Open-source alternatives to Sam Hq

This project is a multi-modal image segmentation framework and a text-to-mask vision model. It serves as a SAM-based visual segmenter designed to isolate distinct objects within images and video by converting natural language prompts and other inputs into pixel-level semantic masks. The system functions as a multi-modal image segmentation framework that integrates text, image, and audio signals to generate masks. It includes an interactive video object tracker that isolates and tracks visual entities across video frames using referring images or textual queries. The framework provides capabi

MedSAM is a deep learning framework designed for automating the segmentation of anatomical structures in 2D and 3D medical imagery. It provides specialized tools for fine-tuning pretrained segmentation weights on custom medical datasets and evaluating the accuracy of those predictions against ground truth labels. The project focuses on adapting the Segment Anything Model architecture for medical use, enabling the isolation of specific anatomical structures through prompt-guided methods such as bounding boxes and point prompts. The system covers a full medical AI workflow, including data engi

Recognize-anything is a multimodal foundation model designed for image recognition, visual tagging, and the generation of descriptive text captions from visual input. It functions as a multimodal embedding model that maps images and text into a shared vector space to enable cross-modal retrieval and recognition. The system implements zero-shot image classification and open-vocabulary object detection, allowing it to recognize object categories not present in the original training data through custom label embeddings. It also features a visual tagging engine and a captioning system that produc

This project is an integrated creative interface for image manipulation that combines generative canvas expansion, object segmentation, and diffusion-based inpainting. It functions as an extension for Stable Diffusion, providing a workflow to isolate specific elements and perform targeted content modifications through prompt-guided synthesis. The tool distinguishes itself by automating the creation of precise selection masks, allowing users to identify and isolate objects by pointing to them rather than manually drawing selections. By chaining segmentation models with generative diffusion pip

This is a structured deep learning curriculum for programmers, delivered as a collection of Jupyter notebooks. It teaches the fundamentals of training neural networks for computer vision, natural language processing, tabular data analysis, and collaborative filtering using PyTorch and the fastai library. The course is designed to be hands-on, guiding learners from building a training loop from scratch to fine-tuning pretrained models for a variety of practical tasks. The curriculum distinguishes itself by covering the full lifecycle of a deep learning project, from data preparation and augmen

This project is a neural network image classifier and a set of tools for building and training convolutional neural networks to recognize and categorize images. It serves as a machine learning educational guide, providing a practical resource for learning neural network fundamentals through an onboarding process. The system includes a dedicated workflow for pretrained model fine-tuning, allowing existing network weights to be adapted to new image categories. This is supported by a transfer learning pipeline that replaces final classification layers and adjusts weights through targeted retrain

BiRefNet is a PyTorch image segmentation framework designed for high-precision binary mask generation. It functions as a bilateral image segmentation model used to isolate foreground objects from complex backgrounds, as well as a specialized tool for camouflaged object detection and industrial defect detection. The project is designed for export to the ONNX format, which facilitates cross-platform deployment and inference. It supports custom model fine-tuning on user-provided image and mask datasets to adapt the model for specialized professional use cases. The system covers high-resolution

This project is a self-supervised contrastive learning framework designed to train deep learning models to learn visual representations from images without using human-provided labels. It provides a system for developing pretrained visual representation models that can be adapted for downstream computer vision tasks. The framework includes tools for semi-supervised image classification, which combines large unlabeled datasets with small labeled sets to improve accuracy. It also features a linear probe evaluation tool to assess the quality of learned image features by training a simple linear

Open NSFW is a deep learning image classifier and content filter designed to identify explicit or unsafe visual content. It uses a neural network to analyze images and generate probability scores that determine whether media is not safe for work. The system employs a Caffe-based model architecture to perform automated visual feature extraction and image classification. It supports the use of pre-trained weights and allows for fine-tuning the network with custom datasets to adjust detection accuracy. The project covers end-to-end image moderation workflows, including automated content scannin

jetson-inference is a set of libraries and tools for executing optimized deep learning models on embedded GPU hardware. Its primary purpose is to enable real-time computer vision and AI inference at the edge with low latency and high throughput. The project distinguishes itself through high-performance streaming analytics and the ability to execute concurrent AI pipelines on auto-grade silicon. It provides specialized support for multi-sensor stream processing, utilizing zero-copy data transport to load camera frames directly into GPU memory. The codebase covers a broad surface of capabiliti

This project is a collection of pre-trained machine learning models and conversion pipelines designed for running inference directly in the browser using TensorFlow.js. It provides a library of ready-to-use models for computer vision, audio classification, and natural language processing tasks. The suite includes specialized tools for transforming Python-based Keras models into JSON formats compatible with web environments. It enables the deployment of these models by fetching architectures and weight shards via HTTP for client-side execution. The project covers a broad range of capabilities

This project is a TensorFlow object detection framework designed for training and deploying Single Shot MultiBox Detector models. It provides a neural network training toolkit for implementing the SSD architecture to achieve real-time image and video object localization. The framework includes a dedicated data pipeline for transforming object detection datasets into binary record formats to increase training speed and performance. It also features utilities for converting model weights between different checkpoint formats to facilitate the reuse of pre-trained networks. The system covers a b

MMSegmentation is an open-source semantic segmentation toolbox built on PyTorch that provides a modular, configurable framework for building, training, evaluating, and deploying segmentation models. At its core, it offers a config-driven pipeline that assembles training, evaluation, and inference workflows by parsing hierarchical configuration files, with a modular component registry that enables plug-and-play composition of neural network modules, optimizers, datasets, and metrics. The framework supports the full model lifecycle through a unified runner interface that controls training, testi

Map-anything is a 3D scene reconstruction framework and neural geometry estimator designed to transform two-dimensional images into metric three-dimensional spatial representations using feed-forward neural networks. It provides a specialized toolkit for predicting camera intrinsics and ray directions from single images without requiring external geometric metadata. The project includes a 3D model benchmarking suite that utilizes a unified model wrapper to standardize outputs from diverse reconstruction models. This allows for consistent evaluation and accuracy measurement across various spat

NSFW detection on the client-side via TensorFlow.js

ComfyUI-nunchaku is a 4-bit diffusion inference engine and a set of nodes for running low-precision quantized diffusion models within ComfyUI visual workflows. It provides a backend that reduces memory overhead and increases generation speed for transformer models. The project includes specialized tools for identity-preserving generation and an image-to-image guidance toolkit that uses depth maps and reference images. It also features a multimodal visual question answering implementation and a utility for merging multiple quantized model files into single unified files. The engine covers a b

Segment Geospatial is a Python toolkit for isolating geographic features in remote sensing imagery using the Segment Anything Model. It functions as a remote sensing image processor that converts map tiles into georeferenced formats to generate segmentation masks from satellite data. The system enables the extraction of geographic objects through automatic mask generation or manual prompts, such as text descriptions, bounding boxes, and interactive markers. It supports timeseries imagery segmentation to track or identify objects across sequences of images over different dates and provides a g

MagicQuill is a suite of interactive tools for image segmentation, diffusion-based editing, layered composition, and prompt-guided visual synthesis. It functions as a diffusion model image editor and a layered visual composition tool, enabling the addition, removal, and recoloring of image elements through a combination of sketches and text prompts. The system features a prompt-guided image generator that predicts editing instructions by analyzing user drawings to automatically populate text prompts. It allows for visual style control by swapping generative model weights to shift outputs betw

MobileSAM is a lightweight image segmenter and promptable vision model designed for fast object isolation on resource-constrained hardware. It functions as an automatic image masking tool capable of detecting and isolating distinct objects across an entire image without manual input. The system enables prompt-based object masking using coordinate points or bounding boxes to generate precise masks. It also supports all-object image segmentation through object-aware prompt sampling to identify every distinct object in a scene. To facilitate mobile and edge deployment, the model is compatible w

FastSAM is an image segmentation framework that uses convolutional neural networks to isolate visual elements and generate masks for detectable objects within images. It provides a system for both automatic all-object segmentation and promptable image segmentation. The project utilizes an inference-optimized architecture to reduce computational overhead, enabling faster mask generation and real-time visual analysis. It supports the creation of precise masks through various prompt inputs, including points, bounding boxes, and text descriptions. The framework covers broader computer vision cap

This project is a multi-label classification pipeline designed for genre prediction. It implements a machine learning workflow that assigns multiple category labels to a single item by processing both textual and visual input data. The system utilizes multimodal feature extraction to transform images and text descriptions into semantic vectors. This process includes using pre-trained networks for visual feature extraction and semantic word averaging for text analysis, allowing the model to integrate different data types into a unified input. The pipeline covers the full machine learning life

Learn_Prompting is an educational project focused on prompt engineering, providing the principles and techniques required to craft effective inputs and improve the quality of generative AI outputs. The project covers advanced prompting strategies to enhance reasoning, reliability, and output quality. This includes techniques for task decomposition, chain-of-thought reasoning, and the use of few-shot and zero-shot guidance. It also addresses model security through the study of prompt hacking, vulnerability analysis, and privacy auditing to prevent sensitive data leaks. The scope extends to th

DeepSeek-VL is a multimodal large language model and image-to-text reasoning engine. It functions as a vision-language model and visual question answering system that integrates visual perception with linguistic reasoning to understand and describe images. The project enables multimodal image understanding and document image analysis, specifically processing screenshots of web pages and technical diagrams. It provides capabilities for visual conversational AI, allowing users to interact with visual data to extract insights and perform complex reasoning across different types of visual informa

ImageBind is a multi-modal embedding model and joint representation learner that maps images, text, audio, and other modalities into a single shared vector space. It functions as a cross-modal retrieval framework designed to bind multiple sensory inputs into one cohesive mathematical embedding. The system uses a contrastive learning architecture to align disparate data types by maximizing the similarity between related samples. This allows the model to perform zero-shot multimodal classification and execute cross-modal data retrieval, such as locating visual content via natural language descr

This project is a standardized set of abstraction and reasoning problems designed for benchmarking the ability of artificial intelligence models to learn new rules. It functions as a fluid intelligence test and a reasoning benchmark, utilizing a collection of grid-based puzzles and a program synthesis dataset to evaluate how agents generate algorithms from examples. The project focuses on measuring general fluid intelligence and the capacity for zero-shot generalization, testing whether a system can apply learned logic to unseen problems without relying on task-specific training. It provides