Add logger for hiopbbpy #739
Conversation
|
here is one example of the log: hiopbbpy Problem name: LpNormProblem |
|
|
||
| y_train_virtual = y_train.copy() # old training + batch_size num of virtual points | ||
| for j in range(self.batch_size): | ||
| # Get a new sample point | ||
| self.logger.scalars(f"In batch {j+1}/{self.batch_size}") |
There was a problem hiding this comment.
This is not the batch number but rather the jth point of a given batch. In batch {j+1}/{self.batch_size} is misleading.
There was a problem hiding this comment.
I don't get this point. j in range(self.batch_size) --> should j mean the batch number? in my example above, this line prints the lines of
hiopbbpy In batch 1/2
hiopbbpy In batch 2/2
| self.logger.critical("===================================") | ||
| self.logger.critical("Bayesian Optimization completed") | ||
| self.logger.critical(f"Total evaluations: {len(self.y_hist)}") | ||
| self.logger.critical(f"Optimal at BO iteration: {self.idx_opt//self.batch_size+1} ") |
There was a problem hiding this comment.
When the batch_size is self.y_hist are obtained in the first BO iteration and so self.idx_opt being
There was a problem hiding this comment.
I don't get this.
if batch_size = 5, and idx_opt = 2, it means the optimal point was obtained in the 1st BO iteration, doesn't it?
The equation used here is a floor division, and hence {self.idx_opt//self.batch_size+1} = 2//5+1 = 0+1 = 1.
nychiang
force-pushed
the
hiopbbpy-log-dev
branch
from
4a0faee to
95014bf
Compare
implement a logger using the code from https://docs.python.org/3/library/logging.html
CLOSE #717