File size: 1,998 Bytes
e79f4f0 c90b9f1 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 |
---
configs:
- config_name: default
data_files:
- split: test
path: qrels/test.jsonl
- config_name: corpus
data_files:
- split: corpus
path: corpus.jsonl
- config_name: queries
data_files:
- split: queries
path: queries.jsonl
---
## Dataset Summary
**CQADupstack-tex-Fa** is a Persian (Farsi) dataset curated for the **Retrieval** task, specifically targeting **duplicate question detection** in community question-answering (CQA) forums. This dataset is a translation of the **"TeX - LaTeX" StackExchange subforum** from the 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 assesses the ability of **text embedding models** to identify **semantically equivalent or duplicate questions** within the domain of **TeX and LaTeX programming**. Evaluation results appear on the **Persian MTEB Leaderboard** (filter by language: Persian).
## Construction
The construction process included:
- Translating the **"TeX - LaTeX"** subforum from the CQADupstack dataset using the **Google Translate API**
- Retaining the original duplicate-retrieval task structure
The **FaMTEB** paper confirms quality through:
- **BM25 retrieval comparisons** across English and Persian versions
- **GEMBA-DA framework**, which employed LLMs to verify translation fidelity
## Data Splits
Per FaMTEB paper (Table 5), all **CQADupstack-Fa** sub-datasets share the same test pool:
- **Train:** 0 samples
- **Dev:** 0 samples
- **Test:** 480,902 samples (aggregate)
The individual split count for `cqadupstack-tex-fa` is not provided separately. For detailed partitioning, refer to FaMTEB resources.
**Total (user-reported):** ~76.2k examples
|