Whether to apply learning rate warmup to this update rule. When true, the learning rate (eta) is scaled by (i+1)/n_iterations.
URI: tvbo:slot/warmup
name: warmup description: Whether to apply learning rate warmup to this update rule. When true, the learning rate (eta) is scaled by (i+1)/n_iterations. from_schema: https://w3id.org/tvbo rank: 1000 owner: UpdateRule domain_of: - UpdateRule range: boolean