FW.run

FW.run(callback=None, **kw) → None

Run the Frank-Wolfe algorithm.

Parameters

callbackcallable, optional

Called after each iteration with the signature callback(res), where res is a OptimizeResult.

kwkeyword arguments

Further options, see FWConfig.

Returns

resOptimizeResult

The optimization result represented as a OptimizeResult object. TODO-PJ: which attributes.

See also

FWConfig

Options accepted by the solver.

scipy.optimize.OptimizeResult

Details on the optimization result.