I'm working on a model trainer that supports per-epoch and per-batch callbacks, progress bars, mini-batching, etc. Basically, just a higher-level wrapper around the training loops in the examples in https://github.com/tensorflow/swift-models
Is this important, and is this aligned with the vision of where this high-level library wants to go?
I'm working on a model trainer that supports per-epoch and per-batch callbacks, progress bars, mini-batching, etc. Basically, just a higher-level wrapper around the training loops in the examples in https://github.com/tensorflow/swift-models
Is this important, and is this aligned with the vision of where this high-level library wants to go?