Dataset Viewer
The dataset could not be loaded because the splits use different data file formats, which is not supported. Read more about the splits configuration. Click for more details.
Couldn't infer the same data file format for all splits. Got {NamedSplit('train'): ('parquet', {}), NamedSplit('test'): ('text', {})}
Error code: FileFormatMismatchBetweenSplitsError
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
MoaLLM Bilingual Technical & Study Corpus (Arabic/Egyptian + English)
A large-scale, high-quality bilingual dataset optimized for training ultra-fast static embedding models (like Model2Vec), cross-lingual retrieval engines, and dense passage retrieval.
Dataset Summary
- Total Unique Passages: 1,186,899
- Total Unique Query-Passage Pairs: 1,138,237
- Language Composition:
- English Technical: 1,091,470 passages
- Arabic / Egyptian Dialect: 86,225 passages
- Bilingual Code-Switched: 9,204 passages
- Multi-Scale Passage Lengths:
- Short Definitions & Notes (15–80 words): 897,145 (75.6%)
- Standard Sections (80–250 words): 229,784 (19.4%)
- Long Comprehensive Sections (250–600 words): 59,970 (5.1%)
Query-Passage Pairs by Source
| Source Category | Count | Description |
|---|---|---|
github |
776,878 | Technical documentation extracted from 358 top GitHub repositories. |
hf_parallel |
136,540 | 100% authentic, non-templated English-Egyptian parallel text & sentences (oddadmix, IbrahimAmin, michsethowusu, HeshamHaroon, Omar-youssef). |
youtube |
134,505 | Timed, cleaned lecture subtitles from 932 technical & university courses. |
hf |
84,377 | Pure-text STEM textbooks and technical QA pairs (cosmopedia-v2, finemath, open-web-math, cornstack, codesearchnet, UltraTextbooks). |
synthetic |
5,877 | Cross-lingual search queries for documentation anchors. |
llm_student_note |
60 | Non-templated bilingual student study notes and conceptual search queries. |
Schemas
1. data/passages/*.parquet (24 shards)
text(string): Multi-scale passage text (15 to 600 words).
2. data/pairs/*.parquet (23 shards)
query(string): Natural search query, question, heading, or Egyptian Arabic cross-lingual query.passage(string): Target text passage or English sentence.source(string): Source provenance (github:<repo>,youtube:<channel>,hf_parallel:<dataset>, etc.).
Quick Start with Hugging Face datasets
from datasets import load_dataset
HF_DATASET_REPO = "MohamedSameh77i/moallm-bilingual-corpus"
# Load passages
passages = load_dataset(HF_DATASET_REPO, data_files="data/passages/*.parquet", split="train")
# Load pairs
pairs = load_dataset(HF_DATASET_REPO, data_files="data/pairs/*.parquet", split="train")
- Downloads last month
- 114