BREAKING
Zero-Shot Object Detector on GitHub
LLM Reasoning Plus Tool Use
How It Narrows the Target
1
Grid split
↓
2
Spatial reasoning
↓
3
Tool calls
↓
4
Bounding box
Operating Modes by Use Case
advanced_reasoning
default
●
Tool use plus grid
●
Highest accuracy
vision_model
●
GroundingDINO plus SAM
●
Vision-only backend
0
stars
0
forks
Benchmarks Still Awaited
AI NEWS BLITZ
A new open-source tool uses LLM reasoning to detect objects in images without any training.