Open Korbrent opened 4 weeks ago
Implemented hyper tuning however after leaving running for 8 hours on my laptop, I've deduced this will take FOREVER on a CPU. Last night once I got home, I set up a new Linux distro on my desktop and got Cudas configured for GPU training. Unsurprisingly, much faster. However, two issues arise:
Next thing to do is figure out how to only auto save the best 5 weights. If possible, disabling the warnings would be pretty nice too.
Using the
``keras_tuner
API, implement hyper tuning so that our model can be tuned and optimized for our data.