TorchDynamo: An Experiment in Dynamic Python Bytecode Transformation

[fortianyou]

@jansel

we laid out a vision of deploying eager mode PyTorch to more production settings and investing in using compilers to make eager mode faster and easier to maintain.

Does that mean TorchDynamo only supports the scenarios in Inference? What is the roadmap of TorchDynamo to support PyTorch compilers for training?