reranking dataset

[nabsabraham]

hello! thank you for this IR dataset!
This excerpt below says there are 100 queries for reranking but on huggingface there are 5000 queries (but only 4991 unique queries). Am I looking in the wrong place for the rerank eval dataset?

Additionally, for a smaller 100-question IR/Rerank evaluation, the dataset contains one positive for each question and 99 false negatives extracted by hard-negative mining using BM25 and sentence vector models.

from datasets import load_dataset
ds = load_dataset("hotchpotch/JaCWIR", "eval")
df = ds['eval'].to_pandas()
print(df['query'].nunique(), len(df['query'])) 
# 4991 5000