Tunisian Arabic NLP Resources: an access-verified inventory

An access-verified inventory of natural language processing resources for Tunisian Arabic (Derja, ISO 639-3 aeb): datasets, speech corpora, models, benchmarks and lexicons. Each entry is checked for whether it can actually be obtained (open, on request, paywalled, paper-only or gated), and for multi-dialect resources the Tunisian share is recorded rather than counting the whole set. Maintained on GitHub, where corrections and additions arrive through issue forms.