aiqbench analysis
Shaer demonstrates significant progress in controllable generative modeling by successfully aligning Large Language Models with the strict prosodic and semantic constraints of Classical Arabic poetry. The study validates a QLoRA fine-tuning approach on a curated dataset, achieving high accuracy in meter adherence while maintaining literary quality without introducing memorization artifacts. This work highlights the growing capability of ML systems to handle complex, rule-based linguistic structures beyond standard natural language tasks.
Shaer: Controlled Arabic Poetry Generation with Meter Subform and Semantic Conditioning
arXiv:2610.09756v1 Announce Type: new Abstract: Classical Arabic poetry generation requires simultaneously satisfying semantic, linguistic, and fine-grained prosodic constraints. Existing systems typically control broad poetic attributes but do not jointly model semantic intent, meter subform, and poem length. We present Shaer, a controllable Classical Arabic poetry generation framework jointly conditioned on natural-language descriptions, meter subforms, and target hemistich counts. To support this task, we construct an enriched corpus of 116,032 classical Arabic poems derived from Ashaar, containing normalized meter-subform labels and automatically generated, validated semantic descriptions. We then adapt Yehia-7B using QLoRA-based supervised fine-tuning with a completion-only objective. Our evaluation combines automatic assessment of base-meter conformity, requested-subform adherence, and length control with three LLM judges, blinded human evaluation, and memorization analysis. Shaer achieves 95.17% base-meter accuracy, 91.75% poem-level meter-subform accuracy, and 83.40% exact count accuracy. Relative to its untuned foundation model, these results represent gains of 68.68, 57.77, and 38.93 percentage points, respectively; Shaer also attains the highest base-meter accuracy among all evaluated systems. Multi-LLM evaluation and a blinded human assessment of top-ranked outputs further indicate competitive semantic and literary quality. Finally, analysis of all 3,481 test generations finds no exact copies from the training corpus or paired source poems. Code, models, and datasets are publicly available.
Abstract:Classical Arabic poetry generation requires simultaneously satisfying semantic, linguistic, and fine-grained prosodic constraints. Existing systems typically control broad poetic attributes but do not jointly model semantic intent, meter subform, and poem length. We present Shaer, a controllable Classical Arabic poetry generation framework jointly conditioned on natural-language descriptions, meter subforms, and target hemistich counts. To support this task, we construct an enriched corpus of 116,032 classical Arabic poems derived from Ashaar, containing normalized meter-subform labels and automatically generated, validated semantic descriptions. We then adapt Yehia-7B using QLoRA-based supervised fine-tuning with a completion-only objective. Our evaluation combines automatic assessment of base-meter conformity, requested-subform adherence, and length control with three LLM judges, blinded human evaluation, and memorization analysis. Shaer achieves 95.17% base-meter accuracy, 91.75% poem-level meter-subform accuracy, and 83.40% exact count accuracy. Relative to its untuned foundation model, these results represent gains of 68.68, 57.77, and 38.93 percentage points, respectively; Shaer also attains the highest base-meter accuracy among all evaluated systems. Multi-LLM evaluation and a blinded human assessment of top-ranked outputs further indicate competitive semantic and literary quality. Finally, analysis of all 3,481 test generations finds no exact copies from the training corpus or paired source poems. Code, models, and datasets are publicly available.
22 pages, 7 figures. Code: this https URL ; models and datasets: this https URL
Computation and Language (cs.CL)
arXiv:2610.09756 [cs.CL]
(or arXiv:2610.09756v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2610.09756 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 311 words
Related coverage
Beyond Poetry: Can Large Language Models Generate Classical Arabic Maqamat?
arXiv · Sep 24, 2026
2026.10: You are here 📍
Home Assistant · Oct 7, 2026
Falcon-Emirati: When an LLM Learns the Dialect, the Culture, and the Nuance
Hugging Face · Oct 6, 2026
Here’s How Delhi Achieved Its Epic Power-Grid Fix
IEEE Spectrum · Sep 28, 2026