Unanswered
Hi! I Am Having Some Problems With A Loss After A Good Amount Of Training, What Would Be The Best Way To Log A Value To Have A Better Idea Of What Is Happening?
I would do something like:
` from clearml import Logger
def forward(...):
self.iteration += 1
weights = self.compute_weights(...)
m = (weights * (target-preds)).mean()
Logger.current_logger().report_scalar(title="debug", series="mean_weight", value=m, iteration=self.iteration)
return m `
169 Views
0
Answers
2 years ago
one year ago