Problem about pytorch1.1 to pytorch1.7

I successfully run forward in pytorch1.7, and the weight trained in pytorch1.7 appeared noise in the images after run forward.The weight trained on pytorch1.1 is no problem after run forward with pytorch1.7. Do you have any suggestions for me to find the problem? Thank you very much.

Could you explain your issue a bit more, please?
Are you able to train a model in a specific version, but the inference in another PyTorch release changes?