToT (2024)

The following benchmarks (& runs) are available. Results are for the dev2 set.:

Benchmark Runfiles NDCG@10 NDCG@1000 MRR@1000 R@1000
BM25 (k1=1, b=1.0) runs 0.0657 0.1033 0.0590 0.3600
Dense Retrieval (SBERT) (DR) runs 0.1040 0.1665 0.0901 0.5600

Note: The current repository only supports the 2024 version of the corpus/queries. For using the 2023 version, refer to the 2023 release, use tot23.py instead, and change the ir_dataset names used by baselines inside the code.

Initial setup

## optional: create new environment using py-env virtual-env
## pyenv virtualenv 3.8.11 trec-tot-benchmarks
# install requirements 
pip install ir_datasets sentence-transformers==2.2.2 pyserini==0.20.0 pytrec_eval faiss-cpu==1.6.5

2024

After downloading the files (see guidelines), set DATA_PATH to the folder which contains the uncompressed files s.t:

DATA_PATH/
  | train-2024
  | | - queries.jsonl
  | |  - qrel.txt
  | dev1-2024
  | | - queries.jsonl
  | | - qrel.txt
  | dev2-2024
  | | - queries.jsonl
  | | - qrel.txt
  | corpus.jsonl

Quick test to see if data is setup properly:

python tot.py

The command above should print the correct number of train/dev queries and the number of documents in the corpus, along with example queries and documents.

2023

After downloading the files (see guidelines), set DATA_PATH to the folder which contains the uncompressed files s.t:

DATA_PATH/
  | train
  | | - queries.jsonl
  | |  - qrel.txt
  | dev 
  | | - queries.jsonl
  | | - qrel.txt

Quick test to see if data is setup properly:

python tot.py

The command above should print the correct number of train/dev queries and the number of documents in the corpus, along with example queries and documents.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages