Linengyao

when i train deformable-detr model at 2080Ti(11G) platform, use batch_size=2, it will tip: RuntimeError: CUDA out of memory. Tried to allocate 56.00 MiB (GPU 0; 10.76 GiB total capacity; 8.54...

I want to train from scratch on a single card, but after testing, I found that there are many errors in loading models. Does your code support training from scratch?...

I've seen your comments on weak or untextured datasets in a lot of issues. I am currently ranging sea surface targets, and do not use the pre-trained model to train...