Pashto instruction-tuned LoRA adaptation of microsoft/Phi-3-mini-4k-instruct (ZamAI Labs).

Model artifacts and experiments published by ZamAI Labs (training results and references).

Core processing utilities and pipelines for Pashto text (normalization, cleaning, preprocessing).

ZamAI Labs — datasets, Pashto processing, models, and training pipelines powering the ecosystem.

Template and starter structure for Pashto language projects in ZamAI Labs.

Pashto-focused work with mT5 (experiments, fine-tuning, references) in ZamAI Labs.

Reusable training and experiment spaces for ZamAI Labs (templates, scripts, and runs).

Curated and processed Pashto datasets for ZamAI Labs (with source attribution and dataset documentation).

Training pipelines and data workflow tools for building and fine-tuning ZamAI models.