Shi hua

you need to modify the config file ` MODEL: WEIGHTS: "your_path/dla pretrained model" `

build_fcos_efficientnet_fpn_backbone epoch:6 batchsize:8 gpus:4 coco results: | AP | AP50 | AP75 | APs | APm | APl | |:------:|:------:|:------:|:------:|:------:|:------:| | 29.017 | 45.281 | 30.911 | 16.333 | 31.203...

sorry, I have no time to do this, it is easy to measure the speed and you can test the model yourself. ``` def compute_speed(model, input_size, device, iteration): print('input_size', input_size)...

In this project, I use 4 GPUs

Could you paste the error messages？

Thanks for your issue, This project is not intended to reproduce the results of the paper, my goal is only to achieve some of the realizable tricks in the paper.

I tried to convert the TensorFlow models to PyTroch, but failed，because I seldom use TensorFlow before. I will use PyTroch to reproduce the results. (if Google open the source code)

Generate 128 resolution image(batch size 64) will use about 10G memory per GPU, so I don't have enough machines to try ...