Closed skeggse closed 6 years ago
@skeggse You're right, code examples are definitely missing. I intend to simplify the API even more and then add examples for various problems and networks, such as CNN, RNN, autoencoders, etc. Will get to it as soon as I have some free time.
What is your use case?
I need to perform nested cross-validation for hyperparameter/model selection and model evaluation, and would like to minimize the work needed to run the inner cross-validation with hyper-engine. Depending on how the tensorflow bindings are implemented, I may not be able to leverage the cluster I have access to appropriately, so my use-case is likely just a module I use to early stop training based on learning curve prediction, and to schedule the right jobs with appropriate hyperparameters outside of an actual script running tensorflow.
Is there anything I can help with as far as implementation goes? I just need a bit of internal documentation to understand what the various parts of hyper-engine are and how they fit together. I'd love to contribute!
Hi @skeggse Thanks for your interest. I've reworked API and uploaded a bunch of examples this weekend, and plan to add more. Please take a look to get a sense how it works in theory. I'm ready to discuss the architecture.
Amazing, many thanks! I'm taking a look at the examples now. I'll get back to you soon on architecture ideas.
I'm also closing this issue as it pertains specifically to usage examples, and the new examples are super helpful.
I found this project from your post on stackoverflow. I'd like to experiment with it, but I'm unclear on how to use it - there isn't any obvious documentation describing the use of hyper-engine.
The README says hyper-engine provides a tensorflow binding. Depending on what that binding entails, it may or may not accommodate my use-case.