add gradnorm and run_gradnorm example
modify mtl_trainer and utils\mtl.py
references：
(1)Chen Z, Badrinarayanan V, Lee C Y, et al. Gradnorm: Gradient normalization for adaptive loss balancing in deep multitask networks[C]//International conference on machine learning. PMLR, 2018: 794-803.
(2)https://github.com/brianlan/pytorch-grad-norm
add gradnorm and run_gradnorm example modify mtl_trainer and utils\mtl.py
references： (1)Chen Z, Badrinarayanan V, Lee C Y, et al. Gradnorm: Gradient normalization for adaptive loss balancing in deep multitask networks[C]//International conference on machine learning. PMLR, 2018: 794-803. (2)https://github.com/brianlan/pytorch-grad-norm