domainbed_measures/measures/held_out_measures.py [286:307]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
                  weight_decay_min=1e-4,
                  callbacks=None,
                  train_split=None):
        """Get heads for optimization.

        Set the maximum learning rate to start from for each head, and multiple
        heads then have a learning rate that is lr_sweep_factor^(i-1) * max_lr
        for the i'th head that we consider.
        """
        if callbacks == None:
            callbacks = []

        heads = []
        for hidx in range(num_batch_heads):
            lr = max_lr * lr_sweep_factor**(hidx)
            weight_decay = float(torch.multinomial(torch.logspace(1, -3, 5),
                                                   1))

            logging.info(
                f"Creating head {hidx} with lr {lr}, weight decay {weight_decay}"
            )
            if self._v_plus == True:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



domainbed_measures/measures/held_out_measures.py [646:667]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
                      weight_decay_min=1e-4,
                      callbacks=None,
                      train_split=None):
        """Get heads for optimization.

        Set the maximum learning rate to start from for each head, and multiple
        heads then have a learning rate that is lr_sweep_factor^(i-1) * max_lr
        for the i'th head that we consider.
        """
        if callbacks == None:
            callbacks = []

        heads = []
        for hidx in range(num_batch_heads):
            lr = max_lr * lr_sweep_factor**(hidx)
            weight_decay = float(torch.multinomial(torch.logspace(1, -3, 5),
                                                   1))

            logging.info(
                f"Creating head {hidx} with lr {lr}, weight decay {weight_decay}"
            )
            if self._v_plus == True:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -