com.intel.analytics.zoo.pipeline.api.keras.optimizers
learning rate
portion of total for the warmup, -1 means no warmup. Default: -1
total number of training steps for the learning rate schedule, -1 means constant learning rate. Default: -1
schedule to use for the warmup. Default: 'linear'
first moment coefficient
second moment coefficient
for numerical stability
weight decay
first moment coefficient
second moment coefficient
for numerical stability
learning rate
a function that takes a single input (X), the point of a evaluation, and returns f(X) and df/dX
the initial point
the new x vector and the function list {fx}, evaluated before the update
schedule to use for the warmup.
schedule to use for the warmup. Default: 'linear'
total number of training steps for the learning rate schedule, -1 means constant learning rate.
total number of training steps for the learning rate schedule, -1 means constant learning rate. Default: -1
portion of total for the warmup, -1 means no warmup.
portion of total for the warmup, -1 means no warmup. Default: -1
weight decay
weight decay
(Since version 0.2.0) Please use clearHistory() instead
(Since version 0.2.0) Please initialize OptimMethod with parameters when creating it instead of importing table
Implements BERT version of Adam algorithm