v1.1

What's new:

• A branch of new methods: WeaSEL, ImplyLoss, ASTRA, MeanTeacher, Meta-Weight-Net, Learning-to-Reweight
• A new EndClassifierModel model which unifies all the classification backbones
• Two new datasets on image classification
• Support torch native amp for inference in the validation step
• Support training on multiple GPUS via torch's DistributedDataParallel and the new parallel_fit function
• fixed some bugs