Project translated automatically.
Log in
or
register,
to view the original
Hello, you need to create a model capable of scaling to approximate the function.
Here is a set of data from 3 columns, the third can be removed. You need to look for the dependency between these two columns. Predict the first column on the basis of the second. Data in decimal format, I think it is worth the data to be presented in binary and adding noise in some lines to the total length of the longest line in the column, which in both should not be more than 256 bits. And it is possible to imagine in the vector space that the neuron does not issue the error of endless signs and NAN in the MSE. If you can create a trained and scalable algorithm, you are ready to rent a server for 2 weeks with 8 video cards of nvidia a10 +-. Whoever is ready to cooperate, I will be happy to communicate.
HTTPS://drive.google.com/file/d/1TipoaRzUEISNMdgpaSiDCEkPX0ikM0uV/view?usp=drivesdk