Learning-NVDLA-Notes icon indicating copy to clipboard operation
Learning-NVDLA-Notes copied to clipboard

Published 20 hours ago • verified publisher icon JunningWu

Object detection

Open DrVijayK opened this issue 6 years ago • 4 comments

@JunningWu Excellent work and it is very helpful. Can you please let me know if you were able to successfully perform object detection using NVDLA with any CNN. If so, it would be of great help if you can let me know the details. Waiting for your response.

DrVijayK avatar Feb 05 '19 16:02 DrVijayK

supporting for other CNNs is heavily depending on NVIDIA's compiler, currently only some few networks.

JunningWu avatar Feb 14 '19 03:02 JunningWu

@JunningWu Thanks for the reply. So far we were not able to successfully do object detection with NVDLA. If you were able to do that, can you please let us know details on which CNN etc. We already tried Yolov1 tiny and it doesn't work with NVDLA. The VP output is givinh very low probabilities, wrong classifications and wrong coordinates of bounding boxes.

DrVijayK avatar Feb 14 '19 09:02 DrVijayK

https://github.com/JunningWu/Learning-NVDLA-Notes/wiki/Learning-NVDLA-Notes-by-Junning#551-%E4%BD%BF%E7%94%A8vp%E4%BB%BF%E7%9C%9Fresnet-18

JunningWu avatar May 08 '19 02:05 JunningWu

https://github.com/JunningWu/Learning-NVDLA-Notes/wiki/Learning-NVDLA-Notes-by-Junning#551-%E4%BD%BF%E7%94%A8vp%E4%BB%BF%E7%9C%9Fresnet-18

English please?

shazib-summar avatar Oct 16 '19 16:10 shazib-summar