Closed ginward closed 4 years ago
@ginward Yes, it can be done. For that, you need to write a custom output
from livelossplot import PlotLosses
from livelossplot.outputs import MatplotlibPlot, ExtramaPrinter, BaseOutput
class YourPrinter(BaseOutput):
def __init__(self):
self.iteration = 0
def send(self, logger):
self.iteration += 1
print(f"{self.iteration}")
plotlosses = PlotLosses(outputs=[MatplotlibPlot(), ExtramaPrinter(), YourPrinter()])
See https://github.com/stared/livelossplot/blob/master/livelossplot/outputs/extrema_printer.py for inspiration.
For numeric values, you can also add batch size and the learning rate to logged variables.
Let me know it if helps.
Ok thanks!
🚀 Feature
Is it possible to add a feature that can send a custom message to display below the plots? For example, I am doing some hyper parameter search and it would be nice to display the current batch size, learning rate, etc.
Thanks!