Drone navigation models
Models from the thesis project LLM-driven drone navigation on edge devices. Code: see the GitHub repository.
| File | Description |
|---|---|
slm/smollm2_drone_Q8_0_imat.gguf |
SmolLM2-135M fine-tuned for drone commands, Q8_0 with importance matrix (used on the STM32MP2) |
slm/smollm2_drone_Q4_0_imat.gguf |
Same model, Q4_0 |
yolo/yolov8n_320_best.pt |
YOLOv8n, 320×320 input, 5 classes (pedestrian, BMW_X5, dog, bench, oil_barrel) |
yolo/yolov8n_320_int8.tflite |
Same detector, INT8 TFLite |
yolo/stm32mp2/yolov8n_320_quant_pt_uf_od_custom.nb |
Same detector compiled for the STM32MP2 NPU |
yolo/rockchip/yolov5s-640-640.rknn |
YOLOv5s, 640×640 input, for the Rockchip NPU |
Training data: edoardochiavazza/drone-datasets.
Usage
hf download edoardochiavazza/drone-navigation-models --local-dir models
llama-server -m models/slm/smollm2_drone_Q8_0_imat.gguf --port 8080
- Downloads last month
- -
Hardware compatibility
Log In to add your hardware
4-bit
8-bit
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support