owl-vit

Run zero-shot prediction models on your data

Enhance your skills in prompt engineering for vision models. Learn to effectively prompt, fine-tune, and track experiments for models like SAM, OWL-ViT, and Stable Diffusion 2.0 to achieve precise image generation, segmentation, and object detection.

Annotation assistant tool that uses CLIP to find described objects in dataset and label them

Vector with Advanced AI compatibilities

OWL-ViT module for Autodistill.

This is a multimodal model design for the Vision Question Answering (VQA) task. It integrates the Llama2 13B, OWL-ViT, and YOLOv8 models.

Interactive OpenCV playground — chain image processing operations with live preview.

LLMs/LVMs/LMMs Prompt Engineering & Quantization

FastAPI service running YOLO instance segmentation on aerial imagery to extract ordered rooftop corner coordinates

POLARIS: staged vision-language pipeline for cross-temporal landmark matching with OWL-ViT detection, FastVLM enrichment, and geometric verification

Back-end model of Efficient Video Analytics, 2023 summer research intern

Implementing a solution for fall detection in videos, with specific emphasis on detecting falls accurately and minimizing false positives.

Zero-shot amenity detection in hotel photos (OWL-ViT) and Gemini-generated questions for solo, couple and group travellers, evaluated in a user study. Master's thesis with trivago N.V.

Zero-shot object detection system for visually impaired users using CLIP, OWL-ViT, and real-time audio feedback.