2 ms·
Sounds like an interesting way to benchmark models. Put two instances of a model in a loop with itself. Run for n cycles. Then look at the output.
by leobg 20d ago
Sounds like an interesting way to benchmark models. Put two instances of a model in a loop with itself. Run for n cycles. Then look at the output.