cnemri/tuni-folk-gemini

RLFT of Gemini 3.5 Flash for Tunisian folk poetry (الشعر الشعبي التونسي): rhyme-topology reward, Cloud Run scorer, calibrated against the national corpus # tuni-folk-gemini **An end-to-end SFT → RL pipeline that teaches Gemini 3.5 Flash to compose Tunisian folk poetry (الشعر الشعبي التونسي / الملحون).** ``` gemini-3.5-flash │ │ STAGE 1 — supervised fine-tuning │ 3,631 (prompt, poem) pairs from the Diwan │ teaches the DISTRIBUTION — what folk poetry sounds like ▼ tunifolk-sft-base │ │ STAGE 2 — reinforcement fine-tuning │ 843 prompts, no answers, graded by a programmatic reward │ teaches the CONSTRAINT — rhyme topology, dialect, novelty ▼ tunifolk-rlft-from-sft ``` --- ## Why this task, and why two stages Tunisian folk poetry is an unusually good tuning target. Its metrical system is *not* Khalilian — Muhammad al-Marzuqi's argument, adopted by Muhyi al-Din Khrayyef and by the editorial committee of the national corpus, is that the Tunisian dialect has `asbab` and `awtad` but no `fawasil`, so classical `buhur` cannot be applied at all: > «الشعر الشعبي الحديث لا يمكن أن نطبّق عليه البحور القديمة… ولا يمكن أن نضبط موازينه > إلا بواسطة الإيقاع» — المرزوقي، *الأدب الشعبي*، ص 83 What *does* hold a poem together is a **rhyme topology** that differs per form and is fully decidable by a program. That is what makes a reinforcement reward possible here at all: an objective, cheap, hard-to-game structural signal on a creative task. The two stages teach different things, and neither substitutes for the other: - **Supervised tuning can only teach what a target exemplifies.** Right for register — diction, imagery, the sound of the Mudawwana. Wrong for "satisfy this topology *and* stay in dialect *and* don't copy the source", because no single poem demonstrates the conjunction, and copying a training poem is zero loss. - **Reinforcement tuning can only teach what a scorer can grade.** It never sees a target, so it can be pushed toward that conjunction and told copying is worthless — but from the base model it wastes its early run learning register that a supervised pass learns in an hour. **The order is enforced, not …