Dataset Viewer
Auto-converted to Parquet
query-id
stringlengths
1
6
corpus-id
stringlengths
1
6
score
stringclasses
1 value
19399
102236
1
19399
91901
1
19399
177507
1
19399
80798
1
19399
112990
1
19399
182056
1
19399
130867
1
19399
17821
1
19399
11810
1
19399
4501
1
19399
175150
1
19399
168653
1
19399
71259
1
19399
167405
1
19399
11476
1
19399
120630
1
19399
192720
1
19399
44459
1
19399
190304
1
19399
167240
1
19399
8545
1
19399
97772
1
19399
1338
1
19399
162608
1
19399
133105
1
19399
74680
1
19399
22082
1
19399
81484
1
19399
134461
1
19399
183003
1
19399
71266
1
19399
41583
1
19399
128973
1
19399
79668
1
19399
152106
1
19399
11589
1
19399
190191
1
19399
43474
1
19399
175596
1
19399
164832
1
19399
147712
1
19399
163800
1
19399
164657
1
19399
119668
1
19399
9053
1
19399
3288
1
19399
132622
1
19399
71154
1
19399
181598
1
19399
144405
1
19399
184525
1
19399
166773
1
19399
56449
1
19399
166173
1
19399
5588
1
19399
130948
1
19399
151807
1
19399
113234
1
19399
46609
1
19399
11023
1
19399
76371
1
19399
172448
1
19399
39838
1
19399
77835
1
19399
166397
1
5987
24519
1
57488
92203
1
97154
5913
1
97154
98854
1
97154
79081
1
97154
191853
1
97154
186539
1
97154
92267
1
97154
194528
1
97154
29552
1
97154
180660
1
97154
145369
1
97154
127072
1
97154
76161
1
97154
62257
1
97154
50686
1
97154
147063
1
97154
7841
1
97154
120709
1
97154
174143
1
97154
119530
1
97154
147905
1
97154
129431
1
97154
13093
1
97154
192402
1
97154
186518
1
97154
85125
1
97154
2964
1
97154
62692
1
97154
161894
1
97154
23421
1
97154
132833
1
97154
19847
1
97154
41889
1
97154
15609
1
End of preview. Expand in Data Studio
YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/datasets-cards)

Dataset Summary

CQADupstack-english-Fa is a Persian (Farsi) dataset developed for the Retrieval task, focused on duplicate question detection in community question-answering (CQA) forums. This dataset is a translated version of the "english" (English Language & Usage) StackExchange subforum from the original English CQADupstack collection and is part of the FaMTEB benchmark under the BEIR-Fa suite.

  • Language(s): Persian (Farsi)
  • Task(s): Retrieval (Duplicate Question Retrieval)
  • Source: Translated from English using Google Translate
  • Part of FaMTEB: Yes — under BEIR-Fa

Supported Tasks and Leaderboards

This dataset is designed to evaluate text embedding models in their ability to retrieve semantically equivalent or duplicate questions within the domain of English language and usage. Results can be compared on the Persian MTEB Leaderboard (filter by language: Persian).

Construction

The dataset was constructed by:

  • Translating the "english" StackExchange subforum using the Google Translate API
  • Maintaining CQA structure for duplicate retrieval evaluation

As described in the FaMTEB paper, translation quality for BEIR-Fa datasets was evaluated via:

  • BM25 performance comparison (original vs translated)
  • GEMBA-DA framework using LLMs for quality assessment

Data Splits

According to the FaMTEB paper (Table 5), all CQADupstack-Fa datasets share a unified test set:

  • Train: 0 samples
  • Dev: 0 samples
  • Test: 480,902 samples (aggregate)

The cqadupstack-english-fa subset specifically includes approximately 45.6k examples (user-provided). Refer to the dataset provider for finer-grained split details.

Downloads last month
39