There are two integration modes for connecting an agent to LiveAvatar. This repo implements Flow 1 end-to-end and leaves the door open for Flow 2.
You are given a randomly-initialized ReLU MLP and a FLOP budget. Your job: predict the per-neuron mean activation under N(0, 1) input, without running anywhere near the budget's worth of forward ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results