smolvlm

A Rust library integrated with ONNXRuntime, providing a collection of Computer Vison and Vision-Language models such as YOLO, FastVLM, and more.

A CPU Realtime VLM in 500M. Surpassed Moondream2 and SmolVLM. Training from scratch with ease.

🎭 Real-time voice-controlled 3D avatar with multimodal AI - speak naturally and watch your AI companion respond with perfect lip-sync

Finetune VLM for Image classification in specific fields

Real-time webcam demo using SmolVLM with vLLM backend

This repository contains the implementation of AlignVLM paper, which proposes a novel method for vision language alignment

SmolVLM2 on the RK3588 NPU

A small VLM that sees everything

A full-stack ChatGPT-like application built (almost) from scratch

𝗜𝗺𝗮𝗴𝗲 𝗖𝗮𝗽𝘁𝗶𝗼𝗻𝗶𝗻𝗴 𝘄𝗶𝘁𝗵 𝗢𝗰𝗰𝗹𝘂𝘀𝗶𝗼𝗻 𝗔𝗻𝗮𝗹𝘆𝘀𝗶𝘀 | 𝗩𝗶𝗧-𝗚𝗣𝗧𝟮 | 𝗦𝗺𝗼𝗹𝗩𝗟𝗠 | 𝗕𝗘𝗥𝗧

SmolVLM2 on the RK3588 NPU

SmolVLM2 on the RK3588 NPU

⭐ Comparing VLMs with CNNs for garbage classification

Tool for multi-modal recommendation

Scripts for combining SmolVLM and LLM

Real-time vision demo using SmolVLM with llama.cpp backend

Build and run powerful offline AI models on Android using Termux, llama.cpp, Qwen2.5-Coder, and SmolVLM. Complete beginner-friendly guide with step-by-step documentation.

Receipt OCR using Fine-tuned VLMs

A benchmark suite for lightweight generative multimodal Vision-Language Models, comparing ViLT and SmolVLM under resource-constrained inference environments. Demonstrates CPU-only deployment, model evaluation, and multimodal reasoning with images and text, highlighting practical GenAI engineering for real-world applications.

On-device AI camera guard you program in one English sentence — Arm CPU only (hand-written NEON gate + KleidiAI INT4 SmolVLM), 100% local, no cloud/NVIDIA. Runs on a spare iPhone or a Mac webcam. Arm AI Optimization Challenge · Track 3 Mobile AI.