ComfyUI nodes for NVIDIA LocateAnything-3B visual grounding
-
Updated
Jun 27, 2026 - Python
#
locateanything
Here are 10 public repositories matching this topic...
Implement of LocateAnything: Fast and High-Quality Vision-Language Grounding with Parallel Box Decoding in windows system, and we also extend the inference result with object detection confidence.
-
Updated
May 31, 2026 - Python
Self-hosted vision model trainer: label images with dots, train YOLO and Apache-licensed challengers, judge every run on a frozen holdout, then serve it — web cockpit, API, and Telegram data-collection bots. Multi-machine training over ssh, Docker-first, AGPL-3.0.
training docker machine-learning computer-vision deep-learning image-annotation telegram-bot self-hosted pytorch yolo object-detection labeling-tool mlops fastapi yolov8 locateanything
-
Updated
Sep 4, 2026 - Python
Natural-language visual localization SDK for AI agents, RPA, GUI automation, LangChain, and data annotation
python appium gui-automation rpa visual-grounding fastapi playwright ai-agent langchain locateanything
-
Updated
May 29, 2026 - Python
Local-first desktop vision agent: YOLO realtime + adaptive LocateAnything-3B reasoning, event memory, Gradio cockpit
python real-time computer-vision desktop screen-capture yolo object-detection supervision gradio vlm ai-agent yolov8 vision-language-model local-ai locateanything
-
Updated
Aug 1, 2026 - Python
Ein Vision-Language-Modell spielt Doom: erkennt Monster, verschont Menschen, und ordnet Unbekanntes ein. Alles gegen die Wahrheit der Engine gemessen.
agent computer-vision doom object-detection vizdoom vision-language-model open-vocabulary-detection locateanything
-
Updated
Aug 29, 2026 - Python
Desktop UI for visual auto-labeling based on NVIDIA LocateAnything, with local and SSH execution.
-
Updated
Jul 15, 2026 - Python
An autonomous macOS agent that replaces fragile web scrapers with visual grounding. Built with Python, it translates text queries into native hardware-level mouse actions using a locally hosted Vision-Language Model.
-
Updated
Jul 5, 2026 - Python
Serverless visual grounding with NVIDIA LocateAnything-3B on Modal — FastAPI + Next.js UI for image and video bounding boxes.
-
Updated
Jun 10, 2026 - TypeScript
Run NVIDIA LocateAnything-3B natively on Apple Silicon with MLX — local vision-language object detection with bounding boxes, single-image and batch modes. ~33 tok/s on M1, 16 GB RAM, zero cloud. Non-commercial research license.
-
Updated
Jul 13, 2026 - Python
Add this topic to your repo
To associate your repository with the locateanything topic, visit your repo's landing page and select "manage topics."