An example of implementing an MLP to fit sin(x)

1. Create a dataset with registration: ./data/sin_example_dataset.py. Import it in ./data/__init__.py.

2. Create a model with registration: ./models/sin_example_mlp.py, and import it in ./models/__init__.py.

3. Create a trainer with registration: ./trainer/sin_example_trainer.py, and import it in ./trainer/__init__.py.

4. Create a config: ./configs/sin_example_config.yaml.

5. Run training: GPU=0 bash scripts/train.sh configs/sin_example_config.yaml. If multiple GPU are specified (e.g. GPU=0,1), distributed parallel training is enabled.