Yohei Nakajima writes about glance, a tool designed to allow users to ask an open vision-language model (VLM) typed questions about images and receive probability data directly on their own machine. Rather than generating new text or training models, it acts as a measurement harness that reads logits from frozen models—such as Qwen3-VL-4B by default—to provide yes/no answers, single-choice selections, and qualitative ratings without any image data leaving the user's device.
- The tool provides three response types: "noul" (yes/no), choice (pick one from a list), and score (a rating on a specified scale).
- It includes an experimental MLX backend to provide faster runtimes specifically for Apple Silicon users.
- Glance can perform self-calibration using unlabeled data or precise calibration through labeled datasets to improve rating accuracy.
- The software is designed with privacy in mind, ensuring all inference and logging stay local on the user's hardware.
- The tool provides three response types: "noul" (yes/no), choice (pick one from a list), and score (a rating on a specified scale).
- It includes an experimental MLX backend to provide faster runtimes specifically for Apple Silicon users.
- Glance can perform self-calibration using unlabeled data or precise calibration through labeled datasets to improve rating accuracy.
- The software is designed with privacy in mind, ensuring all inference and logging stay local on the user's hardware.
This Python code demonstrates a neural network application on a CircuitPython board, utilizing a camera (OV7670) for image capture, preprocessing, and inference using a digit classifier. It includes image conversion, auto-cropping, and normalization steps.
This article details how to train an image classification model on an ESP32 using both the SenseCraft AI platform and a custom TensorFlow Lite implementation. It covers setting up binary classification, training the model, and deploying it on ESP32-S3 devices.
Multi-class zero-shot embedding classification and error checking. This project improves zero-shot image/text classification using a novel dimensionality reduction technique and pairwise comparison, resulting in increased agreement between text and image classifications.
This study examines the effectiveness of employing advanced machine learning techniques, particularly YOLO models, to solve Google's reCAPTCHAv2 captchas. The research finds that the use of machine learning achieves a 100% success rate in solving captchas, indicating that image-based captchas are becoming more vulnerable to AI technologies.
Learn how to perform image classification on edge devices like the Raspberry Pi using TensorFlow Lite and Mobilenet V2 models.
This page provides information on how to connect and utilize the SenseCAP Watcher as a Grove sensor using UART communication.
Use large language models embedded in single-file executables from the command line to perform tasks like renaming images based on their visual content
