//
sign in
Post
by @danabra.mov
PostEmbed
by @danabra.mov
Record
by @jimpick.com
Record
by @atsui.org
+ new component
Post
The weirdest observation: I generated movies visualizing the polytope boundaries for ReLU networks using Muon and AdamW. Same experiment, same data, same random seed. The difference is the "crease pattern" that the optimizers produce.