How about the training details? about tf-cpn

[chenyilun95]

1. 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.
2. All models are trained on 8 GPUs.
3. For resnet101 and 384x288 inputs, I trained it with Nvidia P40 (24GB memory). For others, I trained it with Nvidia Pascal GPU (12GB memory).

[7color94]

Got it. Thank you very much.