Detect encoding problems, hidden content, and dialect mixing in Arabic JSONL datasets.
npx skills add https://github.com/Moshe-ship/hurmoz --skill majal-moshe-ship
Please help me install this Agent Skill. Skill: majal Source: https://github.com/Moshe-ship/hurmoz/tree/main/majal Command: npx skills add https://github.com/Moshe-ship/hurmoz --skill majal-moshe-ship
This Skill helps data teams quickly identify encoding problems, hidden content, and dialect mixing in Arabic JSONL training data, preventing quality issues during model fine-tuning.
Scan your training_data.jsonl with Majal to surface encoding and dialect issues.
High-intent search queries and answers about installing and using this skill.
To detect mojibake and encoding issues in Arabic JSONL training data, scan the dataset to identify mixed encodings, null bytes, and invisible characters to ensure strict UTF-8 consistency before model fine-tuning.
Identifying dialect mixing in Arabic NLP datasets involves scanning JSONL records to flag inappropriate language trends and ensure dialect balance. This improves data quality and model performance before fine-tuning.
Yes, you can check for hidden content and PII leakage in JSONL files by flagging empty Arabic fields, duplicate records, and potential PII leakage. This prevents quality issues during model fine-tuning.
Yes, running these Arabic data quality checks on JSONL files requires the Majal tool installed. Once installed, it can surface encoding, hidden content, and dialect issues in your training data.
To scan large-scale Arabic training data for null bytes and empty fields, run a data quality check on your JSONL inputs. This will flag encoding problems and empty Arabic fields across your records.
rayaninoHandle Arabic scholarly text with UTF-8 encoding and diacritic preservation.
jackquelinunpredictable827Detect encoding issues, hidden content, and dialect mixing in Arabic JSONL training data.
jackquelinunpredictable827Scrape and clean Arabic web content into JSONL training datasets.
Moshe-shipCollect and clean Arabic web content into JSONL training data.
