Ubuntu 16.04, CUDA9 and caffe installed properly:

visionssddetect node runs (using provided launch) without any warnings nvidia-smi says that it uses approximately 54MB of 540M's video memory regardless of image_raw resolution (compared 1920x1200 with 320x200). COCO model was used.

But nevertheless, visionssddetect does not return any detected objects:

rostopic echo /detection/imagedetector/objects command returns messages with empty list of objects, here's full message example: header: seq: 426 stamp: secs: 1562840561 nsecs: 831216836 frameid: "camfcnear" objects: []