jiaxiang-cheng/PyTorch-Transformer-for-RUL-Prediction ? reverse-engineered prompt

Build me a Python project that predicts remaining useful life for turbofan engines using the NASA CMAPSS data. I want it to train a PyTorch model that uses a Transformer encoder with some convolution layers, then reports training loss and test RMSE as it learns.

Please include code to load the CMAPSS files, prepare the sequences, add the remaining useful life labels, train the model, and run testing. It should be easy to start with something like one dataset such as FD001, and I’d like a simple way to switch to the others too.

Also add a small visualization script so I can look at the results and compare predictions with the ground truth. Keep it practical and runnable from the command line, and if you need to check anything current in the docs or package behavior, look it up online first.