Update README with installation guide and model table

Fix LayerDecayOptimizer applying decay every step instead of once

Add ImageNet normalization to inference preprocessing

Add grayscale image handling and fix type hints for Python 3.8 compatibility

Fix resize function returning None

Ensure model is in eval() mode during inference

Create LICENSE

.gitignore

Fix minor error

Converting to onnx

Delete pycache

Add validation loss

Add amp training

Fix minor error for training

Add lr scheduler (multi-step-lr, warm-up)