This is a Phi Family of SLMs book for getting started with Phi Models. Phi a family of open sourced AI models developed by Microsoft. Phi models are the most capable and cost-effective small language models (SLMs) available, outperforming models of the same size and next size up across a variety of language, reasoning, coding, and math benchmarks

A ternary, zero-heap tiny language model that runs inside a $2 microcontroller — bit-exact Python <-> C99 <-> Cortex-M3 (QEMU) parity. Apache-2.0.

Wonderful Matrices to Build Small Language Models

SLM-SQL: An Exploration of Small Language Models for Text-to-SQL

'Zest' Natural Language to CLI commands tool from Spicy Lemonade

An Android App recreating the Simon Says game. Uses MediaPipe to run an LLM on device

This repository provides everything you need to perform Supervised Fine-Tuning (SFT) of the Qwen2.5-Coder-1.5B-Instruct model—or any of its larger variants (7B, 14B, 32B)—on the Qwen Models, using the nvidia/OpenCodeReasoning dataset.

Code for paper "Achieving Sparse Activation in Small Language Models"

A Streamlit demo using an SLM (Phi) and RAG to showcase how an AI can help users with learning functionality on a website

Enhancing Reasoning Skills in Small Persian Medical Language Models Can Outperform Large-Scale Data Training

Small intelligent agent powered by IBM Granite.

PocketLFM — run Liquid AI's LFM2.5 large language model fully on-device on Android. Offline, private, no cloud. Open-source edge AI via llama.cpp (GGUF). First independent open-source Android app to run LFM2.5 outside Liquid's own Apollo/LEAP.

Small GPT trained from scratch on Tang poetry + Song ci

The frontend code used by the 'Zest' Natural Language to CLI commands tool from Spicy Lemonade

TinyGPT - 9M Parameters

Pruning, Quantization recipe for Small Language Model

Small Language Models

GraphRAG-powered Small Language Model for experimentation and statistical analysis Q&A

Applied Decision Architecture Matrix - Small Language Model

Edge-native AI companion that lives 24/7 on a Raspberry Pi 5 + AI HAT+ 2 (Hailo-10H NPU, 40 TOPS). Fully offline: local 1.5B SLM (qwen2.5), self-hosted SearXNG, no cloud. Biologically inspired — heartbeat, vital signs, 5-channel mood, STM/LTM hierarchy, nightly Sleep Metabolism, proactive research, and a self-reflection loop.