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

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

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

Real-time webcam demo using SmolVLM with vLLM backend

Finetune VLM for Image classification in specific fields

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

A small VLM that sees everything

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

Scripts for combining SmolVLM and LLM

SmolVLM2 on the RK3588 NPU

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

⭐ Comparing VLMs with CNNs for garbage classification

A simple web application for real-time AI vision analysis using SmolVLM-500M-Instruct with live camera feed processing and text-to-speech.

Receipt OCR using Fine-tuned VLMs

A Flask-based web app for managing multimodal datasets text and images with CRUD operations via SQLite, and seamless export as a structured Parquet dataset to Hugging Face Hub.

This blog post introduces SmolVLM, a 2B VLM, SOTA for its memory footprint. SmolVLM is small, fast, memory-efficient, and fully open-source. All model checkpoints, VLM datasets, training recipes and tools are released under the Apache 2.0 license.

SmolVLM2 on the RK3588 NPU

SmolVLM2 on the RK3588 NPU

Vision-Language Model (VLM) for real-time video analysis and description via webcam.

A some what optimized implementation of some light weight and popular models