Thanks you very much for your work!
And could you tell me the details about your training:
How about the batch size setting in your training. Unfortunately, my GPU can only set 16, 24 is Out Of Memory, and I get 72.5 (yours is 72.9) AP in the COCO minival dataset. And, I think the larger batch size can get better performance.
How many GPUs used and how much time spent for your training ? I used 4 GPUs, spending 3.x days.
What's more, how about the memory of your one GPU card?
I trained these models with the setting in code. Batch_size can get a little better performance because of BN. But when I trained resnet101.256x192 model, I got better results with 24 batch_size than 44 batch_size. It's weird.
All models are trained on 8 GPUs.
For resnet101 and 384x288 inputs, I trained it with Nvidia P40 (24GB memory). For others, I trained it with Nvidia Pascal GPU (12GB memory).
Thanks you very much for your work! And could you tell me the details about your training:
Thanks!