Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Akhlaghi, Amir Mohammad, Shabani, Amirhossein, Abdolmaleki, Mostafa, Kheradpisheh, Saeed Reza
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2512.07454
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908698078609408
author Akhlaghi, Amir Mohammad
Shabani, Amirhossein
Abdolmaleki, Mostafa
Kheradpisheh, Saeed Reza
author_facet Akhlaghi, Amir Mohammad
Shabani, Amirhossein
Abdolmaleki, Mostafa
Kheradpisheh, Saeed Reza
contents The democratization of AI is currently hindered by the immense computational costs required to train Large Language Models (LLMs) for low-resource languages. This paper presents Persian-Phi, a 3.8B parameter model that challenges the assumption that robust multilingual capabilities require massive model sizes or multilingual baselines. We demonstrate how Microsoft Phi-3 Mini -- originally a monolingual English model -- can be effectively adapted to Persian through a novel, resource-efficient curriculum learning pipeline. Our approach employs a unique "warm-up" stage using bilingual narratives (Tiny Stories) to align embeddings prior to heavy training, followed by continual pretraining and instruction tuning via Parameter-Efficient Fine-Tuning (PEFT). Despite its compact size, Persian-Phi achieves competitive results on Open Persian LLM Leaderboard in HuggingFace. Our findings provide a validated, scalable framework for extending the reach of state-of-the-art LLMs to underrepresented languages with minimal hardware resources. The Persian-Phi model is publicly available at https://huggingface.co/amirakhlaghiqqq/PersianPhi.
format Preprint
id arxiv_https___arxiv_org_abs_2512_07454
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Persian-Phi: Efficient Cross-Lingual Adaptation of Compact LLMs via Curriculum Learning
Akhlaghi, Amir Mohammad
Shabani, Amirhossein
Abdolmaleki, Mostafa
Kheradpisheh, Saeed Reza
Computation and Language
Artificial Intelligence
The democratization of AI is currently hindered by the immense computational costs required to train Large Language Models (LLMs) for low-resource languages. This paper presents Persian-Phi, a 3.8B parameter model that challenges the assumption that robust multilingual capabilities require massive model sizes or multilingual baselines. We demonstrate how Microsoft Phi-3 Mini -- originally a monolingual English model -- can be effectively adapted to Persian through a novel, resource-efficient curriculum learning pipeline. Our approach employs a unique "warm-up" stage using bilingual narratives (Tiny Stories) to align embeddings prior to heavy training, followed by continual pretraining and instruction tuning via Parameter-Efficient Fine-Tuning (PEFT). Despite its compact size, Persian-Phi achieves competitive results on Open Persian LLM Leaderboard in HuggingFace. Our findings provide a validated, scalable framework for extending the reach of state-of-the-art LLMs to underrepresented languages with minimal hardware resources. The Persian-Phi model is publicly available at https://huggingface.co/amirakhlaghiqqq/PersianPhi.
title Persian-Phi: Efficient Cross-Lingual Adaptation of Compact LLMs via Curriculum Learning
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2512.07454