Move x2num preprocessings into the summary.py

[ruotianluo]

Now all the x2num preprocessings are in writer.py, like:

def add_scalar(self, tag, scalar_value, global_step=None):
        """Add scalar data to summary.
        Args:
            tag (string): Data identifier
            scalar_value (float): Value to save
            global_step (int): Global step value to record
        """
        scalar_value = makenp(scalar_value)
        assert(scalar_value.squeeze().ndim==0), 'input of add_scalar should be 0D'
        self.file_writer.add_summary(scalar(tag, scalar_value), global_step)

Why not put scalar_value = makenp(scalar_value) in scalar function.

The reason I'm proposing this, is I'm using FileWriter, and I collect the summaries during forwarding the network. (Like the pipeline in tensorflow) So I will call the scalar histogram etc functions by myself. (The reason I don't use SummaryWriter is I don't want the SummaryWriter to be exposed to the model itself.)

Now, to call scalar or histogram I have to convert to numpy first, which is ok but not necessary because you can move all the x2num to summary.py.

What do you think.

[lanpa]

Looks good. Would you like to send a PR?

[ruotianluo]

I will do this.