Closed tangjiuqi097 closed 3 years ago
Hi,
We followed the FPS calculation code provided by detr: https://github.com/facebookresearch/detr/issues/110, and get the following results:
Experiemnt enviornment: 1* Tesla V100 GPU, CUDA 11.0, PyTorch 1.7.0, torchscript not enabled
Results: DETR-DC5-R50: 10.204 FPS Conditional DETR-DC5-R50: 9.708FPS
The inference speed is pretty close. Hope I answered your question.
Thank you very much.
Hi, I am interested in the speed of the model Conditional DETR-DC5-R50.