The viewer is disabled because this dataset repo requires arbitrary Python code execution. Please consider removing the loading script and relying on automated data support (you can use convert_to_parquet from the datasets library). If this is not possible, please open a discussion for direct help.

h-corpus 领域的 Retrieval 评价数据集。

Leaderboard

new/data_sample1k

Model NDCG@5 NDCG@10 NDCG@15 NDCG@20 NDCG@30
IYun-large-zh 66.70±27.29 59.67±26.05 56.69±25.36 56.58±25.32 57.97±25.48
acge_text_embedding 64.60±28.04 57.80±25.88 55.54±25.166 55.77±25.17 57.31±25.18
bce-embedding-base_v1 60.66±28.37 53.44±26.13 51.11±25.10 51.18±25.16 52.84±25.45
Dmeta-embedding 52.12±29.83 45.38±26.65 43.20±25.33 43.41±25.10 44.87±25.42
random 0.07±1.24 0.09±1.01 0.10±0.97 0.12±0.99 0.14±1.03

data_sample5k

Model NDCG@10
IYun-large-zh 38.75
tao-8k 38.37
bce-embedding-base_v1 35.81
acge-large-zh 34.26
gte-large-zh 33.07
PEG 24.82
Dmeta-embedding 23.45
Downloads last month
2