abhijitbetigeri/dc-ops-dataset
DC-Ops: Data Center Components Dataset On-device data center operations assistant dataset for the Qualcomm x Meta ExecuTorch Hackathon. Overview 319 images of data center infrastructure (server racks, NVIDIA NVL72, cables, ports, etc.) 3,045 polygon annotations in YOLO-seg format 16 component classes auto-labeled with Grounding DINO + SAM, for fine-tuning YOLOv8n-seg Classes ID Class Count 0 server rack rack enclosures 1 compute tray… See the full description on the dataset page: https://huggingface.co/datasets/abhijitbetigeri/dc-ops-dataset.
Add YOLOv8n-seg QNN HTP .pte for Snapdragon NPU (INT8, 3.2MB, raw-map export)
Add RetinaNet state dict .pth (140MB)
Add RetinaNet full model .pt (140MB, loss=0.85, with moire augmentation)
Add normalized confusion matrix
Add confusion matrix for YOLOv8n-seg v3
Add RetinaNet QNN HTP .pte for Snapdragon 8 Elite NPU (36MB)
Add ExecuTorch .pte v3 for Snapdragon NPU (mAP50=0.749)
Add YOLOv8n-seg v3 trained on 2036 Roboflow images (mAP50=0.749)
Add ExecuTorch .pte model for Snapdragon NPU
Add dataset card
Add YOLOv8n-seg trained weights (6.5MB)
Add YOLO-seg polygon labels (16 DC classes, 3045 objects)
Add 319 DC infrastructure images
initial commit