Update README.md

updated .gitignore

updated to support inference with GPU

fixed device code

added eval() to model creation

added inference timing to logs

updated neo nb

added layer names to onnx model

updated src files

updated to support correct neo pre and post processing

updated train.py

first test with neo

deploy_pytorch_model.ipynb

added example of deploying with jit to standard PyTorch inference container

made jit version work