3 ms·
This sounds like it could explain many ML models ;)
by bflesch 5y ago
This sounds like it could explain many ML models ;)
- mr_toad 5y agoIt’s actually possible in theory for part of a neural net to simply encode the training data and use it to regurgitate responses. In practice, more efficient parts of the network should win in the end.
- NavinF 5y agoHas nobody in this thread heard of https://en.m.wikipedia.org/wiki/Cross-validation_(statistics) https://en.m.wikipedia.org/wiki/Cross-validation_(statistics... Every ml tutorial start with splitting your data in into a training set and validation set.